<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: latentsea</title><link>https://news.ycombinator.com/user?id=latentsea</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Thu, 19 Nov 2026 05:34:38 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=latentsea" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by latentsea in "Genie 4 is now available"]]></title><description><![CDATA[
<p>Having tried it a little, yes.</p>
]]></description><pubDate>Thu, 19 Nov 2026 00:49:46 +0000</pubDate><link>https://news.ycombinator.com/item?id=50554569</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50554569</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50554569</guid></item><item><title><![CDATA[New comment by latentsea in "What is the best harness for each model?"]]></title><description><![CDATA[
<p>This is why for open-weights models at least harness aware post-training is becoming a thing. Also closed source harnesses that aren't fully editable are basically an anti-pattern. The labs should just hurry up and productionize self-evolving harnesses, and have the models internalize harness engineering, so they can optimise the harness for the given task. These are solvable problems.</p>
]]></description><pubDate>Tue, 17 Nov 2026 16:08:49 +0000</pubDate><link>https://news.ycombinator.com/item?id=50533426</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50533426</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50533426</guid></item><item><title><![CDATA[New comment by latentsea in "My Quest to push native Three.js quality match AAA Game Engines"]]></title><description><![CDATA[
<p>I wish we would. It's like we keep prompting one another in an endless circle, and we can't not respond to the prompts.</p>
]]></description><pubDate>Tue, 17 Nov 2026 03:00:46 +0000</pubDate><link>https://news.ycombinator.com/item?id=50526849</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50526849</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50526849</guid></item><item><title><![CDATA[New comment by latentsea in "OpenCode Browse"]]></title><description><![CDATA[
<p>Just use the spare R9700 you have lying around and run locally.</p>
]]></description><pubDate>Tue, 17 Nov 2026 01:14:59 +0000</pubDate><link>https://news.ycombinator.com/item?id=50526121</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50526121</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50526121</guid></item><item><title><![CDATA[New comment by latentsea in "My Quest to push native Three.js quality match AAA Game Engines"]]></title><description><![CDATA[
<p>Reading the same criticism of the same comment I feel tired boss.</p>
]]></description><pubDate>Tue, 17 Nov 2026 00:28:24 +0000</pubDate><link>https://news.ycombinator.com/item?id=50525718</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50525718</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50525718</guid></item><item><title><![CDATA[New comment by latentsea in "Microsoft 365 outage leads to chlorine gas poisoning of 15 hotel guests"]]></title><description><![CDATA[
<p>PagerDuty.</p>
]]></description><pubDate>Mon, 16 Nov 2026 16:19:50 +0000</pubDate><link>https://news.ycombinator.com/item?id=50519304</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50519304</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50519304</guid></item><item><title><![CDATA[New comment by latentsea in "Microsoft 365 outage leads to chlorine gas poisoning of 15 hotel guests"]]></title><description><![CDATA[
<p>>More properly said, you may cause panic. People genuinely don't know how to respond when they hear an alarm in an unfamiliar place.<p>It's a well known irony that every system intended to prevent accidents eventually itself causes one.</p>
]]></description><pubDate>Mon, 16 Nov 2026 16:06:21 +0000</pubDate><link>https://news.ycombinator.com/item?id=50519139</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50519139</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50519139</guid></item><item><title><![CDATA[New comment by latentsea in "Qwen4-Max"]]></title><description><![CDATA[
<p>Then perhaps I should point out something different. Namely, I suspect this will thoroughly confuse the machines and instead of optimizing the universe to generate paperclips they will instead optimize it to generate pelicans.</p>
]]></description><pubDate>Mon, 16 Nov 2026 15:13:07 +0000</pubDate><link>https://news.ycombinator.com/item?id=50518446</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50518446</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50518446</guid></item><item><title><![CDATA[New comment by latentsea in "Qwen4-Max"]]></title><description><![CDATA[
<p>It's part of culture at this point.</p>
]]></description><pubDate>Mon, 16 Nov 2026 15:07:33 +0000</pubDate><link>https://news.ycombinator.com/item?id=50518391</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50518391</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50518391</guid></item><item><title><![CDATA[New comment by latentsea in "Found: A protein that drives ageing without unleashing inflammation"]]></title><description><![CDATA[
<p>I'd give anything to be 14 again.</p>
]]></description><pubDate>Sun, 15 Nov 2026 16:25:45 +0000</pubDate><link>https://news.ycombinator.com/item?id=50508154</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50508154</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50508154</guid></item><item><title><![CDATA[New comment by latentsea in "Found: A protein that drives ageing without unleashing inflammation"]]></title><description><![CDATA[
<p>Islands have been known to be sources of youth.</p>
]]></description><pubDate>Sun, 15 Nov 2026 16:18:32 +0000</pubDate><link>https://news.ycombinator.com/item?id=50508097</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50508097</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50508097</guid></item><item><title><![CDATA[New comment by latentsea in "Found: A protein that drives ageing without unleashing inflammation"]]></title><description><![CDATA[
<p>Honestly it gets kind of boring and mentally painful. You either die a hero or living long enough to see yourself become the villain. If you live forever, you're guaranteed to become the villain.</p>
]]></description><pubDate>Sun, 15 Nov 2026 15:52:22 +0000</pubDate><link>https://news.ycombinator.com/item?id=50507843</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50507843</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50507843</guid></item><item><title><![CDATA[New comment by latentsea in "Jonathan Blow – If AI solved coding, why is my software still the same? [video]"]]></title><description><![CDATA[
<p>I disagree that architecture and taste aren't trainable in the exact same way. Sure it's getting better at basically everything across the board, so I take your point that it's not all down to RLVR. That being said for anything you can RLVR, you can improve directly and quite fast. So can you figure out ways to make things verifiable to some degree that you previously considered unverifiable? That is actually an interesting question.<p>How do we go about verifying an architecture is good in the first place? For the most part we do policy rollouts and check. Guess and check has been the strat for as long as I can remember. We then swap war stories with one another. Hence, I think we can get AI to just do the same. It'd be difficult to put together tasks and environments by hand, so I've been wondering if it might be possible to leverage those "simulator" video games as pre-existing strong verifiers of a certain narrow domain that you could then develop software against to kind of bootstrap the process if you will. Then if you can figure out a way to introduce changing requirements over time, you could then potentially just hillclimb on finding architectures which minimize changes to core components even as new requirements get supported etc.<p>I think most people would put that firmly in the too hard basket, but I wouldn't. That's the sort of thing that intrigues me enough to try. Sadly, I don't work at a frontier lab though, so... I just play with things at home.</p>
]]></description><pubDate>Sun, 15 Nov 2026 08:35:36 +0000</pubDate><link>https://news.ycombinator.com/item?id=50505142</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50505142</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50505142</guid></item><item><title><![CDATA[New comment by latentsea in "Anthropic upgrades Claude Code Projects to run parallel AI workers"]]></title><description><![CDATA[
<p>"Solve The Riemann Hypothesis"</p>
]]></description><pubDate>Sat, 14 Nov 2026 05:37:42 +0000</pubDate><link>https://news.ycombinator.com/item?id=50494330</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50494330</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50494330</guid></item><item><title><![CDATA[New comment by latentsea in "Anthropic upgrades Claude Code Projects to run parallel AI workers"]]></title><description><![CDATA[
<p>> What future do you want?<p>This is why in my private life I run only open weights models and procure all the hardware I can get my hands on. I don't want to be completely beholden to this company and its whims.</p>
]]></description><pubDate>Sat, 14 Nov 2026 05:36:31 +0000</pubDate><link>https://news.ycombinator.com/item?id=50494326</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50494326</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50494326</guid></item><item><title><![CDATA[New comment by latentsea in "Anthropic upgrades Claude Code Projects to run parallel AI workers"]]></title><description><![CDATA[
<p>Why can't it just be aware of usage limits and act accordingly?</p>
]]></description><pubDate>Sat, 14 Nov 2026 05:32:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=50494313</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50494313</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50494313</guid></item><item><title><![CDATA[New comment by latentsea in "Anthropic upgrades Claude Code Projects to run parallel AI workers"]]></title><description><![CDATA[
<p>What if we don't want it in the cloud? What if we want it to orchestrate what it can locally?</p>
]]></description><pubDate>Sat, 14 Nov 2026 05:30:16 +0000</pubDate><link>https://news.ycombinator.com/item?id=50494295</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50494295</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50494295</guid></item><item><title><![CDATA[New comment by latentsea in "Anthropic upgrades Claude Code Projects to run parallel AI workers"]]></title><description><![CDATA[
<p>Does it work to ask them to pace themselves so they can work at a pace that stays within the usage limit windows, so that it can just keep going but perhaps slower?</p>
]]></description><pubDate>Sat, 14 Nov 2026 05:13:49 +0000</pubDate><link>https://news.ycombinator.com/item?id=50494200</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50494200</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50494200</guid></item><item><title><![CDATA[New comment by latentsea in "Jonathan Blow – If AI solved coding, why is my software still the same? [video]"]]></title><description><![CDATA[
<p>A lot of the performance comes down to how you're holding it. I think that speaks to current limits and failure modes of the tech, but the results are definitely better the more you understand the failure modes and engineer them out.<p>I think humans are by and large just not great at this, especially in groups. I also think over the course of 2027 frontier labs will work on RL'ing harness engineering skills of the models themselves, so these systems treat engineering the environment in which they're engineering as a first class citizen since that's what makes a big difference.<p>Already they're doing RLCR to train swarms, so coordination is getting very good and this feeds into them being able to loop pretty quickly on code reviews and feedback without human input.<p>If your goal is verifiable and sufficiently well defined, and you have a strong deterministic set of constraints in place, then they can write a lot of code that's pretty usable.<p>From Nov 2025 to Nov 2026 we went from "we've only just gotten agents to work" to "we've only just gotten agents swarms to work" in publically available models. I'm rather interested to see who's still saying what in November 2027.</p>
]]></description><pubDate>Sat, 14 Nov 2026 04:59:14 +0000</pubDate><link>https://news.ycombinator.com/item?id=50494107</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50494107</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50494107</guid></item><item><title><![CDATA[New comment by latentsea in "Jonathan Blow – If AI solved coding, why is my software still the same? [video]"]]></title><description><![CDATA[
<p>I agree with this sentiment, though it's a slightly odd future. I've been running custom forks of various open source software that I just customise through pure vibe coding testing it into a working solution for some custom thing I want to do. But it seems inefficient everyone doing that.</p>
]]></description><pubDate>Sat, 14 Nov 2026 04:43:48 +0000</pubDate><link>https://news.ycombinator.com/item?id=50494018</link><dc:creator>latentsea</dc:creator><comments>https://news.ycombinator.com/item?id=50494018</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=50494018</guid></item></channel></rss>