<?xml version="1.0" encoding="UTF-8"?>
<?xml-stylesheet type="text/xsl" href="/css/rss.xsl"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
	<channel>
		<title>Software and Tech stories from an Insider - iDiallo.com</title>
		<atom:link href="https://idiallo.com/feed.rss" rel="self" type="application/rss+xml" />
		<image>
			<url>https://cdn.idiallo.com/images/id-self-br.jpg</url>
			<title>Software and Tech stories from an Insider - iDiallo.com</title>
			<link>https://idiallo.com</link>
		</image>
		<description>
			<![CDATA[
		Throughout the years, I have decide to put all my the knowledge I have accumulated in my Blog. Hopefully it will serve others as well as it serves me.
		]]>
		</description>
		<link>https://idiallo.com</link>
		
			<item>
				<title><![CDATA[SO Is Back! No, Not Stack Overflow, but Its Spirit ]]></title>
				<link>https://idiallo.com/byte-size/stackoverflow-is-back-in-spirit</link>
				<author>rss@idiallo.com</author>
				
				<description>
					<![CDATA[
			<p>One thing people complained about most on Stack Overflow was having their questions closed and marked as duplicates of other questions that were only slightly related. The poster would argue in the comments, trying to make the case that their question was different, only for it to fall on deaf ears.</p>
<p>Then LLMs came along and ate Stack Overflow's lunch. An LLM will answer your question, and with even more vigor if a related question exists online. Everyone was happy, until the LLMs suddenly stopped responding.</p>
<p>Along the way, we decided that LLMs are too dangerous and shouldn't answer every development question. If your prompt sounds, even vaguely, like you are trying to circumvent security, the model drops your request and flags your account. Do it enough times, and your account is closed for good.</p>
<p>Now it's no longer smug developers closing questions. You are left with an unanswered prompt and no one to pin the fault on.</p>
<p>Even when they don't blatantly stop you in your tracks, the answers can feel gated, as if the model knows the answer but deliberately won't tell you.</p>
<p>Now, I was never so much against marking questions as duplicates. In the majority of cases I encountered, the person asking could have benefited from reading the original question. You don't get a full answer, but it serves as a guide that helps you find a solution. With LLMs withholding answers, the fix is often to pay for a more expensive model. So maybe it's not so much Stack Overflow as ExpertSexChange that is back.</p>						]]>
				</description>
				<pubDate>Thu, 19 Nov 2026 07:27:09 GMT</pubDate>
				<guid>https://idiallo.com/byte-size/stackoverflow-is-back-in-spirit</guid>
											</item>
		
			<item>
				<title><![CDATA[People will lose their jobs over GTA 6 ]]></title>
				<link>https://idiallo.com/blog/losing-your-job-over-a-video-game</link>
				<author>rss@idiallo.com</author>
				
				<description>
					<![CDATA[
			<p>GTA 6 is about to come out, and I've seen many people online joking that they'll take the week off to play it. I'm not a big fan of GTA, but the series did leave an impression on me. The last one I played was Vice City. It was an amazing game, and I still listen to its songs. It's where I first heard "Your Love" by The Outfield. I love the game, but I don't have the luxury of playing video games the way I used to.</p>			<p>All this to say that I understand people taking time off. When Metal Gear Solid 5 came out, I didn't just take the day off. I slowly prioritized the game over the work I was supposed to be doing. I had a contract with a client who couldn't make up his mind about a project. I delivered the project, a flower subscription service, in the span of a month. He had paid a team in Ukraine $25k to hand him a broken Drupal website. I rewrote it for half the price, and he finally had the ability to manage the website on his own.</p>
<p>But a website isn't complete once it's delivered. There are bugs and feature requests, or the client might just change his mind about things. So he retained me for maintenance. We launched in the summer, and he couldn't have been happier. Customers flocked in, and all was well. Now it was time for maintenance.</p>
<p>The problem was that the client called me every day. He wanted to change everything, all the time. My expert opinion had no effect on him, he just wanted things to be different. I was getting paid for it, so I made the changes. I was constantly on the phone with him that summer. It was annoying, but that's the life of a contractor. I could have kept going, but September 1st was the release date for MGS 5.</p>
<p>I could have managed if I had put my mind to it. But I had already pre-ordered the game, and I had played Ground Zeroes, the prelude that came out earlier, at least a thousand times. Nothing was going to get between me and my game.</p>
<p>When the game came out, I stopped answering calls. Instead, I replied to emails. Whenever the title "A Hideo Kojima Game by Hideo Kojima Produced by Hideo Himself" appeared on the screen, I used the time to respond to email. Some of the work he requested was never completed, because I knew he would change his mind. The more I ignored the client's calls, the more frustrated he became, and the more of the story I missed. "Who is Skull Face again?" I didn't want to miss a second of the dialogue.</p>
<p>At one point, the client called and I didn't pick up, as usual. He sent a text message: "PICK UP NOW!" I knew it was serious. So I didn't pick up right away, because doing so would have made it look like I had been intentionally ignoring him. I called him back some minutes later, only to find out it was nothing urgent. The call turned into a therapy session. He was frustrated that people had flocked to the website during the grand opening, yet it had translated into few subscriptions. The website was only a few months old, mind you.</p>
<p>I could have given better advice, but Quiet was shooting at me from a distance, and I hadn't figured out how to win that battle yet. Besides, I was there to do technical work, not to solve his business problems. He sensed it. When he found a bug on the website, I was eager to fix it. When he wanted to discuss ways to make the subscriptions more appealing to customers, I checked out.</p>
<p>Eventually, he confronted me and asked whether I was all in on this business. I wasn't. I was busy thinking that the Huey character design was lazy. Why did they recycle Otacon? Couldn't they come up with something new instead of playing on our nostalgia? Anyway, I knew a couple of friends who had just started a web consulting business. They could solve business problems and do the development work, so they would be more helpful than I was. Plus, it wouldn't make sense for him to retain me while working with them.</p>
<p>I found them their very first client, and the client found a better team.</p>
<p>My consulting pattern used to be to deliver a job and then, during the maintenance phase, look for other clients. That way, I never had downtime. But this time, I used the downtime to play my game. Remember, this was 2015. $13k went so much further then than it does now.</p>
<p>You can imagine my disappointment when the game ended abruptly, as if the producer had been kicked out mid-development. Kojima had been ousted from Konami. Still, it was a good game. But I needed money to survive, so for my next contract, I ended up charging <a href="https://idiallo.com/blog/18000-dollars-static-web-page">$18,000 for a single static HTML page</a>.</p>
<p>What I mean is, as an MGS fan, I understand the GTA fans.</p>			]]>
				</description>
				<pubDate>Mon, 16 Nov 2026 00:40:21 GMT</pubDate>
				<guid>https://idiallo.com/blog/losing-your-job-over-a-video-game</guid>
												<enclosure url="https://cdn.idiallo.com/images/assets/695/hero.jpg" type="image/jpeg" length="102400"/>
							</item>
		
			<item>
				<title><![CDATA[So you want to buy AI for your company ]]></title>
				<link>https://idiallo.com/byte-size/buying-ai-for-your-company</link>
				<author>rss@idiallo.com</author>
				
				<description>
					<![CDATA[
			<p><em>Where do I start?</em></p>
<p>We have Gemini 4 Hydrogen. But it sounds like you could use Claude Sonnet 5.5. Or maybe 5.4; it really depends. Have you tried Opus? We have Opus 5.2 right now. It's similar to Fable 5.2, but better for day-to-day work. It's not quite as powerful as Mythos 5.2, though. I hear the gap will narrow eventually. Opus scores 94.8 on a benchmark, while Mythos gets 96.1. Fable is a respectable 92.7, assuming you don't ask it to do anything too ambitious.</p>
<p>Have you tried the Chinese models? DeepSeek is pretty good. You might want to start with the R1 family to test the waters before moving on to V4. Personally, I'm a Flash kind of guy, but Pro is pretty good too. The crowd favorite is GLM these days. You've got to try 5.3 Flash. Its Mixture-of-Experts architecture is state of the art, in my opinion. It scores an impressive 87.3 on the Benchmark Performance Index.</p>
<p>But let's be serious for a minute. You're probably looking for ChatGPT. Astra is overkill unless you want to do some damage. I'd stick with GPT-6.0 Terra, not Sol! That one ate through my budget in three days. Terra gets 98.2 on the Enterprise Productivity Benchmark, while Sol gets 99.1. That extra 0.9 will cost you $14,000 a month. So go with Luna if you must.</p>
<p>If you're feeling more adventurous, you could go with Qwen, though. I'd stick with 3.8 MoE. Don't even think about 4 Plus. It's a nightmare. Kimi K is dear to my heart, though. You can't go wrong with it. I use it alongside Codex to complement my workflow. Really great. It scored 91.6 on SWE.</p>
<p>Those are just some of the options.</p>
<p>If you're a programmer, you probably don't need any of that. You just want to use Claude Code. Or get Cursor. Make sure to remove Grok from the auto-selection options. You don't want Grok. If you don't know why, then you don't need to know why.</p>
<p>Anyway, I use Jules with Antigravity now. In VS Code, I use Continue with a local R1 Flash running on Ollama. I use Qwen3.5-397B-A17B for setting up machines, and Qwen-VL to read through my camera feeds. The latter is particularly good at detecting forklifts.</p>
<p>And that's pretty much it. A fairly simple setup, really.</p>
<p>For the layperson, though, it's much easier: they just use the free ChatGPT tier and call it a day... Or Argon. It depends really. Google 3.6 Flash has the lowest score, but it is great for email. I mean how mean do you need your email game to be? Anyway, that's it.</p>
<p>Nano Banana! You probably need that too. It's great with Omni or Carbon. Or wait next week when Hydrocarbon comes out.</p>
<p>That's it!</p>						]]>
				</description>
				<pubDate>Fri, 13 Nov 2026 00:34:41 GMT</pubDate>
				<guid>https://idiallo.com/byte-size/buying-ai-for-your-company</guid>
											</item>
		
			<item>
				<title><![CDATA[Writing English is Solved ]]></title>
				<link>https://idiallo.com/byte-size/english-is-solved</link>
				<author>rss@idiallo.com</author>
				
				<description>
					<![CDATA[
			<p>Maybe you still struggle to write in Pulaar, or in your rinky-dink language. But here in the AI-connected, English-speaking world, we don't need to write anymore. In fact, I don't remember the last time I wrote English. I have one master agent that I talk to and tell what I need to get done. It breaks down the tasks and serves them to spawned agents that get the work done.</p>
<p>I spend $3,000 a month on tokens. I use them mostly for:</p>
<ul>
<li>Research and search</li>
<li>Summarization and transcription</li>
<li>Writing (the big one)</li>
<li>Art and music</li>
</ul>
<p>My great transition this year, as a programmer, was that once I accepted that AI can merge code without needing my review, I could let go. What this meant was that when AI writes English for me, I no longer need to review it before publishing. It can write English, and it writes it well.</p>
<p>Okay, let me add this here before I go too far: <code>/s</code>.</p>
<hr />
<p>I read this blog post yesterday, and I'm always baffled to hear technical people <a href="https://ben.stolovitz.com/posts/how-use-ai-oct-2026/">write these words</a>:</p>
<blockquote>
<p>Writing code is solved.</p>
</blockquote>
<p>As a software engineer who uses AI, I feel like I must be holding it wrong. I've found a correlation between the number of lines of code written and the number of bugs found. Whether a person or a model generates the code, the longer the PR, the higher the odds of finding an issue. So when someone says coding is solved, I automatically assume they were mediocre at best before using these tools.</p>
<p>If we look at it technically, we could say, in the same vein, that writing English is solved. The AI can write correct sentences, write paragraphs, or publish an entire document. If you have doubts about the content, you can have a second AI verify it and fix potential hallucinations. But of course, that's not how it works in the real world. The moment people detect AI-generated text, even if it is after reading the whole text, they reject it.</p>
<p>It feels empty, deceitful. What is an anecdote from an LLM to convey, when the experience never actually happened? Yes, code is not to be consumed like English. It is for the machine. But it is subjective in its purpose because we humans have to read it eventually. </p>
<p>Writing code is not solved, just as writing English cannot be solved. Writing is subjective and has no solution because it isn't a problem. If anything, AI for writing is a testament to this saying:</p>
<blockquote>
<p>"Je n'ai fait celle-ci plus longue que parce que je n'ai pas eu le loisir de la faire plus courte." &mdash; Blaise Pascal</p>
</blockquote>						]]>
				</description>
				<pubDate>Fri, 06 Nov 2026 08:00:00 GMT</pubDate>
				<guid>https://idiallo.com/byte-size/english-is-solved</guid>
											</item>
		
			<item>
				<title><![CDATA[It's unbreakable in theory ]]></title>
				<link>https://idiallo.com/byte-size/unbreakable-in-theory</link>
				<author>rss@idiallo.com</author>
				
				<description>
					<![CDATA[
			<p>I've seen the picture of the iPhone Duo, still in the store, perched on a stand, its screen flickering with broken white lines.</p>
<div class="art-image">
  <img src="https://cdn.idiallo.com/images/assets/daily/121/duo.jpg" alt="broken duo" />
</div>
<p>Apple has touted testing the screen hinge, opening it mechanically thousands of times. The phone survived every test it was put through. Yet here it is, on the stand, with a broken screen.</p>
<blockquote>
<p>Wait till the product meets its first customer.</p>
</blockquote>
<p>I'm sure the tests worked in theory and in the delicate hands of its engineers. But customers' hands are anything but delicate. Remember Elon Musk's wrecking ball? They must have tested the Cybertruck's window beforehand, but it took just one person outside the test group to break the glass with a single throw.</p>
<p>I remember when my nephew was barely a year old, not even able to walk. I handed him a wooden spoon in the living room as a distraction while I ran to the kitchen to get some food. As soon as I got there, I heard a snap. I came back to find this little boy holding the spoon, broken in two. It was a hard spoon, and there was no way his one-year-old hands could break it in half, yet there was the proof in front of me.</p>
<p>So yes, Apple tested the rigidity of the screen with fancy machines, but the iPhone Duo never stood a chance against the customer's hands. Unless we are holding it wrong?</p>						]]>
				</description>
				<pubDate>Thu, 05 Nov 2026 20:20:10 GMT</pubDate>
				<guid>https://idiallo.com/byte-size/unbreakable-in-theory</guid>
											</item>
		
			<item>
				<title><![CDATA[Goodhart's Law, but We Already Spent All Our Money ]]></title>
				<link>https://idiallo.com/blog/we-spent-all-our-money-goodharts-law</link>
				<author>rss@idiallo.com</author>
				
				<description>
					<![CDATA[
			<p>Here is how to make any sports fan angry. Basketball is just a bunch of grown men bouncing a ball and throwing it into a basket. Football is a bunch of grown men running on grass trying to kick a ball into a net, yes <em>that</em> football. And baseball... Oh boy, baseball.</p>			<p>I've used these lines many times to piss people off and to justify my complete lack of athletic feats. But the counter-argument I always get is that football is a multi-billion-dollar industry, or that a specific player makes more money than I will ever see in a lifetime. If I counter that video games are just for kids, because of course I'd say that, I get a similar response: <em>Video games make more money than all of these sports combined.</em></p>
<p>It's all true. But nobody is telling me <em>why</em> basketball is important. Nobody tells me how good the game is, how people bond over it, how it brings a community together, or that it is simply healthy to stay active. Whether it's basketball, football, video games, or (God forgive me) baseball, it all boils down to the exact same measuring stick: how much money the activity generates.</p>
<blockquote>
<p><em>If the importance of a thing is only measured by how much money it makes, then we will aim to make more money despite the thing itself.</em> &mdash; An AI Somewhere.</p>
</blockquote>
<p>Goodhart's Law states that when a measure becomes a target, it ceases to be a good measure. Enter Artificial Intelligence.</p>
<p>This wave of generative AI has supposedly generated billions of dollars since ChatGPT first launched. But in the process, vastly more money was spent to make it possible. Money was spent to make AI look good. AI is deemed "good" because it is expected to make lots of money. Because AI isn't actually making that money back yet, it relies on inflated valuations to define its worth in monetary terms.</p>
<p>It's like running a pizzeria where the goal is to make a million pizzas, but you don't actually have the ingredients to bake them. Instead of buying flour and tomatoes, you use AI to generate high-resolution photos of delicious pizzas and print them on empty boxes where the future pizza will be. You sell the hype of the box, promising that the actual pizza will arrive once the funding hits.</p>
<p>There is this simple argument, almost naive I would say: <em>If AI is going to take all our jobs, who is going to pay for it?</em></p>
<p>If workers are displaced, they have no income. If they have no income, they have no purchasing power. Who is this mythical customer who doesn't work, receives no paycheck, yet seamlessly pays a monthly subscription for enterprise AI tools?</p>
<p>Currently, the primary consumer of AI isn't the public buying a finished product. Instead, it's corporations selling AI back and forth to each other in an enterprise circular economy, funded by venture capital that expects infinite returns on a product nobody has fully figured out how to monetize at scale.</p>
<p>This may be an anecdote where <code>N = {"my coworkers over the past four years"}</code>, but in that short timespan, we were visited by the Four Horsemen of Corporate AI. </p>
<p>The First Horseman told us never to touch AI, warning of IP theft and security leaks. The Second Horseman told us to use it cautiously "where needed." The Third Horseman turned AI into a game, tracking metrics to see who could use it the most, (the winner got a free iphone). The Fourth Horseman arrived recently, asking us to stop using Fable, and get written managerial approval to exceed a reduced token budget.</p>
<p>Each horseman was a paid consultant hired to provide training. In just a few years, corporate sentiment swung from paranoia to forced obsession, and finally to budget panic. Employees went from being pushed to "Tokenmaxx" everything to being scolded that tokens cost real money.</p>
<blockquote>
<p>AI is important because we spend so much money on it. </p>
</blockquote>
<p>When an industry measures success by the dollar amount poured into it rather than the actual utility derived from it, we start living in a fantasy. We spent astronomical sums of money to build AI because we believed it was valuable. Now, we must prove it is valuable simply because we spent all that money.</p>
<p>If we keep measuring the value of technology solely by market capitalizations and speculative hype rather than usefulness, we shouldn't be surprised when people stop buying empty pizza boxes. Beautifully printed empty pizza boxes I should say. Right now, we are running out of money to buy the ingredients to actually make the pizza.</p>			]]>
				</description>
				<pubDate>Wed, 04 Nov 2026 08:00:00 GMT</pubDate>
				<guid>https://idiallo.com/blog/we-spent-all-our-money-goodharts-law</guid>
												<enclosure url="https://cdn.idiallo.com/images/assets/694/hero.jpg" type="image/jpeg" length="102400"/>
							</item>
		
			<item>
				<title><![CDATA[Prove you are not a child ]]></title>
				<link>https://idiallo.com/blog/prove-you-are-not-a-child</link>
				<author>rss@idiallo.com</author>
				
				<description>
					<![CDATA[
			<p>I'm 100% for banning social media for children. I would even extend it to banning it all together in its current form. But I am also against the methods for enforcement.</p>			<p>The problem with banning social media for one group and not another is that you have to differentiate the different groups somehow. And right now, in an age where information and personal data means everything, there are several entities that are salivating to get this data.</p>
<p>I've written before about all the problems that come from <a href="https://idiallo.com/blog/your-id-online-and-offline">uploading sensitive data online</a>. It doesn't even matter if they delete the data afterward, as long as the data goes through the wire in the first place, there is a chance for a leak.</p>
<blockquote>
<p>What appears to be the same simple request "please verify your identity", becomes fundamentally different when mediated by technology. The question isn't whether these digital systems will be compromised, but when. And unlike that movie theater clerk who can't perfectly recall my birthdate minutes after seeing it, computers have perfect memory. They store, copy, backup, and transmit our most sensitive information through networks we don't control, to companies we've never heard of, under policies we'll never read.</p>
</blockquote>
<p>When kids are not allowed to use social media, it's the adult who pays the price. It's the adult who has to prove that they are not children. They have to risk uploading their government ID to a third party, or biometric data that may be saved and backed up forever. </p>
<p>I reject this method. </p>
<p>I also reject the method where I have to provide my age to the operating system. No system should compel me to provide personal data in order to use it. Especially if this feature is intended to provide the information to a third party. And what happens if I don't provide this information to my freebsd machine? Or worse, what if I provide false information? Am I going to get punished? How does that protect children in any way?</p>
<blockquote>
<p>Breaking news: Grown up arrested when OS reported forged identity in his laptop.</p>
</blockquote>
<p>One question I asked myself was what happens when a minor circumvents the ban. What are the consequences? At least in Australia, a country that has implemented a social media ban for those under 16, <a href="https://www.politico.com/news/2026/08/16/australia-social-media-ban-explained-01038852">there are no consequences</a>.</p>
<blockquote>
<p>The law places legal responsibility on platforms, not children or parents, and it currently applies to Facebook, Instagram, Kick, Reddit, Snapchat, Threads, TikTok, Twitch, X and YouTube.</p>
</blockquote>
<p>This almost sounds reasonable. Why should you care, if it becomes facebook responsibility when you don't even own a facebook account? But there is always a caveat. When facebook becomes a tool your school uses to communicate with parents. When Google classroom is the service students use to do their homework. When you are required to login on a system to receive government service. You may not care about a particular social media, but they raise the bar for every other entrant in the game. Every website frequented by a large number of people inherently become social media, and demand data you didn't need to provide before.</p>
<p>Technology is not a magical thing that can just solve every problem. It is especially weak when it comes to social problems. However, society can use technology to solve their problems. The responsibility of children falls onto their parents. And there are great tools for parents that their kids access on their devices. </p>			]]>
				</description>
				<pubDate>Tue, 03 Nov 2026 00:00:00 GMT</pubDate>
				<guid>https://idiallo.com/blog/prove-you-are-not-a-child</guid>
												<enclosure url="https://cdn.idiallo.com/images/assets/693/hero.jpg" type="image/jpeg" length="102400"/>
							</item>
		
			<item>
				<title><![CDATA[College will only cost $20 a month ]]></title>
				<link>https://idiallo.com/blog/college-will-cost-20-dollars</link>
				<author>rss@idiallo.com</author>
				
				<description>
					<![CDATA[
			<p>From 2020 to 2022, I've encouraged a steady stream of students to pursue their education in computer science. Tech companies were hiring like crazy. I would interview 3 candidates a week on some occasions. The future of tech seemed to be paved with gold. Any day now, several of those students will graduate. I dread opening my inbox and get an email from them asking:</p>			<p>"So... about those jobs?"</p>
<p>Things have changed since those crazy old days. Hiring has come to a stop. Those graduates will step into a market that looks nothing like what they signed up for. Getting a job will be a struggle. If I could see the future, I would have recommended a different path. But then again, how could I have known? Hindsight is always 20/20.</p>
<p>So this leaves me with a dilemma as a parent. What do I tell my own kids? Am I going to push them to pursue the same traditional colleges? Will I disburse life savings so they could attend a prestigious university? Or will I just buy them a $20 a month AI Subscription?</p>
<p>Imagine handing a student a textbook straight from a university syllabus. They read a lesson, and then you have an AI tutor to quiz them, debate concepts, and break down difficult ideas until they fully grasp the material. Is that a good replacement for college?</p>
<p>Some argue that college is about the experience. It's the friendships you make along the way, the networking, the independence. These same people often omit the lifetime of crippling debt that hangs over those friends twenty years later.</p>
<p>When I was in college, I focused almost entirely on mastering the material in the books, completing all assignments quickly. I spent the rest of the semesters tutoring my peers and applying the material I had learned. For me replacing the entire college system, the brick-and-mortar university with a set of books and an AI subscription seems like it would have done the job. I was the target candidate for this model.</p>
<p>Yet, now that I have these AIs to experiment it myself, I found two flaws in the model:</p>
<ol>
<li>Understanding takes time.</li>
<li>Not having an immediate answer forces you to strain your brain until you figure things out.</li>
</ol>
<p>When I wrestled with dense textbooks in college, solutions didn't arrive at the click of a button. Instead, I struggled. I hit dead ends. I tried solutions that seemed promising only to abruptly stop two hours in the process only to realize that the whole thing was wrong. Sitting with the problem, the trials, the errors, giving up and going for a walk to clear my head, that was the process that taught me how to think. It taught me how to debug.</p>
<p>Even when I found solutions online and pasted them into my code, I had to encounter the same problem repeatedly before the concept finally made sense. You have to work your way through a solution to <a href="https://idiallo.com/blog/i-know-kung-fu">understand a problem</a>.</p>
<p>This is why I feel concerned when we throw Millennium Prize Mathematical problems at AI. Imagine an AI presents a complete proof to one of these unsolved problems tomorrow? If or when AI solves it, we still have two problems. </p>
<ol>
<li>We still have to understand the solution. Having an AI produce a proof doesn't mean we understand how or why it works.</li>
<li>We don't get to discover new problems. </li>
</ol>
<p>The second one is often overlooked, but ask yourself, where do these problems in Mathematics come from? They come from mathematicians attempting to solve hard problems, taking detours and failing along the way. Remove that step, and you... Well, you don't find any new problems.</p>
<p>Not just in mathematics, I feel we will see something similar in code where fewer applications rely on <a href="https://idiallo.com/blog/writing-code-is-easy-reading-is-hard">frameworks</a>. Frameworks solve real issues in programming, but why use a framework when the AI can just give you all the answers. I lived through the JS framework mania. I am not suggesting we create more, but I think we will miss out on discovering new problems that need solving.</p>
<p>If we rely on a system that just hands us a solution, we lose the ability to uncover new questions because we never went through the steps of trying something new.</p>
<hr />
<p>I don't think a $20-a-month AI subscription paired with a textbook is the ultimate solution. Not when we still need some friction in order to inculcate concepts in our minds. But traditional higher education has changed so drastically and the skyrocketing costs and changing employment realities don’t help. There is no turning back to the old model either.</p>
<p>The challenge for learning remains understanding. How we grasp concepts and develop critical thinking in an era of instant answers. Maybe that's where we should turn our focus to. Figuring out how to design learning that somehow preserves productive struggle. If AI can solve this issue and turn students into masters of crafts instead of prompt engineers, I mean, I'll gladly keep that college tuition in my pocket.</p>			]]>
				</description>
				<pubDate>Fri, 30 Oct 2026 23:35:13 GMT</pubDate>
				<guid>https://idiallo.com/blog/college-will-cost-20-dollars</guid>
											</item>
		
			<item>
				<title><![CDATA[Productivity ]]></title>
				<link>https://idiallo.com/byte-size/productivity-video</link>
				<author>rss@idiallo.com</author>
				
				<description>
					<![CDATA[
			<blockquote>
<p>I'd say in a given week I probably only do about fifteen minutes of real, actual, work. &mdash; Office Space</p>
</blockquote>
<p>I was watching a video where a guy said he has 8 agents working in the background. He was stressed out and was trying to "get shit done." At one point he said "I have no time to myself." </p>
<p>That was very confusing. Let's assume for a second that the AI agents were doing all the work for him. Maybe it was some tedious work that would have taken him so much time to do manually. Now, he doesn't have to do the work because someone else is doing it. Why is he still busy?</p>
<div class="art-video"><iframe width="720" height="480" src="https://www.youtube.com/embed/Eb9HgqKbCDk"></iframe><p>The Future we were promised</p></div>
<p>It seems like the goal of productivity is to reduce the amount of work needed to create a set of results. But the moment you create the same results in less time, you are incentivised to use the new gained time to produce even more results. </p>
<blockquote>
<p>You will rest when you grow up &mdash; My Mom.</p>
</blockquote>
<p>But now that we are grown up, our employer changed the tune just a bit:</p>
<blockquote>
<p>You will rest when you are dead. </p>
</blockquote>						]]>
				</description>
				<pubDate>Thu, 22 Oct 2026 16:56:14 GMT</pubDate>
				<guid>https://idiallo.com/byte-size/productivity-video</guid>
											</item>
		
			<item>
				<title><![CDATA[Can they write a function? ]]></title>
				<link>https://idiallo.com/blog/can-they-write-a-function</link>
				<author>rss@idiallo.com</author>
				
				<description>
					<![CDATA[
			<p>For a software developer position, there isn't a trick to finding the best candidate. Your criteria do not determine how solid the candidates will be. Everyone is submitting 500 applications a month now, and those who automate the process submit thousands.</p>			<p>So how do you hire a great candidate? Pick a few from the lot, ask them if they can write a function, have them do a live FizzBuzz test, then ask them to deploy a service from localhost to an actual public server. That's it. Hire them after that.</p>
<p>As far as experience is concerned, they will learn on the job. If you think they don't know how to write tests or use a particular paradigm, they can ask ChatGPT and apply what it tells them a few times until they get familiar with it. <a href="https://idiallo.com/blog/work-taught-me-how-to-think">People learn on the job</a>. Nobody knows what the hell back pressure is until they encounter it in a particular service.</p>
<p>Everything else is really just random. We insist on things like culture, but we have no way of measuring it. It's all so arbitrary.</p>
<p>For example, early in my career I had an in-person interview where the head of HR brought me close to tears. She would make a joke, I would laugh, and while I was laughing she would stop and vigorously take notes. While she wrote, she would casually ask, "Why do you look so stressed out?" I nearly choked. I had just lost my father, and I had rescheduled the first interview so I could bury him.</p>
<p>The VP of engineering pulled a similar stunt on me. I did 6 interviews in 3 days, and they called me for a 7th. I lay on my bed, ready to give up. But I had a whole family to take care of and a giant shoe to fill, so I decided to do it. That same evening, I got an email saying the interview was cancelled; instead, they were ready to hire me. I didn't know whether I was supposed to be happy or mad.</p>
<p>It took 2 years on the job to figure out why they had done this to me. The person I was hired to replace had gone from a great employee to a complete psycho. Once, a teammate made the mistake of rejecting his PR. When he saw the rejection, he got up from his desk, climbed over the cubicles, and walked from desk to desk all the way to her cubicle. With his foot on her screen, he said, "What the fuck did you just do?"</p>
<p>So I learned that the culture and psychological test I was put through was meant to filter out this type of person. It could have been called a success, because I was well liked and never threatened anyone in my 4-year tenure at the company. However, that same test couldn't stop them from hiring two employees who ended up being convicted of murder. One of them had a <a href="https://www.courttv.com/news/gareth-pursehouse-to-be-sentenced-for-ex-girlfriends-murder/">well-televised trial</a> and is currently serving life without parole.</p>
<p>So those culture interviews are only designed to please a manager somewhere. They have no documented effectiveness.</p>
<p>On Reddit's recruiting hell topic, you'll find mind-blowing stories of candidates being rejected for no good reason.</p>
<p>You might ask:</p>
<blockquote>
<p>“What about experts? Surely the bar is higher than just writing a function, right?”</p>
</blockquote>
<p>It's much easier to find an expert than people tend to admit. If you sort through your resume list right now and use only the criterion of 10-plus years in the field, you'll have your expert. Pick one, ask them to talk about their field of expertise, then hire them. If you're afraid they might not be the best candidate, I'm afraid to say that there is no way to determine that before they start working for you.</p>
<p>The problem is that companies want those experts at the price of an undergrad. Or they think they need to build the next Google when they run a small business with a manageable number of customers.</p>
<p>Hiring is only hard because we approach it like a dating app. We have unrealistic criteria for made-up scenarios. If they can write a function, you can hire them.</p>			]]>
				</description>
				<pubDate>Wed, 21 Oct 2026 19:00:40 GMT</pubDate>
				<guid>https://idiallo.com/blog/can-they-write-a-function</guid>
											</item>
		
	</channel>
</rss>