<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Hacker News: tech_ken</title><link>https://news.ycombinator.com/user?id=tech_ken</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Tue, 25 Aug 2026 03:04:33 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=tech_ken" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by tech_ken in "I were 17, I'd learn how to build LLMs from scratch"]]></title><description><![CDATA[
<p>Oh yeah I absolutely agree that there's lots of greenfield for relative newcomers to find some big improvements! I'm just saying that reasoning about 'best practices' in terms of what works for something like a CPU is missing some important and relevant differences.</p>
]]></description><pubDate>Mon, 24 Aug 2026 19:44:57 +0000</pubDate><link>https://news.ycombinator.com/item?id=49424901</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=49424901</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49424901</guid></item><item><title><![CDATA[New comment by tech_ken in "I were 17, I'd learn how to build LLMs from scratch"]]></title><description><![CDATA[
<p>But in the case of CPU architecture there are SOME people who understand how things work 100%, and they've built and vetted abstractions/mental models that enable other engineers and scientists to have that kind of mixed shallow/deep understanding in a way that works. On the side of LLMs we're still lacking an expertise which could flawlessly explain how these things operate; the abstractions that we're using are instead derived inductively and are totally unvetted.</p>
]]></description><pubDate>Mon, 24 Aug 2026 17:56:57 +0000</pubDate><link>https://news.ycombinator.com/item?id=49423535</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=49423535</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49423535</guid></item><item><title><![CDATA[New comment by tech_ken in "I were 17, I'd learn how to build LLMs from scratch"]]></title><description><![CDATA[
<p>Developing an intuition about high dimensional systems is pretty different from understanding the character of some specific point on a 1e9+ dimensional manifold of parameters, in my professional opinion (setting aside all the degrees of freedom that come from the structure of the thing). Sure one can understand generic principles like the curse of dimensionality, but truly groking how an LLM works is basically an open problem as far as I'm aware. I'm not saying there's no benefit for amateurs  to study how LLMs work, but let's be realistic about how far mere intuition can truly take <i>anyone</i> in this space.</p>
]]></description><pubDate>Mon, 24 Aug 2026 17:27:04 +0000</pubDate><link>https://news.ycombinator.com/item?id=49423120</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=49423120</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49423120</guid></item><item><title><![CDATA[New comment by tech_ken in "Project Cybersyn (2022)"]]></title><description><![CDATA[
<p>> it makes me wonder what more democratic, free market companies would be like.<p>I was wondering the same after PRW lol. Recently I had the chance to ask a friend of mine who worked in Google Infra (years ago) if they had ever tried to 'liberalize' their infrastructure planning; something like having internal business units bid on space, power, and compute instead of distributing it top-down. He said that yes, this was actually something like their original operating model, and the outcome was that YouTube basically bought up every piece of hardware in the company and then leased them out to other verticals at a profit.</p>
]]></description><pubDate>Thu, 20 Aug 2026 19:23:23 +0000</pubDate><link>https://news.ycombinator.com/item?id=49378965</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=49378965</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49378965</guid></item><item><title><![CDATA[New comment by tech_ken in "Project Cybersyn (2022)"]]></title><description><![CDATA[
<p>I'd honestly recommend checking out the People's Republic of Walmart; it's a super short read and anything I say would be basically a warmed up rehash of whatever I remember from it.<p>Basic point is just that operations research/supply chain management was heavily developed by the USSR to support the central economy and now it's used by every major retailer (among others) to manage supply chains that are massively larger than anything the USSR ever got close to.</p>
]]></description><pubDate>Thu, 20 Aug 2026 19:19:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=49378931</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=49378931</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49378931</guid></item><item><title><![CDATA[New comment by tech_ken in "Project Cybersyn (2022)"]]></title><description><![CDATA[
<p>Back when I read "Cybernetic Revolutionaries" I followed it up with "People's Republic of Walmart" and came away with a bit of a different conclusion to your point 1. IMO lots of US-based hyperscalars are managing what are effectively command economies, and it's allowed them to ex. scale their infrastructure so aggressively these last ~2 years.</p>
]]></description><pubDate>Thu, 20 Aug 2026 18:51:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=49378586</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=49378586</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49378586</guid></item><item><title><![CDATA[New comment by tech_ken in "Shinjuku Station in 3D"]]></title><description><![CDATA[
<p>Oh duh</p>
]]></description><pubDate>Wed, 22 Jul 2026 13:44:31 +0000</pubDate><link>https://news.ycombinator.com/item?id=49006722</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=49006722</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49006722</guid></item><item><title><![CDATA[New comment by tech_ken in "Shinjuku Station in 3D"]]></title><description><![CDATA[
<p>Finally I'll be able to properly follow JJK season 2</p>
]]></description><pubDate>Mon, 20 Jul 2026 16:34:51 +0000</pubDate><link>https://news.ycombinator.com/item?id=48981180</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=48981180</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48981180</guid></item><item><title><![CDATA[New comment by tech_ken in "Antares achieves criticality of Mark-0 reactor"]]></title><description><![CDATA[
<p>> It's an abstraction, not an error.<p>Extending an abstraction beyond its expressive capabilities is an error IMO.</p>
]]></description><pubDate>Tue, 30 Jun 2026 18:14:29 +0000</pubDate><link>https://news.ycombinator.com/item?id=48736870</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=48736870</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48736870</guid></item><item><title><![CDATA[New comment by tech_ken in "Antares achieves criticality of Mark-0 reactor"]]></title><description><![CDATA[
<p>Right but “upgrading to the next level” is exactly the type of cognitive error that I’m against, it’s forward-looking Whig history. There is no “levels” of civilization, there’s only the arrangements we have now and the ones that will occur in the future. Maybe those arrangements will have different pros and cons than we have now, maybe they’ll have more sophisticated engineering practices, but there’s no objective “development score” that is being maximized it’s just more humanity.</p>
]]></description><pubDate>Tue, 30 Jun 2026 14:52:21 +0000</pubDate><link>https://news.ycombinator.com/item?id=48733535</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=48733535</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48733535</guid></item><item><title><![CDATA[New comment by tech_ken in "Antares achieves criticality of Mark-0 reactor"]]></title><description><![CDATA[
<p>Tech development != tech tree, history is nonlinear and ecological. Linear, unlockable "tech trees" are an ahistorical fantasy necessary to make some video games fun and digestible. Applying them to the real world is like trying to win a war by replicating the Battle of Helm's Deep or something.</p>
]]></description><pubDate>Tue, 30 Jun 2026 14:36:07 +0000</pubDate><link>https://news.ycombinator.com/item?id=48733294</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=48733294</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48733294</guid></item><item><title><![CDATA[New comment by tech_ken in "Antares achieves criticality of Mark-0 reactor"]]></title><description><![CDATA[
<p>From the article:
> We said criticality in 2026, electricity production in 2027, and power to the warfighter in 2028.<p>Are there any other examples of land-based militaries using nuclear power? Seems kind of like since they can't talk about the energy transition or w/e this has to be a military thing instead.</p>
]]></description><pubDate>Tue, 30 Jun 2026 14:04:16 +0000</pubDate><link>https://news.ycombinator.com/item?id=48732908</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=48732908</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48732908</guid></item><item><title><![CDATA[New comment by tech_ken in "Antares achieves criticality of Mark-0 reactor"]]></title><description><![CDATA[
<p>> moving up the tech tree of civilization<p>Life isn't Civ 6, there is no tech tree to climb.</p>
]]></description><pubDate>Tue, 30 Jun 2026 13:59:30 +0000</pubDate><link>https://news.ycombinator.com/item?id=48732859</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=48732859</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48732859</guid></item><item><title><![CDATA[New comment by tech_ken in "Antares achieves criticality of Mark-0 reactor"]]></title><description><![CDATA[
<p>It's more of an engineering call than something that can be purely determined from inductive reasoning. I think most engineers working in the space would say "both" are needed, but partisans exist on both sides.</p>
]]></description><pubDate>Tue, 30 Jun 2026 13:56:24 +0000</pubDate><link>https://news.ycombinator.com/item?id=48732824</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=48732824</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48732824</guid></item><item><title><![CDATA[New comment by tech_ken in "Tacky men with ridiculous glasses want you to wear them too"]]></title><description><![CDATA[
<p>I think there’s a difference between full-on “bullying” and what I’d call “blunt criticism”. I think bullying requires an intent to silence or otherwise coerce through fear. Blunt criticism is more about skipping the niceties and saying it plainly, and accepting that your subject might get offended as a result. Calling someone “tacky” is way more in the blunt criticism camp I think. It’s certainly not “nice”, but there’s no way that I read this blog article as trying to change Mark Zuckerberg’s mind through fear. Nobody is getting pilloried for being tacky, and tackiness can even be repositioned as desirable through the lens of nostalgia or similar.</p>
]]></description><pubDate>Mon, 22 Jun 2026 22:45:17 +0000</pubDate><link>https://news.ycombinator.com/item?id=48637457</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=48637457</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48637457</guid></item><item><title><![CDATA[New comment by tech_ken in "NSF slashes research programs to support new tech initiative, insiders say"]]></title><description><![CDATA[
<p>> The government is going to suck at funding the right things<p>I'm pretty sure you have this totally backwards. People who study scientific development seem to think that the government is actually a really effective funder of research, and covers gaps that would never be addressed by private industry. See for example:<p>* <a href="https://www.aau.edu/newsroom/leading-research-universities-report/new-research-suggests-returns-federal-investments-rd" rel="nofollow">https://www.aau.edu/newsroom/leading-research-universities-r...</a><p>* <a href="https://news.northwestern.edu/stories/2022/07/government-funded-scientific-research-reflects-public-interest-northwestern-study-finds" rel="nofollow">https://news.northwestern.edu/stories/2022/07/government-fun...</a><p>* <a href="https://www.americanscientist.org/article/%E2%80%9Cwhy-are-we-funding-this%E2%80%9D" rel="nofollow">https://www.americanscientist.org/article/%E2%80%9Cwhy-are-w...</a></p>
]]></description><pubDate>Mon, 22 Jun 2026 20:25:28 +0000</pubDate><link>https://news.ycombinator.com/item?id=48635640</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=48635640</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48635640</guid></item><item><title><![CDATA[New comment by tech_ken in "Chevron signs 20-year power agreement with Microsoft for West Texas data center"]]></title><description><![CDATA[
<p>I can think of a few angles that might have pushed them towards gas, mainly (a) they wanted on-demand generation cap, (b) they didn't want to get into the batteries game at the volume they'd require, or (c) they didn't want to deal with securing the space needed to produce 2.6GW of solar. Also yeah they're definitely not price-sensitive, any of the hyperscalars is more than happy to pay extra to get exactly what they want.<p>edit: for example that EIA list of new solar projects you linked indicates that the largest battery installations going up in '26 are all ~500MW, and that there are only four of them (of that size). I think the energy intensity of a multi-GW datacenter is the main reason that they're not going for solar here.</p>
]]></description><pubDate>Mon, 22 Jun 2026 15:36:27 +0000</pubDate><link>https://news.ycombinator.com/item?id=48631706</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=48631706</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48631706</guid></item><item><title><![CDATA[New comment by tech_ken in "Can the stockmarket swallow Anthropic, SpaceX and OpenAI?"]]></title><description><![CDATA[
<p>> Starship will bill NASA 1/20th what SLS does<p>Is that before or after the program achieves profitability?</p>
]]></description><pubDate>Tue, 02 Jun 2026 16:23:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=48372381</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=48372381</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48372381</guid></item><item><title><![CDATA[New comment by tech_ken in "Private Equity Bought America's Essential Services"]]></title><description><![CDATA[
<p>Because the buyer is the one monopolizing industries and stripping them for parts</p>
]]></description><pubDate>Wed, 27 May 2026 13:47:34 +0000</pubDate><link>https://news.ycombinator.com/item?id=48294326</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=48294326</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48294326</guid></item><item><title><![CDATA[New comment by tech_ken in "Uber, Lyft drivers in Massachusetts form first US ride-share union"]]></title><description><![CDATA[
<p>> Selling labor for money is the exact same transaction as walking into Walmart and buying a banana.<p>In the same comment you say that the government has to help guarantee that Walmart won't exploit you in this banana purchase transaction by outlawing price gouging and monopolization. If labor is the same type of transaction, it stands to reason that certain types of employment can be exploitative, despite the 'voluntary nature' of the transaction. Is your issue with the 'Uber exploits their contractors' framing simply based in the fact that Uber has not broken any labor laws?<p>Since you're having trouble following me I'll give a quick summary here. You initially said:<p>> if all these drivers are getting horribly exploited why are they doing it?<p>My point is simply that this is crap reasoning: people voluntarily participate in exploitative interactions all the time if they lack genuine alternatives.</p>
]]></description><pubDate>Tue, 26 May 2026 22:02:44 +0000</pubDate><link>https://news.ycombinator.com/item?id=48286627</link><dc:creator>tech_ken</dc:creator><comments>https://news.ycombinator.com/item?id=48286627</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48286627</guid></item></channel></rss>