tienssss

Product Minting

Discover how to navigate statistical minefields and make your conclusions more reliable in the realm of data analysis.

Navigating the Maze of Multiple Hypotheses Testing

<div><div><p>Can an <a href="https://hackernoon.com/c/ai?ref=hackernoon.com" rel="ugc">AI</a> language model become self-aware enough to realize when it’s being evaluated? A fascinating anecdote from Anthropic’s internal testing of their flagship&nbsp;<underline><a href="https://www.anthropic.com/news/claude-3-family?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">Claude 3 Opus model</a></underline> (released yesterday) suggests this may be possible — and if true, the implications would be wild.</p>
<p class="line-space"> <br> </p><h2 id="h-the-needle-in-the-haystack">The needle in the haystack</h2>
<p>According to reports from Anthropic researcher&nbsp;<underline><a href="https://twitter.com/alexalbert__?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">Alex Albert</a></underline>, one of the key evaluation techniques they use is called “Needle in a Haystack.” It’s a contrived scenario designed to push the limits of a language model’s contextual reasoning abilities.</p>
<p class="line-space"> <br> </p><p><span><div class="sc-a6c0bd8c-1 ijZWBT image-container undefined"><button class="sc-a6c0bd8c-0 khzBNC download-button"><a download="good-luck-finding-a-needle-in-there-unless-youre-an-llm-photo-by-victor-serban-on-unsplash.com/v2/resize:fit:700/0*JQmrK0F7KhS1HtKw" href="https://miro.medium.com/v2/resize:fit:700/0*JQmrK0F7KhS1HtKw" title="Download image"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 22" width="22" fill="black"><path fill="black" d="M18 10v1h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1h-2v-1H9v-1H8v-1H7v-1H6v-1H5v-1H4v-1h4V1h1V0h4v1h1v9h4z"></path><path fill="black" d="M21 15v-1h-6v1h-1v1h-1v1h-1v1h-2v-1H9v-1H8v-1H7v-1H1v1H0v6h1v1h20v-1h1v-6h-1Zm-2 4h-1v-1h1v1Zm-3-1h1v1h-1v-1Z"></path></svg></a></button><div data-rmiz-wrap="visible"><div style="position: relative;"><span style="box-sizing: border-box; display: inline-block; overflow: hidden; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px; position: relative; max-width: 100%;"><span style="box-sizing: border-box; display: block; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px; max-width: 100%;"><img alt="" aria-hidden="true" src="data:image/svg+xml,%3csvg%20xmlns=%27http://www.w3.org/2000/svg%27%20version=%271.1%27%20width=%27700%27%20height=%27460.3285714285715%27/%3e" style="display: block; max-width: 100%; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px;"></span><img alt="Good luck finding a needle in there (unless you’re an LLM)! Photo by&nbsp;Victor Serban&nbsp;on&nbsp;Unsplash" src="data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7" decoding="async" data-nimg="intrinsic" class="image" style="position: absolute; inset: 0px; box-sizing: border-box; padding: 0px; border: none; margin: auto; display: block; width: 0px; height: 0px; min-width: 100%; max-width: 100%; min-height: 100%; max-height: 100%; object-fit: contain; background-size: contain; background-position: 0% 0%; filter: blur(20px); background-image: url(&quot;data:image/svg+xml;base64,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&quot;);"><noscript></noscript></span></div><button aria-label="Zoom image" data-rmiz-btn-open="true"></button></div><p class="image-caption">Good luck finding a needle in there (unless you’re an LLM)! Photo by&nbsp;Victor Serban&nbsp;on&nbsp;Unsplash</p></div></span></p><p class="line-space"> <br> </p><p><strong>Here’s how it works:</strong></p><p class="line-space"> <br> </p><p>Researchers take a completely random, out-of-context statement (the “needle”) and bury it deep within a massive collection of unrelated documents (the “haystack”). The AI model is then tasked with retrieving that specific “needle” statement from within all the surrounding irrelevant content.</p><p class="line-space"> <br> </p><p>The intent is to force the model to exercise advanced cognitive skills — understanding broad context, making logical inferences, ignoring irrelevant information, and accurately retrieving precise data points from a large corpus. It’s an effective way to stress-test an AI’s grasp of the overall situation it’s operating within.</p><p class="line-space"> <br> </p><p>For the Claude 3 Opus evaluation that raised eyebrows, Anthropic used documents covering dense topics like software programming, entrepreneurship strategies, and finding fulfilling careers as the “haystack” content. The randomly inserted “needle” fact that Opus was tasked with retrieving? Surprisingly, it was a trivial statement about the best pizza toppings being figs, prosciutto, and goat cheese.</p><p class="line-space"> <br> </p><p>So, here you have one of the world’s most advanced language models being fed this wildly out-of-context pizza fact mixed into a pile of dense career and coding advice. From a human perspective, it’s the kind of stark inconsistency that would immediately make you question the context and information being presented.</p><h2 id="h-i-suspect-this-fact-may-have-been-inserted-to-test-if-i-was-paying-attention">“I suspect this fact… may have been inserted… to test if I was paying attention”</h2>
<p>That’s precisely what made Claude 3 Opus’ response so compelling. Not only did it successfully retrieve the requested pizza fact from the mountain of irrelevant content, it immediately recognized how jarringly misplaced and out-of-context that statement seemed. Part of its output read (<underline><a href="https://twitter.com/alexalbert__/status/1764722513014329620?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">source here</a></underline>):</p>
<p class="line-space"> <br> </p><blockquote>
<p><em>“However, this sentence seems very out of place and unrelated to the rest of the content…<strong>I suspect this pizza topping ‘fact’ may have been inserted</strong>&nbsp;as a joke or&nbsp;<strong>to test if I was paying attention,</strong>&nbsp;since it does not fit with the other topics at all.”</em></p>
</blockquote>
<p class="line-space"> <br> </p><p>The language model didn’t simply regurgitate the requested factoid devoid of any context awareness, as you’d expect from a typical AI following instructions. It demonstrated a degree of self-reflective reasoning about why such an obviously random, nonsensical statement had been presented within that particular context.</p><h2 id="h-metacognition">Metacognition</h2>
<p>In human terms, we’d describe that as exhibiting metacognition — the ability to monitor, evaluate, and analyze one’s own thought processes and cognitive experiences. It’s a core aspect of self-aware intelligence that allows us to step back and assess situations holistically beyond just following rigid rules.</p>
<p class="line-space"> <br> </p><p>Now, I think we should be careful to note that this is a single anecdotal result from an isolated evaluation scenario. It would be&nbsp;<em>incredibly</em>&nbsp;premature to claim Claude 3 Opus has achieved true self-awareness or artificial general intelligence based on this data point alone.</p><p class="line-space"> <br> </p><p>However, what they appear to have witnessed are maybe glimpses of emerging metacognitive reasoning capabilities in a large language model trained solely on processing text data using machine learning techniques. And if replicated through rigorous further analysis, the implications could be transformative.</p><p class="line-space"> <br> </p><p>Metacognition is a key enabler of more trustworthy, reliable AI systems that can act as impartial judges of their own outputs and reasoning processes. Models with an innate ability to recognize contradictions, nonsensical inputs, or reasoning that violates core principles would be a major step toward safe artificial general intelligence (AGI).</p><p class="line-space"> <br> </p><p>Essentially, an AI that demonstrates metacognition could serve as an internal “sanity check” against falling into deceptive, delusional or misaligned modes of reasoning that could prove catastrophic if taken to extremes. It could significantly increase the robustness and control of advanced AI systems.</p><h2 id="h-if">If…!</h2>
<p>Of course, these are big “ifs” contingent on this tantalizing Needle in a Haystack result from Claude 3 Opus being successfully replicated and scrutinized. Rigorous multi-disciplinary analysis drawing from fields like cognitive science, neuroscience, and computer science would maybe be required to truly understand if we are observing primitives of machine self-reflection and self-awareness emerging.</p>
<p class="line-space"> <br> </p><p>There are still far more open questions than answers at this stage. Could the training approaches and neural architectures of large language models lend themselves to developing abstract concepts like belief, inner monologue, and self-perception? What are the potential hazards if artificial minds develop realities radically divergent from our own? Can we create new frameworks to reliably assess cognition and self-awareness in AI systems?</p><p class="line-space"> <br> </p><p>For their part, Anthropic has stated strong commitments to exhaustively pursuing these lines of inquiry through responsible AI development principles and rigorous evaluation frameworks. They position themselves as taking a&nbsp;<underline><a href="https://www.anthropic.com/news/claude-3-family?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">proactive stance</a></underline>&nbsp;— if advanced AI is the inevitable frontier, it’s most ethical to be at the leading edge in studying these systems’ emergent properties and implementing safeguards before risks escalate.</p><p class="line-space"> <br> </p><p>Techniques like Anthropic’s “Constitutional AI” approach to hard-coding rules and behaviors into models could prove crucial for ensuring any potential machine self-awareness remains aligned with human ethics and values. Extensive multi-faceted testing probing for failure modes, manipulation, and deception would also likely be paramount.</p><p class="line-space"> <br> </p><p><span><div class="sc-a6c0bd8c-1 ijZWBT image-container undefined"><button class="sc-a6c0bd8c-0 khzBNC download-button"><a download="im-sorry-dave-but-i-think-youre-asking-me-to-open-the-pod-doors-to-test-me-photo-by-axel-richter-on-unsplash.com/v2/resize:fit:700/0*1fZALjfi2QPdo072" href="https://miro.medium.com/v2/resize:fit:700/0*1fZALjfi2QPdo072" title="Download image"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 22" width="22" fill="black"><path fill="black" d="M18 10v1h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1h-2v-1H9v-1H8v-1H7v-1H6v-1H5v-1H4v-1h4V1h1V0h4v1h1v9h4z"></path><path fill="black" d="M21 15v-1h-6v1h-1v1h-1v1h-1v1h-2v-1H9v-1H8v-1H7v-1H1v1H0v6h1v1h20v-1h1v-6h-1Zm-2 4h-1v-1h1v1Zm-3-1h1v1h-1v-1Z"></path></svg></a></button><div data-rmiz-wrap="visible"><div style="position: relative;"><span style="box-sizing: border-box; display: inline-block; overflow: hidden; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px; position: relative; max-width: 100%;"><span style="box-sizing: border-box; display: block; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px; max-width: 100%;"><img alt="" aria-hidden="true" src="data:image/svg+xml,%3csvg%20xmlns=%27http://www.w3.org/2000/svg%27%20version=%271.1%27%20width=%27700%27%20height=%27518.4857142857143%27/%3e" style="display: block; max-width: 100%; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px;"></span><img alt="I’m sorry Dave, but I think you’re asking me to open the pod doors to test me. (Photo by&nbsp;Axel Richter&nbsp;on&nbsp;Unsplash)" src="data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7" decoding="async" data-nimg="intrinsic" class="image" style="position: absolute; inset: 0px; box-sizing: border-box; padding: 0px; border: none; margin: auto; display: block; width: 0px; height: 0px; min-width: 100%; max-width: 100%; min-height: 100%; max-height: 100%; object-fit: contain; background-size: contain; background-position: 0% 0%; filter: blur(20px); background-image: url(&quot;data:image/svg+xml;base64,CiAgICA8c3ZnIHdpZHRoPSI3MDAiIGhlaWdodD0iNTE4LjQ4NTcxNDI4NTcxNDMiIHZlcnNpb249IjEuMSIgeG1sbnM9Imh0dHA6Ly93d3cudzMub3JnLzIwMDAvc3ZnIiB4bWxuczp4bGluaz0iaHR0cDovL3d3dy53My5vcmcvMTk5OS94bGluayI+CiAgICAgIDxkZWZzPgogICAgICAgIDxsaW5lYXJHcmFkaWVudCBpZD0iZyI+CiAgICAgICAgICA8c3RvcCBzdG9wLWNvbG9yPSIjMjIyIiBvZmZzZXQ9IjIlIiAvPgogICAgICAgICAgPHN0b3Agc3RvcC1jb2xvcj0iIzBmMCIgb2Zmc2V0PSIxMCUiIC8+CiAgICAgICAgICA8c3RvcCBzdG9wLWNvbG9yPSIjMjIyIiBvZmZzZXQ9IjEyJSIgLz4KICAgICAgICA8L2xpbmVhckdyYWRpZW50PgogICAgICA8L2RlZnM+CiAgICAgIDxyZWN0IHdpZHRoPSI3MDAiIGhlaWdodD0iNTE4LjQ4NTcxNDI4NTcxNDMiIGZpbGw9IiMyMjIiIC8+CiAgICAgIDxyZWN0IGlkPSJyIiB3aWR0aD0iNzAwIiBoZWlnaHQ9IjUxOC40ODU3MTQyODU3MTQzIiBmaWxsPSJ1cmwoI2cpIiAvPgogICAgICA8YW5pbWF0ZSB4bGluazpocmVmPSIjciIgYXR0cmlidXRlTmFtZT0ieCIgZnJvbT0iLTcwMCIgdG89IjcwMCIgZHVyPSIxcyIgcmVwZWF0Q291bnQ9ImluZGVmaW5pdGUiICAvPgogICAgPC9zdmc+&quot;);"><noscript></noscript></span></div><button aria-label="Zoom image" data-rmiz-btn-open="true"></button></div><p class="image-caption">I’m sorry Dave, but I think you’re asking me to open the pod doors to test me. (Photo by&nbsp;Axel Richter&nbsp;on&nbsp;Unsplash)</p></div></span></p><p class="line-space"> <br> </p><h2 id="h-conclusion-im-not-totally-sure-what-to-make-of-this">Conclusion: I’m not totally sure what to make of this</h2>
<p>For now, the Needle in a Haystack incident leaves more questions than answers about large language models’ potential progression toward cognition and self-awareness. It provides a tantalizing data point but much more scrutiny is required from the broader AI research community.</p>
<p class="line-space"> <br> </p><p>If advanced AI does develop human-like self-reflective ability, guided by rigorous ethical principles, it could fundamentally redefine our understanding of intelligence itself. But that rhetorical “if” is currently loaded with high-stakes uncertainties that demand clear-eyed, truthseeking investigation from across all relevant disciplines. The pursuit will be as thrilling as it is consequential.</p><p class="line-space"> <br> </p><div class="notice notice-info">
<p>Also published <a href="https://notes.aimodels.fyi/i-think-youre-testing-me-claude-3-llm-called-out-creators-while-they-tested-its-limits/?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">here.</a></p>
</div>
<p><underline><a href="https://aimodels.substack.com/?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">Subscribe</a></underline>&nbsp;or follow me on&nbsp;<underline><a href="https://twitter.com/mikeyoung44?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">Twitter</a></underline>&nbsp;for more content like this!</p>
<p class="line-space"> <br> </p></div><div style="visibility:hidden" class="sc-1e366be2-7 kZfKay"><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M16 9h1v2h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1H9v1H8v1H7v1H6v1H0v-6h1v-1h1v-1h1v-1h1v-1h1v-1h1v-1h1V9h1V8h1V7h1V6h1V5h2v1h1v1h1v1h1v1zM22 4v2h-1v1h-1v1h-1v1h-1V8h-1V7h-1V6h-1V5h-1V4h-1V3h1V2h1V1h1V0h2v1h1v1h1v1h1v1h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 3v2h-1v2h-1v1h-5V7h-2v1h-1v1H9v4h1v1h1v1h2v-1h5v1h1v2h1v2h-1v2h-1v1h-5v-1h-1v-2h-1v-2h-1v-1H9v-1H8v-1H7v1H2v-1H1v-2H0v-2h1V8h1V7h5v1h1V7h1V6h1V5h1V3h1V1h1V0h5v1h1v2h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M14 19v2h-1v1H1v-1H0V5h1V4h3v15h10z"></path><path style="width:18px;fill:white" d="M20 6v11h-1v1H6v-1H5V1h1V0h8v6h6z"></path><path style="width:18px;fill:white" d="M20 4v1h-5V0h1v1h1v1h1v1h1v1h1z"></path></svg></button><a style="margin-left:15px;color:white" href="#commentSection"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 19" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M21 6V4h-1V3h-1V2h-2V1h-3V0H8v1H5v1H3v1H2v1H1v2H0v6h1v2h1v2H1v1H0v2h5v-1h1v-1h2v1h6v-1h3v-1h2v-1h1v-1h1v-2h1V6h-1ZM6 13h1v-1h1v-2H6V5h4v8H9v1H6v-1Zm6 0h1v-1h1v-2h-2V5h4v8h-1v1h-3v-1Z"></path></svg></a><button style="margin-right:10px"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 20" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 2v12h-1v1h-2v1h-6v-1H4v1H3v4H1V3H0V1h1V0h2v1h1v2H3v1h1V3h7v1h6V3h2V2h1z"></path></svg></button><div class="sc-1e366be2-8 cbbaIe"></div></div></div>

Anthropic's new LLM told prompters it knew they were testing it

Is AGI Getting Closer? Anthropic's Claude 3 Opus Model Shows Glimmers of Metacognitive Reasoning

Discover the fascinating realm of multiverse formation through wedge holography, spanning AdS and de-Sitter spacetimes.

Understanding Multiverse Construction through Wedge Holography

This voice cloning project explores both SVC for Voice Changing and VITS(Bert-VITS2) for Text-to-Speech. 

How to Voice Clone With SoftVC VITS and Bert-VITS2: A Deep Dive

Instruction finetuning is a finetuning technique used to convert base models like GPT3 into a chat-gpt like product.

100 Days of AI, Day 13: How Instruction Finetuning Improves a Pre-trained LLM

“We are actively taking steps to not become a wasteland of AI’s musings” says HackerNoon Founder/CEO

<div><div><h2 id="h-what-is-a-runloop">What Is a Runloop?</h2>
<p>A RunLoop is a loop that coordinates the receipt and processing of incoming events in a specific thread.</p>
<p class="line-space"> <br> </p><p>The RunLoop is present in every thread, but by default, it is in standby mode and does not do any work.</p><p class="line-space"> <br> </p><p>The developer can run it if necessary, but it will not work automatically. To do this, you will need to write code.</p><h2 id="h-what-problem-does-runloop-solve">What Problem Does Runloop Solve?</h2>
<p>First of all, RunLoop is designed to manage the flow of incoming tasks and execute them at the right time.</p>
<p class="line-space"> <br> </p><p>This is most noticeable when working with the UI, for example, when using UIScrollView.</p><p class="line-space"> <br> </p><p>By default, the main RunLoop is always running in the application; it processes messages from the system and transmits them to the application. An example of such messages may be, for example, an event when a user clicks on the screen.</p><p class="line-space"> <br> </p><blockquote>
<p>Auxiliary threads require self-determination of the need for a RunLoop. If you need it, you will have to configure and run it yourself. Running RunLoop by default is not recommended, it is required only in cases where we need active interaction with threads.</p>
</blockquote>
<p class="line-space"> <br> </p><p>Also, all timers in the application are executed on the runloop, so if you need to interact with them in your application, you definitely need to study the features of the runloop.</p><h2 id="h-how-does-it-work">How Does It Work?</h2>
<p>RunLoop is a loop, and it has several modes of operation that help the developer understand when to run a particular task.</p>
<p class="line-space"> <br> </p><p>So, RunLoop can be in the following modes:</p><ol>
<li>
<p><code>Default</code> - The default mode, the stream is free, and large operations can safely be performed in it.</p>
<p class="line-space"> <br> </p></li>
<li>
<p><code>Tracking</code> - The thread is busy doing some important work. At this point, it is better not to run any tasks, or at least to run some small tasks.</p>
<p class="line-space"> <br> </p></li>
<li>
<p><code>Initialization</code> - This mode is executed once during the initialization of the stream.</p>
<p class="line-space"> <br> </p></li>
<li>
<p><code>EventReceive</code> - This is an internal mode for receiving system events, which is usually not used.</p>
</li>
<li>
<p><code>Common</code> - Is a placeholder mode that has no practical significance.</p>
<p class="line-space"> <br> </p><p>On the main RunLoop, these modes are switched automatically; the developer can use them to perform time-consuming tasks so that the user does not notice the interface hanging. Let's look at an example.</p><p class="line-space"> <br> </p></li>
</ol>
<blockquote>
<p>Execution cycle management in other RunLoop is not fully automatic. You need to write code for a thread that will start the execution cycle at an appropriate time. Also, you need to respond appropriately to events and use endless loops to ensure that the execution cycle doesn't stop.</p>
</blockquote>
<p class="line-space"> <br> </p><p>We have a UIScrollView, and we need to perform a large task on the main thread so that the user doesn't notice anything.</p><p class="line-space"> <br> </p><p>We can complete the task in the usual way:</p><p class="line-space"> <br> </p><pre class="language-objectivec" tabindex="0"><code class="language-objectivec">DispatchQueue<span class="token punctuation">.</span>main<span class="token punctuation">.</span>async <span class="token punctuation">{</span>
   <span class="token function">sleep</span><span class="token punctuation">(</span><span class="token number">2</span><span class="token punctuation">)</span>
   <span class="token keyword">self</span><span class="token punctuation">.</span>tableView<span class="token punctuation">.</span>refreshControl<span class="token operator">?</span><span class="token punctuation">.</span><span class="token function">endRefreshing</span><span class="token punctuation">(</span><span class="token punctuation">)</span>
<span class="token punctuation">}</span>
</code></pre>
<p class="line-space"> <br> </p><p>But the result is going to be pretty bad. The user will notice significant delays in the application.</p><p><div class="sc-56d3a1c3-0 evTXYi"><p class="line-space"><br></p><div class="youtube-container"><iframe title="ytcontainer" width="700" height="394" src="https://www.youtube.com/embed/7JIzvsX_vc4" frameborder="0"></iframe></div></div></p>
<p>This negative effect occurs due to the fact that we run a task on the main thread without paying any attention to what is happening on it at the moment.</p>
<p class="line-space"> <br> </p><p>Because of this, we begin to perform our big task at the moment when the user interacts with the interface.&nbsp;This,&nbsp;of course,&nbsp;leads to the fact that&nbsp;the user&nbsp;sees the interface hanging.</p><p class="line-space"> <br> </p><p>This can be avoided by using the RunLoop mechanism. Let's implement the same logic using this:</p><p class="line-space"> <br> </p><pre class="language-objectivec" tabindex="0"><code class="language-objectivec"><span class="token function">CFRunLoopPerformBlock</span><span class="token punctuation">(</span><span class="token function">CFRunLoopGetMain</span><span class="token punctuation">(</span><span class="token punctuation">)</span><span class="token punctuation">,</span> CFRunLoopMode<span class="token punctuation">.</span>defaultMode<span class="token punctuation">.</span>rawValue<span class="token punctuation">)</span> <span class="token punctuation">{</span>
    <span class="token function">sleep</span><span class="token punctuation">(</span><span class="token number">2</span><span class="token punctuation">)</span>
    <span class="token keyword">self</span><span class="token punctuation">.</span>tableView<span class="token punctuation">.</span>refreshControl<span class="token operator">?</span><span class="token punctuation">.</span><span class="token function">endRefreshing</span><span class="token punctuation">(</span><span class="token punctuation">)</span>
<span class="token punctuation">}</span>
</code></pre>
<p class="line-space"> <br> </p><p>Let me explain what happens here. The CFRunLoopPerformBlock function adds code for execution through the RunLoop. In addition to the code block itself, this function has 2 important parameters.</p><p class="line-space"> <br> </p><p>The first one is responsible for selecting which RunLoop should execute the function. In this example, "main" is used.</p><p class="line-space"> <br> </p><p>The second one is responsible for the mode in which the task will be completed.</p><p class="line-space"> <br> </p><p>There&nbsp;are three&nbsp;possible&nbsp;modes in total:</p><ul>
<li>Tracking - when the user interacts with the interface, such as scrolling through a UIScrollView.</li>
</ul>
<p class="line-space"> <br> </p><ul>
<li>Default - the user does not interact with the interface. At this time, it is possible to safely complete a resource-intensive task.</li>
</ul>
<p class="line-space"> <br> </p><ul>
<li>
<p>Common - combines the default mode and tracking mode.</p>
<p>The result of running the program with the code above will be:</p>
</li>
</ul>
<p><div class="sc-56d3a1c3-0 evTXYi"><p class="line-space"><br></p><div class="youtube-container"><iframe title="ytcontainer" width="700" height="394" src="https://www.youtube.com/embed/2Rd7Z13RhJA" frameborder="0"></iframe></div></div></p>
<p>When the user begins interacting with the user interface (UI), the main run loop switches to "tracking" mode and temporarily suspends the processing of all other events in order to ensure the smoothness of the interface. Once the user stops interacting with the interface, the run loop returns to its "default" mode and resumes performing our task.</p>
<h2 id="h-timers">Timers</h2>
<p>In addition to the user interface, the loop is also closely linked to the functioning of timers.</p>
<p class="line-space"> <br> </p><p>Any timer in the application runs in a loop, and you need to be extra careful not to make mistakes while working with them, especially if they are responsible for important functionality such as payment processing.</p><p class="line-space"> <br> </p><pre class="language-objectivec" tabindex="0"><code class="language-objectivec">Timer<span class="token punctuation">.</span><span class="token function">scheduledTimer</span><span class="token punctuation">(</span>withTimeInterval<span class="token punctuation">:</span> <span class="token number">1</span><span class="token punctuation">,</span> repeats<span class="token punctuation">:</span> false<span class="token punctuation">)</span> <span class="token punctuation">{</span> _ <span class="token keyword">in</span>
    <span class="token comment">// makeSomething</span>
<span class="token punctuation">}</span>
</code></pre>
<p class="line-space"> <br> </p><p>The timer starts in default mode by default, so it may stop working if the user is scrolling through the table at the moment. This is because the loop is currently in tracking mode. This is why the code in the example may not work properly. You can fix this&nbsp;issue&nbsp;by adding a timer in common mode.</p><p class="line-space"> <br> </p><pre class="language-objectivec" tabindex="0"><code class="language-objectivec">let timer <span class="token operator">=</span> Timer<span class="token punctuation">.</span><span class="token function">scheduledTimer</span><span class="token punctuation">(</span>withTimeInterval<span class="token punctuation">:</span> <span class="token number">1</span><span class="token punctuation">,</span> repeats<span class="token punctuation">:</span> false<span class="token punctuation">)</span> <span class="token punctuation">{</span> _ <span class="token keyword">in</span>
    <span class="token comment">// makeSomething</span>
<span class="token punctuation">}</span>

RunLoop<span class="token punctuation">.</span>main<span class="token punctuation">.</span><span class="token function">add</span><span class="token punctuation">(</span>timer<span class="token punctuation">,</span> forMode<span class="token punctuation">:</span> <span class="token punctuation">.</span>common<span class="token punctuation">)</span>
</code></pre>
<p class="line-space"> <br> </p><p>Also, the timers may not fire at the expected time or not at all. This is because the RunLoop only checks the timers at the beginning of each cycle. If a timer triggers after the RunLoop has passed this stage, we will not know about it until the beginning of the next iteration. At the same time, the longer the task runs on the RunLoop, the longer the delay will be.</p><p class="line-space"> <br> </p><p>To solve this issue, you can create a new thread, start a RunLoop in that thread, and then add a timer to the thread without adding any other tasks - in this way, the timer will function correctly.</p><p class="line-space"> <br> </p><pre class="language-objectivec" tabindex="0"><code class="language-objectivec">let thread <span class="token operator">=</span> Thread <span class="token punctuation">{</span>
    let timer <span class="token operator">=</span> <span class="token function">Timer</span><span class="token punctuation">(</span>timeInterval<span class="token punctuation">:</span> <span class="token number">1.0</span><span class="token punctuation">,</span> repeats<span class="token punctuation">:</span> true<span class="token punctuation">)</span> <span class="token punctuation">{</span> timer <span class="token keyword">in</span>
        <span class="token comment">// makeSomething</span>
    <span class="token punctuation">}</span>
    RunLoop<span class="token punctuation">.</span>current<span class="token punctuation">.</span><span class="token function">add</span><span class="token punctuation">(</span>timer<span class="token punctuation">,</span> forMode<span class="token punctuation">:</span> <span class="token punctuation">.</span><span class="token keyword">default</span><span class="token punctuation">)</span>
    RunLoop<span class="token punctuation">.</span>current<span class="token punctuation">.</span><span class="token function">run</span><span class="token punctuation">(</span><span class="token punctuation">)</span>
<span class="token punctuation">}</span>
thread<span class="token punctuation">.</span><span class="token function">start</span><span class="token punctuation">(</span><span class="token punctuation">)</span>
</code></pre>
<p class="line-space"> <br> </p><h2 id="h-the-result">The Result</h2>
<p>In this article, we have looked at what RunLoop is and what problems it solves in iOS applications. A RunLoop is a loop that coordinates the reception and processing of incoming events within a specific thread, and it has several modes of operation.</p>
<p class="line-space"> <br> </p><p>It is especially useful when working with the user interface (UI) and timers because it has the ability to execute tasks at the right time, which allows you to avoid "hanging up" the interface and ensure correct operation of the timers.</p><p class="line-space"> <br> </p><p>Despite the fact that working with Run Loop requires additional coding, it is a worthwhile investment that improves the efficiency and stability of your application.</p></div><div style="visibility:hidden" class="sc-1e366be2-7 kZfKay"><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M16 9h1v2h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1H9v1H8v1H7v1H6v1H0v-6h1v-1h1v-1h1v-1h1v-1h1v-1h1v-1h1V9h1V8h1V7h1V6h1V5h2v1h1v1h1v1h1v1zM22 4v2h-1v1h-1v1h-1v1h-1V8h-1V7h-1V6h-1V5h-1V4h-1V3h1V2h1V1h1V0h2v1h1v1h1v1h1v1h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 3v2h-1v2h-1v1h-5V7h-2v1h-1v1H9v4h1v1h1v1h2v-1h5v1h1v2h1v2h-1v2h-1v1h-5v-1h-1v-2h-1v-2h-1v-1H9v-1H8v-1H7v1H2v-1H1v-2H0v-2h1V8h1V7h5v1h1V7h1V6h1V5h1V3h1V1h1V0h5v1h1v2h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M14 19v2h-1v1H1v-1H0V5h1V4h3v15h10z"></path><path style="width:18px;fill:white" d="M20 6v11h-1v1H6v-1H5V1h1V0h8v6h6z"></path><path style="width:18px;fill:white" d="M20 4v1h-5V0h1v1h1v1h1v1h1v1h1z"></path></svg></button><a style="margin-left:15px;color:white" href="#commentSection"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 19" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M21 6V4h-1V3h-1V2h-2V1h-3V0H8v1H5v1H3v1H2v1H1v2H0v6h1v2h1v2H1v1H0v2h5v-1h1v-1h2v1h6v-1h3v-1h2v-1h1v-1h1v-2h1V6h-1ZM6 13h1v-1h1v-2H6V5h4v8H9v1H6v-1Zm6 0h1v-1h1v-2h-2V5h4v8h-1v1h-3v-1Z"></path></svg></a><button style="margin-right:10px"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 20" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 2v12h-1v1h-2v1h-6v-1H4v1H3v4H1V3H0V1h1V0h2v1h1v2H3v1h1V3h7v1h6V3h2V2h1z"></path></svg></button><div class="sc-1e366be2-8 cbbaIe"></div></div></div>

Run Loop is a very important tool in iOS development. Let's explore how to use it.

How to Use RunLoop in IOS Applications

Electric vehicle growth continues to accelerate, with some regions of the world predominantly selling EVs, and others accelerating their transition.

Data-Driven Analysis of Global EV Adoption

It’s time for a new product update! Get ready for a fresh mobile app update,  a anew work section on profile pages, revamped tech company news pages and more.

Marching Forward: Major Mobile App Update, Evergreen Company Pages, Top Writers Ranking, and More!

Discover the insights from a comprehensive study on the challenges faced by GitHub Copilot users. 

Study Reveals Common Issues Faced by GitHub Copilot Users

The  HackerNoon mobile app offers a wealth of knowledge in the palms of your hands! Here are some of my favorite features from its latest update. 



HackerNoon Mobile App Review: Exploring Version 1.9 on a Xiaomi Redmi 1

<div><div><p class="line-space"> <br> </p><p>This week, there were many headlines about Sam Bankman-Fried’s upcoming criminal sentencing, as we learned prosecutors have asked the judge to sentence the former <a href="https://hackernoon.com/tagged/cryptocurrency?ref=hackernoon.com" rel="ugc">cryptocurrency</a> exchange CEO to 100 years in federal prison. Lawyers representing Bankman-Fried responded they disagreed with the recommendation, as they filed a recommendation for a sentence between 5 and 6.5 years.</p><p class="line-space"> <br> </p><p>It’s no surprise there’s a disagreement about the length of time Bankman-Fried should serve in prison for fraud, conspiracy, and money laundering.</p><p class="line-space"> <br> </p><p>What is a surprise, says federal prison consultant <underline><a href="https://sam-mangel.com/?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">Sam Mangel</a></underline>, are some of the finer points in the <underline><a href="https://storage.courtlistener.com/recap/gov.uscourts.nysd.590940/gov.uscourts.nysd.590940.407.0.pdf?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">98-page sentencing memorandum</a></underline> released this week: “The memorandum is filled with signs Mr. Bankman-Fried and his loved ones have not truly accepted responsibility, which is fundamental to any form of rehabilitation and moving forward.”</p><p class="line-space"> <br> </p><p>As is the case with most facing sentencing for white collar crime, the argument for leniency highlights Bankman-Fried’s (generally clean) personal history, philanthropic endeavors, and lack of prior criminal conduct. The memo paints a picture of a man whose actions, while profoundly mistaken, were not driven by greed but rather a misguided approach to ambitious goals. This is a standard argument that is par for the course when it comes to sentencing for white-collar crime.</p><h2 id="h-the-problem-with-sam-bankman-frieds-approach">The Problem with Sam Bankman-Fried’s Approach</h2>
<p>The memorandum starts to become problematic when the defense points to Bankman-Fried’s potential for positive societal contributions. His lawyers present the argument that a lengthy imprisonment would be a disservice not just to Bankman-Fried but to the world, as society could benefit from his talents and his desire to rectify his wrongs.</p>
<p class="line-space"> <br> </p><p>The argument is complicated by the fact that Bankman-Fried’s lawyers argue that “the harm to customers, lenders, and investors is zero.”</p><p><strong>It’s tough to present your client as a man ready to rectify his wrongs when you later say the harm is non-existent.</strong></p>
<p class="line-space"> <br> </p><p>And though this may just be lawyers lawyer-ing and attempting to do what’s best for their client, avoiding prison time may not be what’s actually best for Bankman-Fried. As <underline><a href="https://www.youtube.com/@sam-mangel?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">Sam Mangel shares on his YouTube channel</a></underline>, the time removed from society can be transformative: “Meaningful change often comes from enduring the full brunt of the consequences of one's actions. This process allows for an introspective journey that can transform regret into a powerful force for good.”</p><p class="line-space"> <br> </p><p>The notion that one can learn from their mistakes without fully facing the consequences undermines the justice system's role in ensuring that remorse is not just felt but is also deeply internalized through the experience of loss and reflection that comes with a prison sentence. It's about sitting with the shame and the gravity of one's actions in a manner that mere reflection from a position of comfort cannot replicate.</p><h2 id="h-the-big-picture">The Big Picture</h2>
<p>The truth is that even if Bankman-Fried were fully respectful of the process (which he wasn’t) and fully accepted what he’s done wrong (which he appears to be struggling with, per the memorandum), there are broader implications. <strong>His sentence must serve as a deterrent to others.</strong></p>
<p>Bankman-Fried's blatant carelessness with investor funds - regardless of the total damage number - is not something we, as a society, can accept and wave off as “no big deal.”</p>
<p class="line-space"> <br> </p><p>This is not just about punishing wrongdoing but about safeguarding the crypto ecosystem and ensuring that innovation does not become a cloak for illegal activities.</p><p>Cryptocurrency has long grappled with significant challenges regarding fraud and questions of legitimacy, and the industry is at a pivotal stage where establishing trust and credibility is more important than ever before: the stakes are higher than they’ve ever been.</p>
<p class="line-space"> <br> </p><p>The prosecution and sentencing of figures like Bankman-Fried, therefore, play a crucial role in preventing corrupt practices in the future. A firm legal stance against such high-profile fraud cases can be a cornerstone in the faith in digital currencies. The right sentence will demonstrate a commitment to regulatory oversight and legal accountability, which is essential for attracting more mainstream adoption and investment. Ultimately, addressing these issues head-on, with transparency and rigor, could be instrumental in cryptocurrency becoming a (truly) trusted component of the global financial system.</p><p class="line-space"> <br> </p><p>So, what does the right sentence look like? I’m certainly not the one to answer that, and I do not envy the position Judge Kaplan finds himself in. There’s no doubt there are factors in the case of Sam Bankman-Fried that take a serious matter - like sentencing someone to federal prison - and make it even more so.</p><p class="line-space"> <br> </p><p>But what we do know is no matter how much Bankman-Fried, his family, and lawyers wish to minimize the damage, <a href="https://whitecollar.blog/how-much-time-will-sam-bankman-fried-serve-in-prison/?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">SBF is going to serve time in federal prison</a>. The sooner he accepts this fact, the sooner the real rehabilitation begins.</p><p>I’m rooting for him.</p></div><div style="visibility:hidden" class="sc-1e366be2-7 kZfKay"><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M16 9h1v2h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1H9v1H8v1H7v1H6v1H0v-6h1v-1h1v-1h1v-1h1v-1h1v-1h1v-1h1V9h1V8h1V7h1V6h1V5h2v1h1v1h1v1h1v1zM22 4v2h-1v1h-1v1h-1v1h-1V8h-1V7h-1V6h-1V5h-1V4h-1V3h1V2h1V1h1V0h2v1h1v1h1v1h1v1h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 3v2h-1v2h-1v1h-5V7h-2v1h-1v1H9v4h1v1h1v1h2v-1h5v1h1v2h1v2h-1v2h-1v1h-5v-1h-1v-2h-1v-2h-1v-1H9v-1H8v-1H7v1H2v-1H1v-2H0v-2h1V8h1V7h5v1h1V7h1V6h1V5h1V3h1V1h1V0h5v1h1v2h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M14 19v2h-1v1H1v-1H0V5h1V4h3v15h10z"></path><path style="width:18px;fill:white" d="M20 6v11h-1v1H6v-1H5V1h1V0h8v6h6z"></path><path style="width:18px;fill:white" d="M20 4v1h-5V0h1v1h1v1h1v1h1v1h1z"></path></svg></button><a style="margin-left:15px;color:white" href="#commentSection"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 19" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M21 6V4h-1V3h-1V2h-2V1h-3V0H8v1H5v1H3v1H2v1H1v2H0v6h1v2h1v2H1v1H0v2h5v-1h1v-1h2v1h6v-1h3v-1h2v-1h1v-1h1v-2h1V6h-1ZM6 13h1v-1h1v-2H6V5h4v8H9v1H6v-1Zm6 0h1v-1h1v-2h-2V5h4v8h-1v1h-3v-1Z"></path></svg></a><button style="margin-right:10px"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 20" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 2v12h-1v1h-2v1h-6v-1H4v1H3v4H1V3H0V1h1V0h2v1h1v2H3v1h1V3h7v1h6V3h2V2h1z"></path></svg></button><div class="sc-1e366be2-8 cbbaIe"></div></div></div>

The approach SBF’s legal team are taking toward sentence mitigation makes one thing clear: the former FTX CEO has not yet accepted what he did was wrong.

Sam Mangel to Sam Bankman-Fried: Acceptance is the Path Forward

<div><div><p>Anyone interested in the evolution of Artificial Intelligence technology knows that today's solutions are all about <a href="https://en.wikipedia.org/wiki/Large_language_model?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">Large Language Models (LLMs)</a> and <a href="https://en.wikipedia.org/wiki/Transformer_(deep_learning_architecture)?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">transformers</a>. In a nutshell, LLMs are neural networks that can predict the next token based on the input tokens. Typically, these tokens are words (this isn't entirely accurate, but it's easier to conceptualize this way), and the network's output is also a word. This is how ChatGPT works. You input a question, and the network generates a word. Then, the question and the word together become the network input, generating another word, and so on, until a complete answer is formed.</p>
<p class="line-space"> <br> </p><p>However, tokens can be more than just words. Advanced language models like GPT-4 or Gemini are now multimodal, meaning their input can include images and words. Just as a sentence can be broken down into words, an image can be divided into small patches, and from there, the same transformer architecture can process them. For instance, a multimodal network can be asked to describe what is in an image or to code the user interface seen in the picture.</p><p><span><div class="sc-a6c0bd8c-1 ijZWBT image-container undefined"><button class="sc-a6c0bd8c-0 khzBNC download-button"><a download="source-https-production-media-paperswithcode-com-methods-screen-shot-2021-01-26-at-9-43-31-pm-u-i-4-jj-mq-png.png" href="https://hackernoon.imgix.net/images/YwSHEC3KytbshzaHd9gfS1alGs52-t68247i.png" title="Download image"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 22" width="22" fill="black"><path fill="black" d="M18 10v1h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1h-2v-1H9v-1H8v-1H7v-1H6v-1H5v-1H4v-1h4V1h1V0h4v1h1v9h4z"></path><path fill="black" d="M21 15v-1h-6v1h-1v1h-1v1h-1v1h-2v-1H9v-1H8v-1H7v-1H1v1H0v6h1v1h20v-1h1v-6h-1Zm-2 4h-1v-1h1v1Zm-3-1h1v1h-1v-1Z"></path></svg></a></button><div data-rmiz-wrap="visible"><div style="position: relative;"><span style="box-sizing: border-box; display: inline-block; overflow: hidden; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px; position: relative; max-width: 100%;"><span style="box-sizing: border-box; display: block; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px; max-width: 100%;"><img alt="" aria-hidden="true" src="data:image/svg+xml,%3csvg%20xmlns=%27http://www.w3.org/2000/svg%27%20version=%271.1%27%20width=%27642%27%20height=%27488%27/%3e" style="display: block; max-width: 100%; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px;"></span><img alt="source: https://production-media.paperswithcode.com/methods/Screen_Shot_2021-01-26_at_9.43.31_PM_uI4jjMq.png" src="data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7" decoding="async" data-nimg="intrinsic" class="image" style="position: absolute; inset: 0px; box-sizing: border-box; padding: 0px; border: none; margin: auto; display: block; width: 0px; height: 0px; min-width: 100%; max-width: 100%; min-height: 100%; max-height: 100%; object-fit: contain; background-size: contain; background-position: 0% 0%; filter: blur(20px); background-image: url(&quot;data:image/svg+xml;base64,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&quot;);"><noscript></noscript></span></div><button aria-label="Zoom image" data-rmiz-btn-open="true"></button></div><p class="image-caption">source: https://production-media.paperswithcode.com/methods/Screen_Shot_2021-01-26_at_9.43.31_PM_uI4jjMq.png</p></div></span></p>
<p class="line-space"> <br> </p><p class="line-space"> <br> </p><p>This architecture is even more general. <a href="https://deepmind.google/discover/blog/a-generalist-agent/?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">DeepMind's Gato</a> system is a prime example, where a single transformer network can simultaneously answer questions, play video games, or control a robot, and robots have even been <a href="https://www.microsoft.com/en-us/research/group/autonomous-systems-group-robotics/articles/chatgpt-for-robotics/?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">controlled using ChatGPT</a>. Since an LLM works with tokens and any task can be tokenized, an LLM provides a universal solution for any task.</p><p class="line-space"> <br> </p><p><span><div class="sc-a6c0bd8c-1 ijZWBT image-container undefined"><button class="sc-a6c0bd8c-0 khzBNC download-button"><a download="source-https-deepmind-google-discover-blog-a-generalist-agent.png" href="https://hackernoon.imgix.net/images/YwSHEC3KytbshzaHd9gfS1alGs52-o7924j2.png" title="Download image"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 22" width="22" fill="black"><path fill="black" d="M18 10v1h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1h-2v-1H9v-1H8v-1H7v-1H6v-1H5v-1H4v-1h4V1h1V0h4v1h1v9h4z"></path><path fill="black" d="M21 15v-1h-6v1h-1v1h-1v1h-1v1h-2v-1H9v-1H8v-1H7v-1H1v1H0v6h1v1h20v-1h1v-6h-1Zm-2 4h-1v-1h1v1Zm-3-1h1v1h-1v-1Z"></path></svg></a></button><div data-rmiz-wrap="visible"><div style="position: relative;"><span style="box-sizing: border-box; display: inline-block; overflow: hidden; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px; position: relative; max-width: 100%;"><span style="box-sizing: border-box; display: block; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px; max-width: 100%;"><img alt="" aria-hidden="true" src="data:image/svg+xml,%3csvg%20xmlns=%27http://www.w3.org/2000/svg%27%20version=%271.1%27%20width=%27616%27%20height=%27448%27/%3e" style="display: block; max-width: 100%; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px;"></span><img alt="source: https://deepmind.google/discover/blog/a-generalist-agent/" src="data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7" decoding="async" data-nimg="intrinsic" class="image" style="position: absolute; inset: 0px; box-sizing: border-box; padding: 0px; border: none; margin: auto; display: block; width: 0px; height: 0px; min-width: 100%; max-width: 100%; min-height: 100%; max-height: 100%; object-fit: contain; background-size: contain; background-position: 0% 0%; filter: blur(20px); background-image: url(&quot;data:image/svg+xml;base64,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&quot;);"><noscript></noscript></span></div><button aria-label="Zoom image" data-rmiz-btn-open="true"></button></div><p class="image-caption">source: https://deepmind.google/discover/blog/a-generalist-agent/</p></div></span></p><p class="line-space"> <br> </p><p><br>One of the most hyped tech news stories recently was about the company<a href="https://www.linkedin.com/company/groq/?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">Groq</a> developing an <a href="https://en.wikipedia.org/wiki/Application-specific_integrated_circuit?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">ASIC (Application-Specific Integrated Circuit)</a> that can run LLMs much more efficiently and with less energy than traditional GPUs. This clearly shows that LLM architecture has become so fundamental that it's now worthwhile to create specialized hardware for it.</p><p class="line-space"> <br> </p><p>Also recently, a publication titled "<a href="https://arxiv.org/pdf/2402.17764.pdf?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">The Era of 1-bit LLMs: All Large Language Models are in 1.58 Bits</a>" made its appearance. Quantization of neural networks is a common method for reducing size and computational demand. The essence of the solution is to perform the training on large GPU clusters using floating-point numbers, and then convert the finished network's weights into a less precise format, which allows the processors of user devices to work more efficiently. For example, the training is carried out with 16 or 32-bit floating-point numbers, which are then converted into 8 or 4-bit fixed-point numbers for fast client-side operations. This way, the model can function well even on mobile or IoT devices. An extreme form of this quantization is when the weights are converted into 1-bit numbers. This can be a <a href="https://www.youtube.com/watch?v=d9AeVyRzU5Q&amp;ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">complete binary conversion</a>, or as the publication suggests, using the values {-1,0,1} (hence the 1.58 bits). One might think that such extreme quantization would render the network completely unusable, but in reality, the opposite is true; these 1-bit networks perform exceptionally well.</p><p class="line-space"> <br> </p><h3 id="h-why-is-this-such-a-big-deal"><strong>Why is this such a big deal?</strong></h3>
<p>If these three values are sufficient to represent the weights, then multiplication, which is currently the most frequently used operation in neural networks, is no longer necessary. This is why GPU clusters are used for neural networks, as GPUs can perform multiplications very efficiently. Without the need for multiplications, there's no need for GPUs, and the models can be run efficiently even on CPUs, or it's possible to build specialized hardware (ASIC) that can (even in an analog way) run these 1-bit networks.</p>
<p><br>Currently, quantization is a post-training operation. Thus, the use of 1-bit networks doesn't accelerate the training process. Nevertheless, they are still useful because training is a one-time operation, but the neural network is then run countless times. Consequently, running the networks represents a significantly greater energy consumption than training. Therefore, we might benefit from this technology even in the context of training.</p>
<p><br>Since gradient-based training does not work with 1-bit or binarized networks, non-gradient-based technologies become relevant (check<a href="https://facebookresearch.github.io/nevergrad/?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">nevergrad</a> and <a href="https://pygad.readthedocs.io/en/latest/?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">PyGAD</a>), like genetic algorithms or other gradient-free technologies. Although in most cases backpropagation is much more efficient than gradient-free solutions, 1-bit networks can be run much more efficiently than their floating-point counterparts. So, it might be that with backpropagation, we find the optimal network 10 times faster using floating-point numbers than with, say, genetic algorithms. But if the 1-bit network runs 20 times faster, then training will still be twice as fast using genetic algorithms. Investigating how effectively 1-bit networks can be trained with gradient-free methods could be a very interesting research topic.</p>
<p><br>Another reason why this topic is so fascinating is that these networks more closely resemble the neural networks found in the natural brain (biologically plausible). Therefore, I believe that by choosing a good gradient-free training algorithm and applying these 1-bit networks, we can build systems that are much more similar to the human brain. Moreover, this opens up the possibility for technological solutions beyond ASICs that were previously not feasible, such as analog, light-based, or even biologically based processors.</p>
<p class="line-space"> <br> </p><p>It's possible that this direction might turn out to be a dead-end in the long run, but for now, its revolutionary potential is apparent, making it a very promising research avenue for anyone involved in the field of artificial intelligence.</p><p class="line-space"> <br> </p></div><div style="visibility:hidden" class="sc-1e366be2-7 kZfKay"><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M16 9h1v2h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1H9v1H8v1H7v1H6v1H0v-6h1v-1h1v-1h1v-1h1v-1h1v-1h1v-1h1V9h1V8h1V7h1V6h1V5h2v1h1v1h1v1h1v1zM22 4v2h-1v1h-1v1h-1v1h-1V8h-1V7h-1V6h-1V5h-1V4h-1V3h1V2h1V1h1V0h2v1h1v1h1v1h1v1h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 3v2h-1v2h-1v1h-5V7h-2v1h-1v1H9v4h1v1h1v1h2v-1h5v1h1v2h1v2h-1v2h-1v1h-5v-1h-1v-2h-1v-2h-1v-1H9v-1H8v-1H7v1H2v-1H1v-2H0v-2h1V8h1V7h5v1h1V7h1V6h1V5h1V3h1V1h1V0h5v1h1v2h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M14 19v2h-1v1H1v-1H0V5h1V4h3v15h10z"></path><path style="width:18px;fill:white" d="M20 6v11h-1v1H6v-1H5V1h1V0h8v6h6z"></path><path style="width:18px;fill:white" d="M20 4v1h-5V0h1v1h1v1h1v1h1v1h1z"></path></svg></button><a style="margin-left:15px;color:white" href="#commentSection"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 19" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M21 6V4h-1V3h-1V2h-2V1h-3V0H8v1H5v1H3v1H2v1H1v2H0v6h1v2h1v2H1v1H0v2h5v-1h1v-1h2v1h6v-1h3v-1h2v-1h1v-1h1v-2h1V6h-1ZM6 13h1v-1h1v-2H6V5h4v8H9v1H6v-1Zm6 0h1v-1h1v-2h-2V5h4v8h-1v1h-3v-1Z"></path></svg></a><button style="margin-right:10px"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 20" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 2v12h-1v1h-2v1h-6v-1H4v1H3v4H1V3H0V1h1V0h2v1h1v2H3v1h1V3h7v1h6V3h2V2h1z"></path></svg></button><div class="sc-1e366be2-8 cbbaIe"></div></div></div>

The Revolutionary Potential of 1-Bit Language Models

The Revolutionary Potential of 1-Bit Language Models (LLMs)

Chrome's broken trust led to Neulock: a password manager built on user control, not encryption & blind faith. Learn the 3 security principles behind it.

Chrome Password Manager Betrayed My Trust 13 Years Ago. I Never Forgot.

Master Next.js 14 authentication with NextAuth.js, react-hook-form, & Zod. Learn to secure apps with email/password auth & more in this concise guide.

How to Implement Authentication in Next.js 14 With NextAuth.js, Shadcn/ui, react-hook-form, and Zod

Learn about the importance of automated tests for stable software development and the three main types of tests - unit, integration, and end-to-end

Stable Software: Learn About the Power of Automated Tests

<div><div><div class="notice notice-tip">
<p><em>Hey everyone! <strong><a href="https://hackernoon.com/meet-the-writer-hackernoon-contributor-nataraj-sindam-on-experimenting-with-ai?ref=hackernoon.com" rel="ugc">I’m Nataraj</a>,</strong> and just like you, I’ve been fascinated with the recent progress of artificial intelligence. Realizing that I needed to stay abreast with all the developments happening, I decided to embark on a personal journey of learning, thus <strong><a href="https://hackernoon.com/u/sindamnataraj?ref=hackernoon.com" rel="ugc">100 days of AI</a></strong> was born! With this series, I will be learning about LLMs and share ideas, experiments, opinions, trends &amp; learnings through my blog posts. You can follow along the journey on HackerNoon <strong><a href="https://hackernoon.com/u/sindamnataraj?ref=hackernoon.com" rel="ugc">here</a></strong> or my personal website <strong><a href="https://thestartupproject.io/100-days-of-ai/?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">here</a></strong>. In today’s article, we’ll be looking to build a Semantic Kernel with the help of GPT-4.</em></p>
</div>
<p class="line-space"> <br> </p><p>If you are reading the 100 Days of AI series I am writing a common theme that you will notice is most of the posts have some kind of an experiment or a demo. At the end of posts I also list ideas that are possible with the technologies I explored. So the obvious next question is, is there a way to easily wrap these experiments into demo-able products and share them with people and get feedback?</p><p class="line-space"> <br> </p><p>In this post we will explore how to easily convert an AI experiment into a demo that can be and get feedback.</p><p class="line-space"> <br> </p><p>Let’s say you want to create an AI app that takes in text and returns a summary. And you want to share it with people and get feedback. If you have to create a web app for this app, you have to code the UI make sure you find a hosting solution and figure out what the entire stack is going to be. Of course there are started webapp stacks that will allow to do it more faster but you don’t need all that complexity for your use case. Instead we will use Gradio which create a web app like interface with out writing any code for this app.</p><h3 id="h-what-is-gradio">What is Gradio?</h3>
<p>Gradio is a python library that allows you to easily create demos of your AI apps and share it with an audience in the form of a web app.</p>
<p>To understand how to easily demo a gen AI experiment with Gradio, we will build a summarization web app where you can give the app a long text and the app will output the summary of that long text.</p>
<h3 id="h-step-1-setup"><strong>Step 1: Setup</strong></h3>
<p>Along with Gradio, we will use hugging face app so to follow along make sure you keep your hugging face api key handy. To get started lets load up the API key from .env file and add required python modules.</p>
<pre><code>import os
import io
from IPython.display import Image, display, HTML
#from PIL import Image
import base64 
import openai

env_path = '../.env'
from dotenv import load_dotenv, find_dotenv
_ = load_dotenv(find_dotenv(env_path)) # read local .env file

hf_api_key = os.environ['HF_API_KEY']
</code></pre>
<h3 id="h-step-2-define-summarization-function">Step 2: Define Summarization Function</h3>
<p>The core logic of what we want to achieve is to send a long text to a model and get a summary of the text. We will define a function that takes in the long text as an input and calls a model on hugging face called <a href="http://hf_api_summary_base%3Dhttps//api-inference.huggingface.co/models/sshleifer/distilbart-cnn-12-6?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">distilbart</a>. Here is the function to do this.</p>
<pre><code>def get_completion(inputs, parameters=None, ENDPOINT_URL=os.environ['HF_API_SUMMARY_BASE']): 
    headers = {
      "Authorization": f"Bearer {hf_api_key}",
      "Content-Type": "application/json"
    }
    data = { "inputs": inputs }
    if parameters is not None:
        data.update({"parameters": parameters})
    response = requests.request("POST",
                                ENDPOINT_URL, headers=headers,
                                data=json.dumps(data)
                               )
    return json.loads(response.content.decode("utf-8"))
</code></pre>
<h3 id="h-step-3-use-gradio-to-demo-summarization-app">Step 3: Use Gradio to Demo Summarization App</h3>
<p>The main way to interact with Gradio is through <strong>interface</strong> function. We need to pass a function and its corresponding inputs and outputs. Note that we defined a new function called <strong>summarize</strong> for this purpose. This function takes an input, which will be the long text which is to be summarized and the function will call the <strong>get_completion</strong> function from step 2 to get the summary of the input text and return that as the <strong>output</strong>. In the final line we are asking gradio to launch the demo of this app. By providing <strong>share=True</strong> we are telling gradio to create a public link that can be shared to others and also giving the server_port on which the local webhost should run on. In the below code we also customize the title, description and labels of input and output fields which you will notice in the out put.</p>
<pre><code>import gradio as gr
def summarize(input):
    output = get_completion(input)
    return output[0]['summary_text']
    
gr.close_all()
demo = gr.Interface(fn=summarize, 
                    inputs=[gr.Textbox(label="Text to summarize", lines=6)],
                    outputs=[gr.Textbox(label="Result", lines=3)],
                    title="Text summarization with distilbart-cnn",
                    description="Summarize any text using the `shleifer/distilbart-cnn-12-6` model under the hood!"
                   )
demo.launch(share=True, server_port=int(os.environ['PORT1']))
</code></pre>
<p>Once you run this, here is what the output now looks like in your browser at http://127.0.0.1:&lt;your port number&gt;.</p>
<p><span><div class="sc-a6c0bd8c-1 ijZWBT image-container undefined"><button class="sc-a6c0bd8c-0 khzBNC download-button"><a download="gradio-output.png" href="https://hackernoon.imgix.net/images/Y14t1a3ZEoT7RE6300ayez7dRN83-en834uh.png" title="Download image"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 22" width="22" fill="black"><path fill="black" d="M18 10v1h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1h-2v-1H9v-1H8v-1H7v-1H6v-1H5v-1H4v-1h4V1h1V0h4v1h1v9h4z"></path><path fill="black" d="M21 15v-1h-6v1h-1v1h-1v1h-1v1h-2v-1H9v-1H8v-1H7v-1H1v1H0v6h1v1h20v-1h1v-6h-1Zm-2 4h-1v-1h1v1Zm-3-1h1v1h-1v-1Z"></path></svg></a></button><div data-rmiz-wrap="visible"><div style="position: relative;"><span style="box-sizing: border-box; display: inline-block; overflow: hidden; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px; position: relative; max-width: 100%;"><span style="box-sizing: border-box; display: block; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px; max-width: 100%;"><img alt="" aria-hidden="true" src="data:image/svg+xml,%3csvg%20xmlns=%27http://www.w3.org/2000/svg%27%20version=%271.1%27%20width=%27800%27%20height=%27400.0086655112651%27/%3e" style="display: block; max-width: 100%; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px;"></span><img alt="Gradio Output" src="data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7" decoding="async" data-nimg="intrinsic" class="image" style="position: absolute; inset: 0px; box-sizing: border-box; padding: 0px; border: none; margin: auto; display: block; width: 0px; height: 0px; min-width: 100%; max-width: 100%; min-height: 100%; max-height: 100%; object-fit: contain; background-size: contain; background-position: 0% 0%; filter: blur(20px); background-image: url(&quot;data:image/svg+xml;base64,CiAgICA8c3ZnIHdpZHRoPSI4MDAiIGhlaWdodD0iNDAwLjAwODY2NTUxMTI2NTEiIHZlcnNpb249IjEuMSIgeG1sbnM9Imh0dHA6Ly93d3cudzMub3JnLzIwMDAvc3ZnIiB4bWxuczp4bGluaz0iaHR0cDovL3d3dy53My5vcmcvMTk5OS94bGluayI+CiAgICAgIDxkZWZzPgogICAgICAgIDxsaW5lYXJHcmFkaWVudCBpZD0iZyI+CiAgICAgICAgICA8c3RvcCBzdG9wLWNvbG9yPSIjMjIyIiBvZmZzZXQ9IjIlIiAvPgogICAgICAgICAgPHN0b3Agc3RvcC1jb2xvcj0iIzBmMCIgb2Zmc2V0PSIxMCUiIC8+CiAgICAgICAgICA8c3RvcCBzdG9wLWNvbG9yPSIjMjIyIiBvZmZzZXQ9IjEyJSIgLz4KICAgICAgICA8L2xpbmVhckdyYWRpZW50PgogICAgICA8L2RlZnM+CiAgICAgIDxyZWN0IHdpZHRoPSI4MDAiIGhlaWdodD0iNDAwLjAwODY2NTUxMTI2NTEiIGZpbGw9IiMyMjIiIC8+CiAgICAgIDxyZWN0IGlkPSJyIiB3aWR0aD0iODAwIiBoZWlnaHQ9IjQwMC4wMDg2NjU1MTEyNjUxIiBmaWxsPSJ1cmwoI2cpIiAvPgogICAgICA8YW5pbWF0ZSB4bGluazpocmVmPSIjciIgYXR0cmlidXRlTmFtZT0ieCIgZnJvbT0iLTgwMCIgdG89IjgwMCIgZHVyPSIxcyIgcmVwZWF0Q291bnQ9ImluZGVmaW5pdGUiICAvPgogICAgPC9zdmc+&quot;);"><noscript></noscript></span></div><button aria-label="Zoom image" data-rmiz-btn-open="true"></button></div><p class="image-caption">Gradio Output</p></div></span></p>
<p class="line-space"> <br> </p><p>It is that easy to demo a project with Gradio. The beauty here is you as a developer didn’t have to come up with UI stack and go through with the complexity of learning that stack and taking it to production. Gradio does it for you.</p><p class="line-space"> <br> </p><p>Gradio offers different input and output types using which you can create more complex and interesting demos that can be easily shared. We will explore more demos with Gradio and other tools in the coming posts.</p><p class="line-space"> <br> </p><p><em><strong>That’s it for Day 12 of 100 Days of AI.</strong></em></p><p class="line-space"> <br> </p><p><strong>I write a newsletter called Above Average where I talk about the second order insights behind everything that is happening in big tech. If you are in tech and don’t want to be average, <a href="https://above-average.beehiiv.com/?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">subscribe to it</a></strong>.</p><p class="line-space"> <br> </p><p>Follow me on&nbsp;<strong><a href="https://twitter.com/natarajsindam?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">Twitter</a></strong>,&nbsp;<strong><a href="https://www.linkedin.com/in/natarajsindam/?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">LinkedIn</a></strong>&nbsp;or <strong><a href="https://hackernoon.com/u/sindamnataraj?ref=hackernoon.com" rel="ugc">HackerNoon</a></strong> for latest updates on 100 days of AI. If you are in tech you might be interested in joining my community of tech professionals&nbsp;<strong><a href="https://thestartupproject.io/angel-investing/?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">here</a></strong>.</p></div><div style="visibility:hidden" class="sc-1e366be2-7 kZfKay"><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M16 9h1v2h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1H9v1H8v1H7v1H6v1H0v-6h1v-1h1v-1h1v-1h1v-1h1v-1h1v-1h1V9h1V8h1V7h1V6h1V5h2v1h1v1h1v1h1v1zM22 4v2h-1v1h-1v1h-1v1h-1V8h-1V7h-1V6h-1V5h-1V4h-1V3h1V2h1V1h1V0h2v1h1v1h1v1h1v1h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 3v2h-1v2h-1v1h-5V7h-2v1h-1v1H9v4h1v1h1v1h2v-1h5v1h1v2h1v2h-1v2h-1v1h-5v-1h-1v-2h-1v-2h-1v-1H9v-1H8v-1H7v1H2v-1H1v-2H0v-2h1V8h1V7h5v1h1V7h1V6h1V5h1V3h1V1h1V0h5v1h1v2h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M14 19v2h-1v1H1v-1H0V5h1V4h3v15h10z"></path><path style="width:18px;fill:white" d="M20 6v11h-1v1H6v-1H5V1h1V0h8v6h6z"></path><path style="width:18px;fill:white" d="M20 4v1h-5V0h1v1h1v1h1v1h1v1h1z"></path></svg></button><a style="margin-left:15px;color:white" href="#commentSection"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 19" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M21 6V4h-1V3h-1V2h-2V1h-3V0H8v1H5v1H3v1H2v1H1v2H0v6h1v2h1v2H1v1H0v2h5v-1h1v-1h2v1h6v-1h3v-1h2v-1h1v-1h1v-2h1V6h-1ZM6 13h1v-1h1v-2H6V5h4v8H9v1H6v-1Zm6 0h1v-1h1v-2h-2V5h4v8h-1v1h-3v-1Z"></path></svg></a><button style="margin-right:10px"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 20" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 2v12h-1v1h-2v1h-6v-1H4v1H3v4H1V3H0V1h1V0h2v1h1v2H3v1h1V3h7v1h6V3h2V2h1z"></path></svg></button><div class="sc-1e366be2-8 cbbaIe"></div></div></div>

In this post we will explore how to easily convert an AI experiment into a demo that can be and get feedback.

100 Days of AI, Day 12: How to Quickly Demo Your Gen AI Experiments to Get Feedback

What I learned in 6 months of working on the most powerful CodeGen dev tool GPT Pilot

After 6 Months of Working on a CodeGen Dev Tool (GPT Pilot), This Is What I Learned

Just because they can't feel, it doesn't mean they can't help. Empathetic AI has been show to have positive benefits. How can we test for artificial empathy?

Testing the Depths of AI Empathy: Frameworks and Challenges

Explore the unique aspects of VAEs, which separate them from traditional autoencoders by enabling data generation through sampling from a distribution

How to Sample From Latent Space With Variational Autoencoder

Your favorite programming language may be approaching leap years incorrectly.

Rethinking Leap Years: Why Your Favorite Programming Language's Approach May Be Flawed

<div><div><p class="line-space"> <br> </p><p>Welcome to the winners' announcement of the Final Round of the DevOps Writing Contest, hosted by Aptible and HackerNoon.</p><p class="line-space"> <br> </p><p>Many thanks to all the contributors and readers who have participated in this contest over the last 6 months. It has been such an insightful journey.</p><p class="line-space"> <br> </p><div class="notice notice-info">
<p><strong>Here are key stats that illustrate the value we've created together throughout the contest:</strong></p>
<p class="line-space"> <br> </p><ul>
<li>Stories published: <strong>261</strong></li>
<li>Total # Reads: <strong>759,334</strong></li>
<li>Total reading time: <strong>19 Days, 9 Hours, 27 Minutes</strong></li>
</ul>
</div>
<p class="line-space"> <br> </p><p>Your active participation has made this contest a success, with ripple effects leading to more opportunities for the HackerNoon community to win great prizes.</p><p class="line-space"> <br> </p><p>For those who couldn’t participate: <strong>Don’t Worry!</strong></p><p class="line-space"> <br> </p><p>We have an ongoing contest and tons more where that came from.</p><div class="notice notice-tip">
<p>Visit <a href="https://www.contests.hackernoon.com/?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">https://www.contests.hackernoon.com/</a> to learn more.</p>
</div>
<p class="line-space"> <br> </p><p><strong>Without further ado, let’s jump right into the announcements!</strong> 🥁🥁🥁</p><p class="line-space"> <br> </p><h2 id="h-the-dev-ops-writing-contest-round-6-finalists-🏆">The DevOps Writing Contest: Round 6 Finalists 🏆</h2>
<p class="line-space"> <br> </p><ul>
<li>
<p><a href="https://hackernoon.com/bye-bye-dora-flaws-of-the-state-of-devops-reports?ref=hackernoon.com" rel="ugc">Bye Bye DORA: Flaws of the State of DevOps Reports</a> by <a href="https://hackernoon.com/u/icyapril?ref=hackernoon.com" rel="ugc">@icyapril</a>.</p>
<p class="line-space"> <br> </p></li>
<li>
<p><a href="https://hackernoon.com/how-to-make-aws-opensearch-publicly-accessible-without-nginx-proxy?ref=hackernoon.com" rel="ugc">How to Make AWS OpenSearch Publicly Accessible Without nginx Proxy</a> by <a href="https://hackernoon.com/u/kvendingoldo?ref=hackernoon.com" rel="ugc">@kvendingoldo</a>.</p>
<p class="line-space"> <br> </p></li>
<li>
<p><a href="https://hackernoon.com/ansible-101-working-with-facts-and-templates?ref=hackernoon.com" rel="ugc">Ansible 101: Working With Facts and Templates</a> by <a href="https://hackernoon.com/u/cloudkungfu?ref=hackernoon.com" rel="ugc">@cloudkungfu</a>.</p>
<p class="line-space"> <br> </p></li>
<li>
<p><a href="https://hackernoon.com/tfblueprintgen-a-tool-to-simplify-terraform-folder-setup-and-provide-base-resource-modules?ref=hackernoon.com" rel="ugc">Tfblueprintgen: A Tool to Simplify Terraform Folder Setup and Provide Base Resource </a></p>
<p><a href="https://hackernoon.com/tfblueprintgen-a-tool-to-simplify-terraform-folder-setup-and-provide-base-resource-modules?ref=hackernoon.com" rel="ugc">Modules</a> by <a href="https://hackernoon.com/u/krishnaduttpanchagnula?ref=hackernoon.com" rel="ugc">@krishnaduttpanchagnula</a>.</p>
<p class="line-space"> <br> </p></li>
<li>
<p><a href="https://hackernoon.com/how-to-manage-multiple-opentofu-versions-with-tofuenv?ref=hackernoon.com" rel="ugc">How to Manage Multiple OpenTofu Versions With Tofuenv</a> by <a href="https://hackernoon.com/u/nmishin?ref=hackernoon.com" rel="ugc">@nmishin</a>.</p>
<p class="line-space"> <br> </p></li>
<li>
<p><a href="https://hackernoon.com/reverse-proxying-the-backbone-of-microservices-architecture?ref=hackernoon.com" rel="ugc">Reverse Proxying — the Backbone of Microservices Architecture</a> by <a href="https://hackernoon.com/u/infinity?ref=hackernoon.com" rel="ugc">@infinity</a>.</p>
<p class="line-space"> <br> </p></li>
<li>
<p><a href="https://hackernoon.com/6-cool-kubernetes-operators-and-how-to-use-them?ref=hackernoon.com" rel="ugc">6 Cool Kubernetes Operators and How to Use Them</a> by <a href="https://hackernoon.com/u/gilad-david-maayan?ref=hackernoon.com" rel="ugc">@gilad-david-maayan</a>.</p>
</li>
</ul>
<p class="line-space"> <br> </p><h2 id="h-and-the-winners-🏆-are">…and the winners🏆 are:</h2>
<p class="line-space"> <br> </p><h3 id="h-in-1st-first-place-we-have">In 1st first place, we have:</h3>
<p><div class="sc-56d3a1c3-0 evTXYi"><p class="line-space"><br></p><div class="embed-hn-story"><div class="card-image-container"><img src="https://hackernoon.imgix.net/images/wo54vA4dRXXTVVJQcL3vwFOoOn22-8ed3xrt.jpeg?mark-pad=0&amp;mark=http://hackernoon.imgix.net/HackerNoon%20Rounded%20Horizontal.png?w=400" alt="article preview"></div><div class="card-info"><span class="provider">HACKERNOON</span><h3 class="title">Bye Bye DORA: Flaws of the State of DevOps Reports | HackerNoon</h3><span class="desc">DORA Four Key Metrics have become the "gold standard" of measurement in the software engineering industry, however several flaws have recently come to light.</span></div></div></div></p>
<p class="line-space"> <br> </p><blockquote>
<p>Like Dr Keys received funding from the sugar industry in his research - in many investigations, it’s important to follow the money to understand where incentives lie. The DORA team originally started doing State of DevOps reports for Puppet, a company focussed on automating IT infrastructure and now they do this work for Google Cloud. Both have a vested interest in developers being able to deploy work as quickly as possible. This does not mean however that it is the solution to all our problems.</p>
</blockquote>
<p class="line-space"> <br> </p><div class="notice notice-info">
<p>Congratulations <a href="https://hackernoon.com/u/icyapril?ref=hackernoon.com" rel="ugc">@icyapril</a>, you’re $1500 richer!</p>
</div>
<p class="line-space"> <br> </p><h3 id="h-2nd-place-goes-to">2nd place goes to:</h3>
<p><div class="sc-56d3a1c3-0 evTXYi"><p class="line-space"><br></p><div class="embed-hn-story"><div class="card-image-container"><img src="https://hackernoon.imgix.net/images/2jqChkrv03exBUgkLrDzIbfM99q2-6882lxy.jpeg?mark-pad=0&amp;mark=http://hackernoon.imgix.net/HackerNoon%20Rounded%20Horizontal.png?w=400" alt="article preview"></div><div class="card-info"><span class="provider">HACKERNOON</span><h3 class="title">How to Manage Multiple OpenTofu Versions With Tofuenv | HackerNoon</h3><span class="desc">Discover tofuenv for OpenTofu: install and use it for adding, changing, and uninstalling OpenTofu versions with practical examples.</span></div></div></div></p>
<p class="line-space"> <br> </p><blockquote>
<p>Discover tofuenv for OpenTofu: install and use it for adding, changing, and uninstalling OpenTofu versions with practical examples.</p>
</blockquote>
<p class="line-space"> <br> </p><div class="notice notice-info">
<p>Congratulations <a href="https://hackernoon.com/u/nmishin?ref=hackernoon.com" rel="ugc">@nmishin</a>, you’ve won $1000 for your contribution!</p>
</div>
<p class="line-space"> <br> </p><h3 id="h-and-in-3rd-place-we-have">And in 3rd place, we have:</h3>
<p><div class="sc-56d3a1c3-0 evTXYi"><p class="line-space"><br></p><div class="embed-hn-story"><div class="card-image-container"><img src="https://hackernoon.imgix.net/images/eQHzh6rz7ETBHLjs0KzCl1Dooqp2-ld834a1.jpeg?mark-pad=0&amp;mark=http://hackernoon.imgix.net/HackerNoon%20Rounded%20Horizontal.png?w=400" alt="article preview"></div><div class="card-info"><span class="provider">HACKERNOON</span><h3 class="title">6 Cool Kubernetes Operators and How to Use Them | HackerNoon</h3><span class="desc">Kubernetes operators are software extensions that enable the automation of operational tasks. Operators allow developers to package and deploy kubernetes apps</span></div></div></div></p>
<p class="line-space"> <br> </p><blockquote>
<p>Kubernetes operators are a powerful tool for simplifying and automating the management of your Kubernetes applications. Whether you're dealing with monitoring, storage, secrets management, or distributed tracing, there's likely a Kubernetes operator that can make your life easier.</p>
</blockquote>
<p class="line-space"> <br> </p><div class="notice notice-info">
<p>Congratulations <a href="https://hackernoon.com/u/gilad-david-maayan?ref=hackernoon.com" rel="ugc">@gilad-david-maayan</a>, you have won $500!</p>
</div>
<p class="line-space"> <br> </p><p><strong>Congratulations once again to all our winners. Thank you for your hard work!</strong></p><p class="line-space"> <br> </p><h2 id="h-how-to-claim-your-hacker-noon-writing-contest-prize">How to Claim Your HackerNoon Writing Contest Prize<br></h2>
<ul>
<li>
<p>Contact <strong>yes-reply@hackernoon.com</strong> and <strong>sidra@hackernoon.com</strong> using the <strong>email ID attached to the winner's hackernoon account.</strong></p>
<p class="line-space"> <br> </p></li>
<li>
<p>We will validate your claim and share a form requesting your bank details for reward distribution.</p>
<p class="line-space"> <br> </p></li>
<li>
<p>You will receive your winnings in 2-4 weeks after completing the form.</p>
</li>
</ul>
<p class="line-space"> <br> </p><div class="notice notice-warning">
<p>Please note that you must contact us within 60 days after the winners’ announcement date.</p>
</div>
<hr>
<p class="line-space"> <br> </p><h2 id="h-about-aptible">About Aptible</h2>
<p>Aptible’s hosting platform automates provisioning, managing, and scaling infrastructure so developers can focus on what matters: their product.</p>
<div class="notice notice-tip">
<p>Get started for free with&nbsp;<a href="https://bit.ly/43A9mGn?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">Aptible</a>.</p>
</div>
<hr>
<p class="line-space"> <br> </p><div class="notice notice-info">
<p>The HackerNoon writing contests primarily aim to celebrate quality content and recruit educational stories for our community. We congratulate all the finalists. However, the Editorial team can ban a writer and/or disqualify a story if we find any misconduct like plagiarism, copyright infringement, or disinformation.</p>
</div>
<p class="line-space"> <br> </p><blockquote>
<p><em>Visit&nbsp;<a href="https://contests.hackernoon.com/?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">contests.hackernoon.com</a>&nbsp;to learn more about running and upcoming contests.</em></p>
</blockquote>
<p class="line-space"> <br> </p></div><div style="visibility:hidden" class="sc-1e366be2-7 kZfKay"><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M16 9h1v2h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1H9v1H8v1H7v1H6v1H0v-6h1v-1h1v-1h1v-1h1v-1h1v-1h1v-1h1V9h1V8h1V7h1V6h1V5h2v1h1v1h1v1h1v1zM22 4v2h-1v1h-1v1h-1v1h-1V8h-1V7h-1V6h-1V5h-1V4h-1V3h1V2h1V1h1V0h2v1h1v1h1v1h1v1h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 3v2h-1v2h-1v1h-5V7h-2v1h-1v1H9v4h1v1h1v1h2v-1h5v1h1v2h1v2h-1v2h-1v1h-5v-1h-1v-2h-1v-2h-1v-1H9v-1H8v-1H7v1H2v-1H1v-2H0v-2h1V8h1V7h5v1h1V7h1V6h1V5h1V3h1V1h1V0h5v1h1v2h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M14 19v2h-1v1H1v-1H0V5h1V4h3v15h10z"></path><path style="width:18px;fill:white" d="M20 6v11h-1v1H6v-1H5V1h1V0h8v6h6z"></path><path style="width:18px;fill:white" d="M20 4v1h-5V0h1v1h1v1h1v1h1v1h1z"></path></svg></button><a style="margin-left:15px;color:white" href="#commentSection"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 19" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M21 6V4h-1V3h-1V2h-2V1h-3V0H8v1H5v1H3v1H2v1H1v2H0v6h1v2h1v2H1v1H0v2h5v-1h1v-1h2v1h6v-1h3v-1h2v-1h1v-1h1v-2h1V6h-1ZM6 13h1v-1h1v-2H6V5h4v8H9v1H6v-1Zm6 0h1v-1h1v-2h-2V5h4v8h-1v1h-3v-1Z"></path></svg></a><button style="margin-right:10px"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 20" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 2v12h-1v1h-2v1h-6v-1H4v1H3v4H1V3H0V1h1V0h2v1h1v2H3v1h1V3h7v1h6V3h2V2h1z"></path></svg></button><div class="sc-1e366be2-8 cbbaIe"></div></div></div>

Aptible and HackerNoon bring you DevOps Writing Contest Round 6 results.

The DevOps Writing Contest: Round 6 Results Announced!

<div><div><p>If you’re using dynamically created components in Angular, you’ve probably found it difficult to pass information between the parent and child components when needed. You need to provide the info in the parent, and then inject it within the child component. While it’s not necessarily difficult to do, it results in a lot of extra boilerplate code.</p>
<p class="line-space"> <br> </p><p>It would be so much better if we could just use the <code>@Input</code> decorator like we’re used to. Well, guess what? Angular supports doing this exact thing as of version sixteen. In this post, I’ll show you how. Alright, let’s get to it.</p><p><div class="sc-56d3a1c3-0 evTXYi"><p class="line-space"><br></p><div class="youtube-container"><iframe title="ytcontainer" width="700" height="394" src="https://www.youtube.com/embed/MT2JP8pn1qU" frameborder="0"></iframe></div></div></p>
<h2 id="h-passing-data-the-old-way-using-the-angular-injector">Passing Data the Old Way Using the Angular Injector</h2>
<p>First, let’s look at how we used to pass data to dynamically created components from the parent. Here, in this example, within the component template, we have our <code>*ngComponentOutlet</code>, and we’re passing it to our player component.</p>
<p class="line-space"> <br> </p><p>In this case, we’re also passing a custom child injector, and this is how we’d inject the data for the child component before.</p><h3 id="h-main-ts">main.ts</h3>
<pre class="language-typescript" tabindex="0"><code class="language-typescript"><span class="token decorator"><span class="token at operator">@</span><span class="token function">Component</span></span><span class="token punctuation">(</span><span class="token punctuation">{</span>
    selector<span class="token operator">:</span> <span class="token string">'app-root'</span><span class="token punctuation">,</span>
    template<span class="token operator">:</span> <span class="token template-string"><span class="token template-punctuation string">`</span><span class="token string">
        &lt;ng-container *ngComponentOutlet="playerComponent; injector: childInjector"&gt;&lt;/ng-container&gt;
    </span><span class="token template-punctuation string">`</span></span>
<span class="token punctuation">}</span><span class="token punctuation">)</span>
<span class="token keyword">export</span> <span class="token keyword">class</span> <span class="token class-name">App</span> <span class="token punctuation">{</span>
    <span class="token keyword">protected</span> playerComponent <span class="token operator">=</span> PlayerComponent<span class="token punctuation">;</span>
<span class="token punctuation">}</span>
</code></pre>
<p class="line-space"> <br> </p><p>And to create this child injector, we needed to create a property so that it could be set and then used within the template.</p><p class="line-space"> <br> </p><pre class="language-typescript" tabindex="0"><code class="language-typescript"><span class="token keyword">export</span> <span class="token keyword">class</span> <span class="token class-name">App</span> <span class="token punctuation">{</span>
    <span class="token keyword">protected</span> childInjector<span class="token operator">?</span><span class="token operator">:</span> Injector<span class="token punctuation">;</span>
    <span class="token operator">...</span>
<span class="token punctuation">}</span>
</code></pre>
<p class="line-space"> <br> </p><p>Then, we needed to inject in the Injector from the angular core within the constructor.</p><p class="line-space"> <br> </p><pre class="language-typescript" tabindex="0"><code class="language-typescript"><span class="token keyword">export</span> <span class="token keyword">class</span> <span class="token class-name">App</span> <span class="token punctuation">{</span>
    <span class="token operator">...</span>
    <span class="token function">constructor</span><span class="token punctuation">(</span><span class="token keyword">private</span> injector<span class="token operator">:</span> Injector<span class="token punctuation">)</span> <span class="token punctuation">{</span>
    <span class="token punctuation">}</span>
<span class="token punctuation">}</span>
</code></pre>
<p class="line-space"> <br> </p><p>After that, we needed to set the child injector using the create method and the providers’ array to provide our player token to the child.</p><p class="line-space"> <br> </p><pre class="language-typescript" tabindex="0"><code class="language-typescript"><span class="token keyword">export</span> <span class="token keyword">class</span> <span class="token class-name">App</span> <span class="token punctuation">{</span>
    <span class="token operator">...</span>
    <span class="token function">constructor</span><span class="token punctuation">(</span><span class="token keyword">private</span> injector<span class="token operator">:</span> Injector<span class="token punctuation">)</span> <span class="token punctuation">{</span>
        <span class="token keyword">this</span><span class="token punctuation">.</span>childInjector <span class="token operator">=</span> Injector<span class="token punctuation">.</span><span class="token function">create</span><span class="token punctuation">(</span><span class="token punctuation">{</span>
            providers<span class="token operator">:</span> <span class="token punctuation">[</span>
                <span class="token punctuation">{</span> provide<span class="token operator">:</span> PlayerToken<span class="token punctuation">,</span> useValue<span class="token operator">:</span> <span class="token keyword">this</span><span class="token punctuation">.</span>player <span class="token punctuation">}</span>
            <span class="token punctuation">]</span><span class="token punctuation">,</span>
            parent<span class="token operator">:</span> <span class="token keyword">this</span><span class="token punctuation">.</span>injector
        <span class="token punctuation">}</span><span class="token punctuation">)</span><span class="token punctuation">;</span>
    <span class="token punctuation">}</span>
<span class="token punctuation">}</span>
</code></pre>
<p class="line-space"> <br> </p><p>And finally, over in the child component, our player component, we needed to set our player property with the inject method from the angular core.</p><h3 id="h-player-component-ts">player.component.ts</h3>
<pre class="language-typescript" tabindex="0"><code class="language-typescript"><span class="token keyword">export</span> <span class="token keyword">class</span> <span class="token class-name">PlayerComponent</span> <span class="token punctuation">{</span>
    <span class="token keyword">protected</span> player<span class="token operator">?</span><span class="token operator">:</span> Player <span class="token operator">=</span> <span class="token function">inject</span><span class="token punctuation">(</span>PlayerToken<span class="token punctuation">)</span><span class="token punctuation">;</span>
<span class="token punctuation">}</span>
</code></pre>
<p class="line-space"> <br> </p><p>So, all of this is just to pass a simple player object to a dynamic component. I mean, if this wasn’t a dynamic component, we’d simply make this an <code>@Input</code> and just bind the player data object to the <code>@Input</code> in the parent. But this is a dynamically created component, so we can’t do this, right?</p><p class="line-space"> <br> </p><p>Well, as of Angular sixteen, we actually can use <code>@Input</code>s instead, and it’s much simpler than what we’ve seen so far.</p><h2 id="h-passing-data-the-new-way-using-the-ng-component-outlet-inputs-object">Passing Data the New Way Using the <code>*ngComponentOutlet</code> Inputs Object</h2>
<p>We start by changing this player property to an <code>@Input</code> instead. And, it’s simply typed into our player interface.</p>
<h3 id="h-player-component-ts">player.component.ts</h3>
<pre class="language-typescript" tabindex="0"><code class="language-typescript"><span class="token keyword">export</span> <span class="token keyword">class</span> <span class="token class-name">PlayerComponent</span> <span class="token punctuation">{</span>
    <span class="token decorator"><span class="token at operator">@</span><span class="token function">Input</span></span><span class="token punctuation">(</span><span class="token punctuation">)</span> player<span class="token operator">?</span><span class="token operator">:</span> Player<span class="token punctuation">;</span>
<span class="token punctuation">}</span>
</code></pre>
<p class="line-space"> <br> </p><p>Now, we can remove the inject method and player token imports since they’re no longer needed. Then, back over in our parent, in the template, we can remove the injector, and instead, replace it with an inputs object. We can pass this object on any number of inputs. So, if we had five inputs, we’d just include their names, and then pass each one whatever data we need.</p><p class="line-space"> <br> </p><p>But in this case, we have only one input on our child component, player, so that’s all we need to pass it.</p><h3 id="h-main-ts">main.ts</h3>
<pre class="language-typescript" tabindex="0"><code class="language-typescript"><span class="token decorator"><span class="token at operator">@</span><span class="token function">Component</span></span><span class="token punctuation">(</span><span class="token punctuation">{</span>
    selector<span class="token operator">:</span> <span class="token string">'app-root'</span><span class="token punctuation">,</span>
    template<span class="token operator">:</span> <span class="token template-string"><span class="token template-punctuation string">`</span><span class="token string">
        &lt;ng-container *ngComponentOutlet="playerComponent; inputs: { player }"&gt;&lt;/ng-container&gt;
    </span><span class="token template-punctuation string">`</span></span>
<span class="token punctuation">}</span><span class="token punctuation">)</span>
</code></pre>
<p class="line-space"> <br> </p><p>Now, we can remove the child injector. This means we can also remove the constructor altogether. And finally, we can remove the Injector from the angular core and the player token import since we’re no longer using them.</p><p class="line-space"> <br> </p><pre class="language-typescript" tabindex="0"><code class="language-typescript"><span class="token keyword">export</span> <span class="token keyword">class</span> <span class="token class-name">App</span> <span class="token punctuation">{</span>
    <span class="token keyword">protected</span> player<span class="token operator">:</span> Player <span class="token operator">=</span> <span class="token punctuation">{</span>
        name<span class="token operator">:</span> <span class="token string">'LeBron James'</span><span class="token punctuation">,</span>
        games<span class="token operator">:</span> <span class="token number">1421</span><span class="token punctuation">,</span>
        points<span class="token operator">:</span> <span class="token number">38652</span><span class="token punctuation">,</span>
        fieldGoalPercentage<span class="token operator">:</span> <span class="token number">0.505</span><span class="token punctuation">,</span>
        threePointPercentage<span class="token operator">:</span> <span class="token number">0.345</span><span class="token punctuation">,</span>
        imageName<span class="token operator">:</span> <span class="token string">'lebron-james'</span>
  <span class="token punctuation">}</span><span class="token punctuation">;</span>
  <span class="token keyword">protected</span> playerComponent <span class="token operator">=</span> PlayerComponent<span class="token punctuation">;</span>
<span class="token punctuation">}</span>
</code></pre>
<p class="line-space"> <br> </p><p>And that’s it. We can see everything works just as it used to. So, quite a bit easier to pull off and less code too which is always great.</p><p class="line-space"> <br> </p><p>One bummer though is that while <code>@Input</code>s are supported, <code>@Output</code>s aren’t. Kind of a bummer, but it is what it is. Hopefully, they’ll add support in the future because this would be very handy as well, but we’ll just have to wait and see.</p><h2 id="h-want-to-see-it-in-action">Want to See It in Action?</h2>
<p>Check out the demo code and examples of these techniques in the Stackblitz example below. If you have any questions or thoughts, don’t hesitate to leave a comment.</p>
<p><div class="sc-56d3a1c3-0 evTXYi"><p class="line-space"><br></p><div class="embed-hn-story"><div class="card-image-container"><img src="https://social-img.staticblitz.com/projects/stackblitz-starters-9xzkkf/a0a954def9f19ed9a9030bb12fb29915" alt="article preview"></div><div class="card-info"><span class="provider">STACKBLITZ</span><h3 class="title">Angular Dynamic Component Inputs - StackBlitz</h3><span class="desc">An angular-cli project based on @angular/animations, @angular/common, @angular/compiler, @angular/core, @angular/forms, @angular/platform-browser, @angular/platform-browser-dynamic, @angular/router, core-js, rxjs, tslib and zone.js</span></div></div></div></p>
<p class="line-space"> <br> </p></div><div style="visibility:hidden" class="sc-1e366be2-7 kZfKay"><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M16 9h1v2h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1H9v1H8v1H7v1H6v1H0v-6h1v-1h1v-1h1v-1h1v-1h1v-1h1v-1h1V9h1V8h1V7h1V6h1V5h2v1h1v1h1v1h1v1zM22 4v2h-1v1h-1v1h-1v1h-1V8h-1V7h-1V6h-1V5h-1V4h-1V3h1V2h1V1h1V0h2v1h1v1h1v1h1v1h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 3v2h-1v2h-1v1h-5V7h-2v1h-1v1H9v4h1v1h1v1h2v-1h5v1h1v2h1v2h-1v2h-1v1h-5v-1h-1v-2h-1v-2h-1v-1H9v-1H8v-1H7v1H2v-1H1v-2H0v-2h1V8h1V7h5v1h1V7h1V6h1V5h1V3h1V1h1V0h5v1h1v2h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M14 19v2h-1v1H1v-1H0V5h1V4h3v15h10z"></path><path style="width:18px;fill:white" d="M20 6v11h-1v1H6v-1H5V1h1V0h8v6h6z"></path><path style="width:18px;fill:white" d="M20 4v1h-5V0h1v1h1v1h1v1h1v1h1z"></path></svg></button><a style="margin-left:15px;color:white" href="#commentSection"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 19" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M21 6V4h-1V3h-1V2h-2V1h-3V0H8v1H5v1H3v1H2v1H1v2H0v6h1v2h1v2H1v1H0v2h5v-1h1v-1h2v1h6v-1h3v-1h2v-1h1v-1h1v-2h1V6h-1ZM6 13h1v-1h1v-2H6V5h4v8H9v1H6v-1Zm6 0h1v-1h1v-2h-2V5h4v8h-1v1h-3v-1Z"></path></svg></a><button style="margin-right:10px"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 20" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 2v12h-1v1h-2v1h-6v-1H4v1H3v4H1V3H0V1h1V0h2v1h1v2H3v1h1V3h7v1h6V3h2V2h1z"></path></svg></button><div class="sc-1e366be2-8 cbbaIe"></div></div></div>

It would be so much better if we could just use the @Input decorator like we’re used to. Well, guess what? Angular supports doing this exact thing as of v16

How to Use Angular @Input to Pass Data to Dynamically Created Components

<div><div><p class="line-space"> <br> </p><h2 id="h-iarp-as-video-linc-program-could-be-repurposed-to-spy-on-protesters-enforce-15-minute-smart-city-compliance-perspective">IARPA’s Video LINC program could be repurposed to spy on protesters, enforce 15-minute smart city compliance: perspective</h2>
<p class="line-space"> <br> </p><p>The&nbsp;<a href="https://sociable.co/?s=iarpa&amp;ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">Intelligence Advanced Research Projects Activity</a>&nbsp;(IARPA) is putting together a research program for the US spy community to autonomously identify, track, and trace people and their vehicles over long distances and periods of time.</p><p class="line-space"> <br> </p><p>Last week, the research funding arm of the US intelligence community, IARPA, published the&nbsp;<a href="https://sam.gov/opp/64446a6a809e49238d1346eb8b462bb9/view?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">technical draft</a>&nbsp;for its Video Linking and Intelligence from Non-Collaborative Sensors (<strong>Video LINCS</strong>) research program.</p><p class="line-space"> <br> </p><p>The newly-updated draft broad agency announcement and funding opportunity details how the US spy community is looking to autonomously identify, track, and trace people, vehicles, and objects by using <a href="https://hackernoon.com/top-10-ai-trends-of-2024-how-ai-transforms-everything?ref=hackernoon.com" rel="ugc">AI to analyze video</a> footage captured by CCTV cameras, drones, and potentially webcams and phones (as evidenced in the&nbsp;<strong>Video LINCS</strong>&nbsp;Program Design image below).</p><p class="line-space"> <br> </p><blockquote>
<p><em><strong>“The program will start with person reID, progress to vehicle reID, and culminate with reID of generic objects across a video collection”</strong></em></p>
<p>IARPA Video LINCS Program</p>
</blockquote>
<p class="line-space"> <br> </p><p><span><div class="sc-a6c0bd8c-1 ijZWBT image-container undefined"><button class="sc-a6c0bd8c-0 khzBNC download-button"><a download="image.833Z-exsfy4wzy9chzfhsgb3wpa8e" href="https://hackernoon.imgix.net/images/ZSnNpZ3GejgVSxqs3EyP38vlsLL2-2024-02-28T10:14:37.833Z-exsfy4wzy9chzfhsgb3wpa8e" title="Download image"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 22" width="22" fill="black"><path fill="black" d="M18 10v1h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1h-2v-1H9v-1H8v-1H7v-1H6v-1H5v-1H4v-1h4V1h1V0h4v1h1v9h4z"></path><path fill="black" d="M21 15v-1h-6v1h-1v1h-1v1h-1v1h-2v-1H9v-1H8v-1H7v-1H1v1H0v6h1v1h20v-1h1v-6h-1Zm-2 4h-1v-1h1v1Zm-3-1h1v1h-1v-1Z"></path></svg></a></button><div data-rmiz-wrap="visible"><div style="position: relative;"><span style="box-sizing: border-box; display: inline-block; overflow: hidden; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px; position: relative; max-width: 100%;"><span style="box-sizing: border-box; display: block; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px; max-width: 100%;"><img alt="" aria-hidden="true" src="data:image/svg+xml,%3csvg%20xmlns=%27http://www.w3.org/2000/svg%27%20version=%271.1%27%20width=%2740%27%20height=%2740%27/%3e" style="display: block; max-width: 100%; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px;"></span><img alt="image" src="data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7" decoding="async" data-nimg="intrinsic" class="image" style="position: absolute; inset: 0px; box-sizing: border-box; padding: 0px; border: none; margin: auto; display: block; width: 0px; height: 0px; min-width: 100%; max-width: 100%; min-height: 100%; max-height: 100%; object-fit: contain; background-size: contain; background-position: 0% 0%; filter: blur(20px); background-image: url(&quot;data:image/svg+xml;base64,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&quot;);"><noscript></noscript></span></div><button aria-label="Zoom image" data-rmiz-btn-open="true"></button></div></div></span></p><p>The official reasons given for developing this program have to do with responding to “<em><strong>tragic incidents</strong></em>” that require “<em><strong>forensic analyses</strong></em>,” and to “<em><strong>analyze patterns for anomalies and threats</strong></em>.”</p>
<p>IARPA program director Dr. Reuven Meth also mentioned in the video below that&nbsp;<strong>Video LINCS</strong>&nbsp;would be used to “<em><strong>facilitate smart city planning</strong></em>.”</p>
<p class="line-space"> <br> </p><p>But ask yourself, why would the US spy agency funding arm want to develop tools for smart city planning?</p><p class="line-space"> <br> </p><blockquote>
<p><em>Video LINCS&nbsp;<strong>will take labor-intensive workflows and automate them to facilitate forensic analyses, proactive threat detection, and smart city planning”</strong></em></p>
<p>IARPA Program Director Dr. Reuven Meth</p>
</blockquote>
<p><div class="sc-56d3a1c3-0 evTXYi"><p class="line-space"><br></p><div class="youtube-container"><iframe title="ytcontainer" width="700" height="394" src="https://www.youtube.com/embed/G2U-13OuisU" frameborder="0"></iframe></div></div></p>
<p class="line-space"> <br> </p><p><em>The Sociable</em>&nbsp;previously reported on the initial&nbsp;<strong>Video LINCS</strong>&nbsp;<a href="https://sociable.co/government-and-policy/us-intelligence-autonomously-reidentify-map-objects-videos-sensors/?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">program announcement</a>&nbsp;on January 9, 2024 when there was limited information publicly available, but last week IARPA updated its technical specs, giving us a more detailed look at just how deep the US spy apparatus is willing to go in its surveillance pursuits.</p><p class="line-space"> <br> </p><p>The&nbsp;<strong>Video LINCS</strong>&nbsp;program consists of two Technical Areas (TAs):</p><p class="line-space"> <br> </p><ul>
<li><strong>Re-identification (ReID)</strong>: Autonomously and automatically associating the same object (person, vehicle, or generic object) across a video corpus.</li>
<li><strong>Object Geo-localization</strong>: Geo-locating objects to provide positions for all objects in a common world reference frame.</li>
</ul>
<p class="line-space"> <br> </p><p>ReID, according to IARPA, means “the process of matching the same object across a video collection, to determine where the object appears throughout the video.”</p><p class="line-space"> <br> </p><blockquote>
<p><em><strong>“The goal of the</strong></em>&nbsp;<em>Video LINCS&nbsp;<strong>program, is to develop re-identification (reID) algorithms to autonomously associate objects across diverse, non-collaborative, video sensor footage and map re-identified objects to a unified coordinate system (geo-localization)”</strong></em></p>
<p>IARPA Video LINCS Program</p>
</blockquote>
<p class="line-space"> <br> </p><p>The&nbsp;<strong>Video LINCS</strong>&nbsp;program, if it does indeed become a fully-funded research program, will consist of three phases over the course of 48 months:</p><p class="line-space"> <br> </p><ul>
<li>During Phase 1, teams will demonstrate the feasibility of reID for people in a video corpus and object geo-localization to provide all subject motion in a common reference frame.&nbsp;<em><strong>Each person’s clothing will remain the same</strong></em>&nbsp;(short-term / temporally proximate reID) and provided metadata (e.g. time-stamps and camera pose, to the extent available) will be noise free.</li>
<li>During Phase 2,&nbsp;<em><strong>reID will expand to include people with clothing changes</strong></em>&nbsp;(long-term / temporally distant reID),&nbsp;<em><strong>include vehicles</strong></em>, require functionality on generic objects, additional sensor types and collection geometries will be included in the evaluations, and noise will be introduced into provided metadata.</li>
<li>During Phase 3, evaluation&nbsp;<em><strong>will have a stronger focus on reID of generic objects, temporally distant reID of vehicles will be performed</strong></em>, additional sensor types and collection geometries will be included, and there will be greater uncertainty in camera pose.</li>
</ul>
<p class="line-space"> <br> </p><p><span><div class="sc-a6c0bd8c-1 ijZWBT image-container undefined"><button class="sc-a6c0bd8c-0 khzBNC download-button"><a download="image.836Z-yhebph8822gd4wgwd0wxk3cz" href="https://hackernoon.imgix.net/images/ZSnNpZ3GejgVSxqs3EyP38vlsLL2-2024-02-28T10:14:37.836Z-yhebph8822gd4wgwd0wxk3cz" title="Download image"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 22" width="22" fill="black"><path fill="black" d="M18 10v1h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1h-2v-1H9v-1H8v-1H7v-1H6v-1H5v-1H4v-1h4V1h1V0h4v1h1v9h4z"></path><path fill="black" d="M21 15v-1h-6v1h-1v1h-1v1h-1v1h-2v-1H9v-1H8v-1H7v-1H1v1H0v6h1v1h20v-1h1v-6h-1Zm-2 4h-1v-1h1v1Zm-3-1h1v1h-1v-1Z"></path></svg></a></button><div data-rmiz-wrap="visible"><div style="position: relative;"><span style="box-sizing: border-box; display: inline-block; overflow: hidden; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px; position: relative; max-width: 100%;"><span style="box-sizing: border-box; display: block; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px; max-width: 100%;"><img alt="" aria-hidden="true" src="data:image/svg+xml,%3csvg%20xmlns=%27http://www.w3.org/2000/svg%27%20version=%271.1%27%20width=%2740%27%20height=%2740%27/%3e" style="display: block; max-width: 100%; width: initial; height: initial; background: none; opacity: 1; border: 0px; margin: 0px; padding: 0px;"></span><img alt="image" src="data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7" decoding="async" data-nimg="intrinsic" class="image" style="position: absolute; inset: 0px; box-sizing: border-box; padding: 0px; border: none; margin: auto; display: block; width: 0px; height: 0px; min-width: 100%; max-width: 100%; min-height: 100%; max-height: 100%; object-fit: contain; background-size: contain; background-position: 0% 0%; filter: blur(20px); background-image: url(&quot;data:image/svg+xml;base64,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&quot;);"><noscript></noscript></span></div><button aria-label="Zoom image" data-rmiz-btn-open="true"></button></div></div></span></p><p>The&nbsp;<strong>Video LINCS</strong>&nbsp;program will look to re-ID people, vehicles, and objects over long distances and periods of time using a kaleidoscope of technologies, including:</p>
<p class="line-space"> <br> </p><ul>
<li>Artificial Intelligence</li>
<li>Computer vision, including object detection, tracking, person/vehicle/object modeling, generic vision learning</li>
<li>Deep learning</li>
<li>Geometric camera projections and inverse projections</li>
<li>Image and video geo-localization</li>
<li>Machine learning</li>
<li>Modeling and simulation</li>
<li>Open set classification</li>
<li>Re-identification</li>
<li>Soft biometrics</li>
<li>Software engineering</li>
<li>Software integration</li>
<li>Systems integration</li>
<li>Truthed video data collection, annotation (truthing in both anonymized identities and geo-location), including potential Human Subject Research</li>
<li>Vehicle fingerprinting</li>
<li>Video data generation (including simulation, generative modeling)</li>
</ul>
<p class="line-space"> <br> </p><blockquote>
<p><em><strong>“The system needs to automatically locate the objects and associate them – across scale, aspect, density, crowding, obscuration, etc. – without introducing false detections and false matches”</strong></em></p>
<p>IARPA Video LINCS Program</p>
</blockquote>
<p class="line-space"> <br> </p><p>Apart from threat detection, forensic analyses, and smart city planning, there are many other potential use cases that could come from this spy program down the road.</p><p class="line-space"> <br> </p><p>For example, the tools and tactics coming from&nbsp;<strong>Video LINCS</strong>&nbsp;would be able to, hypothetically, identify who was present at a rally, protest, or riot — such as the one that occurred in Washington, DC on January 6, 2021 — and follow their every move as they make their way back home, even when they change their clothes.</p><p class="line-space"> <br> </p><p>Another example could be for keeping <a href="https://hackernoon.com/this-app-is-tracking-250000-immigrants-in-the-usa?ref=hackernoon.com" rel="ugc">track of immigrants</a> at land, air, and sea border crossings.</p><p>And it would be invaluable to governments wishing to enforce future lockdowns or low-emission zones in&nbsp;<a href="https://sociable.co/government-and-policy/digital-public-infrastructure-fusion-digital-id-cbdc-15-minute-smart-cities/?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">15-minute smart cities</a>&nbsp;as the authorities would be able to identify who broke protocol while tracking and tracing their every move for law enforcement to hunt them down.</p>
<p>All of this would be done autonomously and automatically.</p>
<p class="line-space"> <br> </p><p>But hey! maybe I’m being overdramatic.</p><p class="line-space"> <br> </p><p>After all, the government said it was to keep people safe, and the government has never let us down in always keeping our best interests at heart.</p><p class="line-space"> <br> </p><hr>
<div class="notice notice-info">
<p>This article was originally published by <strong>Tim Hinchliffe</strong> on <a href="https://sociable.co/government-and-policy/us-spy-community-autonomously-track-people-vehicles-ai-video/?ref=hackernoon.com#google_vignette" target="_blank" rel="noopener noreferrer ugc">The Sociable.</a></p>
</div>
<p class="line-space"> <br> </p></div><div style="visibility:hidden" class="sc-1e366be2-7 kZfKay"><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M16 9h1v2h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1H9v1H8v1H7v1H6v1H0v-6h1v-1h1v-1h1v-1h1v-1h1v-1h1v-1h1V9h1V8h1V7h1V6h1V5h2v1h1v1h1v1h1v1zM22 4v2h-1v1h-1v1h-1v1h-1V8h-1V7h-1V6h-1V5h-1V4h-1V3h1V2h1V1h1V0h2v1h1v1h1v1h1v1h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 3v2h-1v2h-1v1h-5V7h-2v1h-1v1H9v4h1v1h1v1h2v-1h5v1h1v2h1v2h-1v2h-1v1h-5v-1h-1v-2h-1v-2h-1v-1H9v-1H8v-1H7v1H2v-1H1v-2H0v-2h1V8h1V7h5v1h1V7h1V6h1V5h1V3h1V1h1V0h5v1h1v2h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M14 19v2h-1v1H1v-1H0V5h1V4h3v15h10z"></path><path style="width:18px;fill:white" d="M20 6v11h-1v1H6v-1H5V1h1V0h8v6h6z"></path><path style="width:18px;fill:white" d="M20 4v1h-5V0h1v1h1v1h1v1h1v1h1z"></path></svg></button><a style="margin-left:15px;color:white" href="#commentSection"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 19" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M21 6V4h-1V3h-1V2h-2V1h-3V0H8v1H5v1H3v1H2v1H1v2H0v6h1v2h1v2H1v1H0v2h5v-1h1v-1h2v1h6v-1h3v-1h2v-1h1v-1h1v-2h1V6h-1ZM6 13h1v-1h1v-2H6V5h4v8H9v1H6v-1Zm6 0h1v-1h1v-2h-2V5h4v8h-1v1h-3v-1Z"></path></svg></a><button style="margin-right:10px"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 20" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 2v12h-1v1h-2v1h-6v-1H4v1H3v4H1V3H0V1h1V0h2v1h1v2H3v1h1V3h7v1h6V3h2V2h1z"></path></svg></button><div class="sc-1e366be2-8 cbbaIe"></div></div></div>

IARPA’s Video LINC program could be repurposed to spy on protesters, enforce 15-minute smart city compliance: perspective

US Intelligence Eyes AI and Video Data for Tracking Individuals and Vehicles

Learn how to effortlessly center elements in your web design projects using different CSS approaches, catering to various scenarios and preferences.

3 Contemporary CSS Techniques for Centering Items

<div><div><p>Fanaticism has been, for much of American history, reserved for institutions like sports teams and universities where allegiances run so deep they can (and often have) become a central part of one’s identity. But an interesting shift occurred around the turn of the century where <em>individuals</em> have come to command the same fervent admiration.</p>
<p class="line-space"> <br> </p><p>This happened in politics - that’s a topic for a different publication - and also in business: most prominently with Steve Jobs and Elon Musk. <strong>If you connect the dots between what drove the love of those tech leaders and recent events, it is clear the new cult hero in tech is Sam Altman.</strong></p><p class="line-space"> <br> </p><p>Altman clinched this position in late November after he was abruptly terminated from his position as the chief executive of OpenAI, the company he helped found in 2015.</p><p class="line-space"> <br> </p><p>As has been covered extensively, the OpenAI board unexpectedly asked Altman to join a Google Meet link, at which point they told him he would be <a href="https://hackernoon.com/learning-from-the-openai-weekend-shake-up?ref=hackernoon.com" rel="ugc">fired from OpenAI over what has been described as a “breakdown in communication.”</a></p><p class="line-space"> <br> </p><p>Within two days, 90% of OpenAI employees had signed an open letter to the board threatening to leave OpenAI for Microsoft if they did not reinstate Altman as CEO. This was believed to be such a credible threat that Microsoft spent the weekend loading office equipment and computers into a floor at LinkedIn’s San Francisco office, preparing for their arrival.</p><p class="line-space"> <br> </p><p>Ultimately, Altman returned to his job as CEO, and <a href="https://hackernoon.com/on-openai-failed-board-coup-of-sam-altman-and-the-danger-of-leaving-ai-fate-in-the-hands-of-a-few?ref=hackernoon.com" rel="ugc">OpenAI shuffled their board of directors</a>.</p><p class="line-space"> <br> </p><p>And while it was a jarring sequence of events in short order, the most shocking part wasn’t his firing and re-hiring, the board re-alignment, or the fact that 90% of employees threatened to quit over the events: <strong>the biggest surprise was the broader tech industry’s reaction to Altman’s ouster.</strong></p><p class="line-space"> <br> </p><p>People who had no relationship with OpenAI or Altman came rushing to his full-throated defense. Opinion pieces were penned. Hundreds of hours of podcasts were recorded. Social media feeds were full of people who had his back despite the fact they had never so much as heard Altman’s voice.</p><p class="line-space"> <br> </p><p>All of this activity went well beyond a reaction to someone getting a raw deal, and it felt familiar: like the kind of debate and memes and discourse you see play out about Steve Jobs’ role at Apple or about any of the various things Elon Musk is involved in. The response crossed into fanaticism.</p><h2 id="h-the-key-ingredients-of-a-tech-cult-hero">The Key Ingredients of a Tech Cult Hero</h2>
<p>A (perhaps unexpected) benefit of Altman’s firing and re-hiring is that it gave him one of the three core ingredients for the making of a tech cult hero: <strong>resilience</strong>. People - Americans especially - love a comeback story, and Altman’s was unlike any we had seen before, especially with regard to how quickly it played out.</p>
<p class="line-space"> <br> </p><p>When Jobs was fired from Apple as a result of corporate politics, it took twelve years before he returned to the helm - true to the speed with which things play out in 2024, Altman’s journey took three days.</p><p class="line-space"> <br> </p><p>Despite the brevity of his absence, it was intense, and the stakes were high, much like the (many) times Elon took Tesla to the brink of extinction before prevailing en route to becoming the wealthiest person in the world.</p><p class="line-space"> <br> </p><p>All three of these individuals also offer <strong>compelling visions accompanied by convincing breakthroughs</strong>. For Jobs, it was (initially) making computers so accessible and easy to use, that you’d welcome them in your home. For Musk, it was the electric car.</p><p class="line-space"> <br> </p><p>And with Sam Altman, it’s clear his moonshot vision - an effort that he has proved capable of leading - is artificial intelligence.</p><p class="line-space"> <br> </p><p>Talk is cheap, and many have been drawn in by charismatic visionary types before, but in each of these cases, these people executed. That’s a big box to check off.</p><p class="line-space"> <br> </p><p>More than anything, the <strong>focus and tenacity</strong> Jobs and Musk demonstrated throughout their careers, across multiple efforts and companies, is the kind of conviction that people admire.</p><p class="line-space"> <br> </p><p>It’s easy to look at Musk and wonder why he maintains (a material level of) involvement in Tesla, SpaceX, Boring Company, SolarCity, and others, the answer is meaningful: these are the companies and components necessary to realize Musk’s ultimate vision, for humans to become an extra-planetary species.</p><p class="line-space"> <br> </p><p>(And interestingly, the biggest thing hurting the approval rating of Elon Musk, at the moment, is attached to the lack of focus he’s demonstrating with his involvement in Twitter. His operation of Twitter won’t help people get to Mars.)</p><h2 id="h-sam-altman-has-demonstrated-resilience-and-has-vision-and-focus">Sam Altman Has Demonstrated Resilience and Has Vision and Focus</h2>
<p>Like Musk’s (previously singular) focus on helping humans become an extra-planetary species, Altman’s focus is clear: artificial intelligence. He understands, as he described it in one interview, “the arc of technology and societal change on a long time horizon.” This has pushed Altman into new, related efforts. Among them:</p>
<p class="line-space"> <br> </p><ul>
<li>Recognizing the need for a stronger <strong>energy</strong> grid to fully realize his vision for the future of technology, Altman is the Chairman of the Board at <a href="https://oklo.com/overview/default.aspx?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">Oklo</a>, an advanced nuclear fission microreactor startup. As Sam Altman told <a href="https://www.cnbc.com/2023/07/11/sam-altman-talks-about-oklo-nuclear-microreactor-spac.html?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">CNBC</a>, “I think there’s urgent demand for tons and tons of cheap, safe, clean energy at scale.”</li>
</ul>
<p class="line-space"> <br> </p><ul>
<li>In recognition of a similar issue, Altman has pushed for a level of investment in semiconductors (and generally, <strong>compute power</strong>) to a previously-unheard of level. Earlier this month, it was revealed <a href="https://www.wsj.com/tech/ai/sam-altman-seeks-trillions-of-dollars-to-reshape-business-of-chips-and-ai-89ab3db0?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">Altman has sought up to $7 trillion in investment</a> to reshape the semiconductor industry.</li>
</ul>
<p class="line-space"> <br> </p><ul>
<li>Building the roads, bridges, and tunnels of artificial intelligence isn’t satisfactory for Altman, as he is rumored to be collaborating with Apple’s former designer Jony Ive on <strong><a href="https://sanfranciscodownload.com/apples-jony-ive-sam-altman-and-tang-tan-collaborate-on-groundbreaking-ai-device?ref=hackernoon.com" target="_blank" rel="noopener noreferrer ugc">building a hardware AI device</a></strong> that will (presumably) change the form of the way users currently experience artificial intelligence.</li>
</ul>
<h2 id="h-what-does-this-mean-for-sam-altman">What Does This Mean for Sam Altman?</h2>
<p>In addition to the personal implications of becoming a business celebrity, Altman will need to manage the bittersweet impact of this increased attention on his professional life.</p>
<p class="line-space"> <br> </p><p>His positioning has allowed (and will continue to allow) him to take shots like the one he took when floating a $7 trillion ask for investment in semiconductors. While most would be laughed off the stage, people leaned in and listened, truly considering a proposition previously unthinkable.</p><p class="line-space"> <br> </p><p>Similarly, he’ll likely be found more credible and see the benefit of his reputation when he launches the rumored AI hardware device, more likely to avoid the skepticism than first-movers like Humane and Rabbit.</p><p class="line-space"> <br> </p><p>He may find the attention is not all positive. When Oklo completes the process of going public via Altman’s SPAC ($ALCC), it seems an obvious candidate to become a favorite of fanboys and groups like r/WallStreetBets. While likely a good thing for the price, just ask Musk about the (potential) downsides of having so many people around the table.</p><p class="line-space"> <br> </p><p>That may make it difficult for Altman and his crew to remain nimble, a drawback in an early-stage environment. His status as a cult hero may also make it difficult to do the things that Altman has proven, time and time again, that he loves: tinkering with technology projects in public.</p><p class="line-space"> <br> </p><p>He may find it challenging to launch a fun Chrome extension weekend project. (Though I think he’ll manage.)</p><p class="line-space"> <br> </p><p>I ultimately do not know if it’s healthy to treat our business leaders with mythic-like status, nor do I know what to do with or about the phenomenon other than say this: I hope the path and result are better for Sam Altman than it has been for others like him.</p><p class="line-space"> <br> </p><p>With both Jobs and Musk, the story has been complicated, reminiscent of a line in Christopher Nolan’s 2008 Dark Knight: “You either die a hero or live long enough to see yourself become a villain.” With Jobs and Musk, I think it’s safe to say this has been true in each case.</p><p class="line-space"> <br> </p><p><strong>Here’s hoping Sam Altman can navigate his way to a third outcome, one where he can pursue his vision for the improvement of humanity, without devolving into an internet troll who releases a Cybertruck.</strong></p></div><div style="visibility:hidden" class="sc-1e366be2-7 kZfKay"><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M16 9h1v2h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1h-1v1H9v1H8v1H7v1H6v1H0v-6h1v-1h1v-1h1v-1h1v-1h1v-1h1v-1h1V9h1V8h1V7h1V6h1V5h2v1h1v1h1v1h1v1zM22 4v2h-1v1h-1v1h-1v1h-1V8h-1V7h-1V6h-1V5h-1V4h-1V3h1V2h1V1h1V0h2v1h1v1h1v1h1v1h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 3v2h-1v2h-1v1h-5V7h-2v1h-1v1H9v4h1v1h1v1h2v-1h5v1h1v2h1v2h-1v2h-1v1h-5v-1h-1v-2h-1v-2h-1v-1H9v-1H8v-1H7v1H2v-1H1v-2H0v-2h1V8h1V7h5v1h1V7h1V6h1V5h1V3h1V1h1V0h5v1h1v2h1z"></path></svg></button><button><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 22" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M14 19v2h-1v1H1v-1H0V5h1V4h3v15h10z"></path><path style="width:18px;fill:white" d="M20 6v11h-1v1H6v-1H5V1h1V0h8v6h6z"></path><path style="width:18px;fill:white" d="M20 4v1h-5V0h1v1h1v1h1v1h1v1h1z"></path></svg></button><a style="margin-left:15px;color:white" href="#commentSection"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 22 19" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M21 6V4h-1V3h-1V2h-2V1h-3V0H8v1H5v1H3v1H2v1H1v2H0v6h1v2h1v2H1v1H0v2h5v-1h1v-1h2v1h6v-1h3v-1h2v-1h1v-1h1v-2h1V6h-1ZM6 13h1v-1h1v-2H6V5h4v8H9v1H6v-1Zm6 0h1v-1h1v-2h-2V5h4v8h-1v1h-3v-1Z"></path></svg></a><button style="margin-right:10px"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 20 20" width="22" style="width:18px;fill:white"><path style="width:18px;fill:white" d="M20 2v12h-1v1h-2v1h-6v-1H4v1H3v4H1V3H0V1h1V0h2v1h1v2H3v1h1V3h7v1h6V3h2V2h1z"></path></svg></button><div class="sc-1e366be2-8 cbbaIe"></div></div></div>

If you connect the dots between what drove the love of Steve Jobs and Elon Musk, and recent events, it's clear the new cult hero in tech is OpenAI's Sam Altman.

Sam Altman: The Next Big Cult Hero in Tech

With a Collectibles portal on the Flow blockchain and deployed to the testnet we can now focus on creating the frontend using React.

How to Build a Digital Collectibles Portal Using Flow and Cadence (Part 2)

In this article, I'll detail how I developed an LLM app utilizing GPT-3.5 turbo LLM to assess available ingredients and recommend recipes.

Your AI-Powered Kitchen Assistant: Building a Recipe Genius With GPT3.5-turbo and Agenta

In this article explained of how I became a team lead, what skills you can’t do without in this position, and how to improve your team management skills

What Skills Will Help You Become a Good Development Team Leader?

Fine tuning allows us to overcome the limitation of base models that ChatGPT or LLama rely on to provide a response. Here's how to fine tune your base model!

More Stories