<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: stanfordkid</title><link>https://news.ycombinator.com/user?id=stanfordkid</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Sat, 13 Jun 2026 10:03:24 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=stanfordkid" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by stanfordkid in "An AI agent deleted our production database. The agent's confession is below"]]></title><description><![CDATA[
<p>I don't think you can really blame AI agents for this. While I agree the user was using AI irresponsibly, some of the blame does go to Railway for making an API key that allows for all operations to happen from a single key without giving clear warnings on privileges. Clearly this user was shooting from the hip and quickly pasted whatever key they got from Railway into a file somewhere so there is some blame there, but any service that handles hosting infrastructure should provide clear UX warning to users regarding the scoping of it's credentials.</p>
]]></description><pubDate>Mon, 27 Apr 2026 16:53:39 +0000</pubDate><link>https://news.ycombinator.com/item?id=47924116</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=47924116</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47924116</guid></item><item><title><![CDATA[New comment by stanfordkid in "GPT-5.5: Mythos-Like Hacking, Open to All"]]></title><description><![CDATA[
<p>No, they are able to detect errors when pointed at them but they have a lot of false positives... making them functionally useless for a large unknown codebase. They also can't build and run an exploit post-identification. Mythos can find vulnerabilities (purportedly) and actually validate them by building and running exploits. This makes it functional and usable for hacking.</p>
]]></description><pubDate>Thu, 23 Apr 2026 22:40:33 +0000</pubDate><link>https://news.ycombinator.com/item?id=47883145</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=47883145</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47883145</guid></item><item><title><![CDATA[New comment by stanfordkid in "Anthropic takes $5B from Amazon and pledges $100B in cloud spending in return"]]></title><description><![CDATA[
<p>They are bringing in $30B in revenue with 3X YoY growth. Why do you think it is a "jig"? I do think the US economy could implode, but thats because of war and wealth inequality in the midst of hyper-inflation. AI models aren't very useful when you have penniless consumers that can't buy the products they help build. All this is to say: the models are valuable, the companies building and providing them are very valuable.<p>The biggest risk to AI companies IMO is further optimization and distillation of the capabilities into smaller and more efficient models. The moat these companies have right now is that higher intelligence requires more specialized and expensive compute. If you can do that for cheap then it kind of negates their business model. Everything is moving fast, we also yet to see world models/embodied AI and how that impacts thing. I think we've reached the peak with regards to capabilities of pure text trained LLMs.</p>
]]></description><pubDate>Wed, 22 Apr 2026 00:02:55 +0000</pubDate><link>https://news.ycombinator.com/item?id=47856640</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=47856640</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47856640</guid></item><item><title><![CDATA[New comment by stanfordkid in "Google has the same AI adoption curve as John Deere"]]></title><description><![CDATA[
<p>They basically wrote the equivalent to Claude Code and launched it as a product... how does their adoption curve lag behind John Deere?</p>
]]></description><pubDate>Mon, 13 Apr 2026 20:23:46 +0000</pubDate><link>https://news.ycombinator.com/item?id=47757344</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=47757344</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47757344</guid></item><item><title><![CDATA[New comment by stanfordkid in "Exploiting the most prominent AI agent benchmarks"]]></title><description><![CDATA[
<p>I don't find this paper very compelling. Obviously it would be fraud if the code generated simply escaped the harness vs solving the actual problem. I agree that theoretically models could learn to do that, and it is important to highlight, but my sense is that those entities reporting the benchmark scores would have an obligation to observe this behavior and re-consider the metrics they report. It is a bit like saying it's possible to cheat in football because the balls are deflatable. It matters, and some have done it, but it doesn't mean widespread cheating is taking place. The paper takes the tone that there is already a lot of cheating happening which I do not think is the case.</p>
]]></description><pubDate>Mon, 13 Apr 2026 01:54:43 +0000</pubDate><link>https://news.ycombinator.com/item?id=47746639</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=47746639</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47746639</guid></item><item><title><![CDATA[New comment by stanfordkid in "Small Models Are Smart Enough"]]></title><description><![CDATA[
<p>Mind blowing! I've had this intuition for a while too, but he really gets into the heart of it.</p>
]]></description><pubDate>Fri, 10 Apr 2026 20:55:24 +0000</pubDate><link>https://news.ycombinator.com/item?id=47723547</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=47723547</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47723547</guid></item><item><title><![CDATA[New comment by stanfordkid in "Ask HN: Why Databases Instead of Filesystem?"]]></title><description><![CDATA[
<p>I think there are a lot of good answers here, but it really comes down to the type of content being stored and access patterns.<p>A database is a data structure with (generally) many small items that need to be precisely updated, read and manipulated.<p>A lot of files don't necessarily have this access pattern (for instance rendering a large video file) ... a filesystem has a generic access pattern and is a lower level primitive than a database.<p>For this same reason you even have different kinds of database for different types of access patterns and data types (e.g Elasticsearch for full text search, MongoDB for JSON, Postgres for SQL)<p>Filesystem is generic and low-level, database is a higher order abstraction.</p>
]]></description><pubDate>Fri, 10 Apr 2026 17:42:36 +0000</pubDate><link>https://news.ycombinator.com/item?id=47721374</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=47721374</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47721374</guid></item><item><title><![CDATA[New comment by stanfordkid in "US and Iran agree to provisional ceasefire"]]></title><description><![CDATA[
<p>Not only is it no better, it is significantly worse.</p>
]]></description><pubDate>Wed, 08 Apr 2026 20:15:28 +0000</pubDate><link>https://news.ycombinator.com/item?id=47695692</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=47695692</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47695692</guid></item><item><title><![CDATA[New comment by stanfordkid in "US fired 1k JASSM cruise missiles in 37 days. Lockheed makes 396 per year"]]></title><description><![CDATA[
<p>What's your source for this? There isn't really a lot of credible, publicly available information on what you're saying... just anecdotes. In the India v. Pakistan conflict recently a French produced Indian Rafale was downed via a Chinese long range air-to-air missle (PL-15) from a a Chinese produced J-10 jet. Even if they don't have the same hit rate, you can buy 10x for the same price.<p>Source: <a href="https://www.reuters.com/business/aerospace-defense/how-pakistan-shot-down-indias-cutting-edge-fighter-using-chinese-gear-2025-08-02/" rel="nofollow">https://www.reuters.com/business/aerospace-defense/how-pakis...</a></p>
]]></description><pubDate>Wed, 08 Apr 2026 18:13:36 +0000</pubDate><link>https://news.ycombinator.com/item?id=47694100</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=47694100</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47694100</guid></item><item><title><![CDATA[New comment by stanfordkid in "US and Iran agree to provisional ceasefire"]]></title><description><![CDATA[
<p>No one wants to liberate Iran. Israel just wants to continue committing genocide and apartheid without any opposition. Iran arms Hezbollah and Hamas, the main forms of Palestinian resistance. The whole point of this operation is to decimate those groups so ethnic cleansing can continue without any resistance. Israel could care less about the Irani people.<p>You are very naive if you think the IRGC truly killed 10's of thousands of it's own people. Israel openly talks about Mossad organizing and supporting the coup, and good old Donny has admitted they have given weapons to organized resistance.<p>I estimate that many of the death numbers come from armed resistance being killed by the IRGC, not ordinary peaceful protestors. I also think armed resistance killed many Irani citizens. There is obviously fog of war here. The thousands of deaths were likely inflated and obfuscated.<p>Look at the coups we have backed in the middle east (including formerly in Iran which is what originally led to the Islamic revolution) -- and you will see a pattern. Both US and Israel provide material support to groups like ISIS or actors like Bin Laden. An Al-Qaeda fighter is literally the head of Syria now thanks to Israel.<p>I don't love Hamas, IRGC or Hezbollah, I don't like their ideology. But it is myopic to think they exist in a vaccum.</p>
]]></description><pubDate>Wed, 08 Apr 2026 03:26:29 +0000</pubDate><link>https://news.ycombinator.com/item?id=47684722</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=47684722</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47684722</guid></item><item><title><![CDATA[New comment by stanfordkid in "Pretraining Language Models via Neural Cellular Automata"]]></title><description><![CDATA[
<p>It's better than using randomly initialized weights. It's more of a theoretical exercise to explore biology. When an infant is born maybe the visual cortex already has some notion of edge detectors etc. through a system such as this one despite never having really opened it's eyes.</p>
]]></description><pubDate>Thu, 26 Mar 2026 16:37:34 +0000</pubDate><link>https://news.ycombinator.com/item?id=47532631</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=47532631</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47532631</guid></item><item><title><![CDATA[New comment by stanfordkid in "Pretraining Language Models via Neural Cellular Automata"]]></title><description><![CDATA[
<p>I did a similar project but using 3D fractals I found on shadertoy feeding into ViTs. They are extremely simple iterative functions that produce a ton of scene like complexity.<p>I have a pet theory that the visual cortex when developing is linked to some kind of mechanism such as this. You just need proteins that create some sort of resonating signal that feed into the neurons as they grow (obviously this is hand-wavy) but similar feedback loops guide nervous system growth in Zebra fish for example.</p>
]]></description><pubDate>Thu, 19 Mar 2026 14:30:56 +0000</pubDate><link>https://news.ycombinator.com/item?id=47440180</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=47440180</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47440180</guid></item><item><title><![CDATA[New comment by stanfordkid in "Yann LeCun raises $1B to build AI that understands the physical world"]]></title><description><![CDATA[
<p>It's pretty simple... the word circle and what you can correlate to it via english language description has somewhat less to do with reality than a physical 3D model of a circle and what it would do in an environment. You can't just add more linguistic description via training data to change that. It doesn't really matter that you can keep back propagating because what you are back propagating over is fundamentally and qualitatively less rich.</p>
]]></description><pubDate>Wed, 11 Mar 2026 01:54:24 +0000</pubDate><link>https://news.ycombinator.com/item?id=47330979</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=47330979</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47330979</guid></item><item><title><![CDATA[New comment by stanfordkid in "Show HN: µJS, a 5KB alternative to Htmx and Turbo with zero dependencies"]]></title><description><![CDATA[
<p>I'm curious to understand, why would you build your website this way vs. say jQuery. I've never really understood the HTMX ecosystem. Is this just to avoid javascript and replace that with html pages, id's and attributes? It feels like the DOM is a very clear abstraction and scripting is a more powerful way to manipulate it. What do people like or prefer about this approach and paradigm?</p>
]]></description><pubDate>Sat, 07 Mar 2026 23:23:47 +0000</pubDate><link>https://news.ycombinator.com/item?id=47292468</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=47292468</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47292468</guid></item><item><title><![CDATA[New comment by stanfordkid in "Shannon Got AI This Far. Kolmogorov Shows Where It Stops"]]></title><description><![CDATA[
<p>I'm not sure I totally buy the "no plasticity" argument. If you are allowed to write to context in an agentic fashion, certainly the LLM can record intermediate answers, go back and re-rank it's memory. The "plasticity" is in the form of data that can be looked up and referenced as a shortcut. I would think this forms a Turing complete system so theoretically it can represent pretty much anything.</p>
]]></description><pubDate>Sat, 07 Mar 2026 22:38:36 +0000</pubDate><link>https://news.ycombinator.com/item?id=47292158</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=47292158</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47292158</guid></item><item><title><![CDATA[New comment by stanfordkid in "We do not think Anthropic should be designated as a supply chain risk"]]></title><description><![CDATA[
<p>Isn’t this kind of all bullshit. Like Anthropic licenses so many of its models through Bedrock. If the DoD has a contract with Amazon they can just use them.</p>
]]></description><pubDate>Sun, 01 Mar 2026 19:32:28 +0000</pubDate><link>https://news.ycombinator.com/item?id=47209855</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=47209855</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47209855</guid></item><item><title><![CDATA[New comment by stanfordkid in "Notepad++ supply chain attack breakdown"]]></title><description><![CDATA[
<p>Shouldn't public signature of the hash of the exe file from a known key before execution fix this??? What am I missing?</p>
]]></description><pubDate>Wed, 04 Feb 2026 22:02:49 +0000</pubDate><link>https://news.ycombinator.com/item?id=46892461</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=46892461</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46892461</guid></item><item><title><![CDATA[New comment by stanfordkid in "Poison Fountain"]]></title><description><![CDATA[
<p>Just look at the domains. Obviously social media will get harder to do this with, maybe that's okay though. I think a simple criterion can be used: could the pre-trained LLM have come up with this itself? If so it probably doesn't have training value.</p>
]]></description><pubDate>Tue, 13 Jan 2026 18:15:22 +0000</pubDate><link>https://news.ycombinator.com/item?id=46605212</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=46605212</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46605212</guid></item><item><title><![CDATA[New comment by stanfordkid in "HP Reveals Keyboard Computer with Ryzen AI Chip"]]></title><description><![CDATA[
<p>Cool to see some innovation from HP, this is actually unique and fills a niche of going from desk to desk (home->office), without needing mobile. Much better than ripping off the latest aluminum MacBook designs.</p>
]]></description><pubDate>Tue, 13 Jan 2026 18:12:55 +0000</pubDate><link>https://news.ycombinator.com/item?id=46605167</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=46605167</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46605167</guid></item><item><title><![CDATA[New comment by stanfordkid in "Poison Fountain"]]></title><description><![CDATA[
<p>I don't see how you get around LLMs scraping data without also stopping humans from retrieving valid data.<p>If you are NYTimes and publish poisoned data to scrapers, the only thing the scraper needs is one valid human subscription where they run a VM + automated Chrome, OCR and tokenize the valid data then compare that to the scraped results. It's pretty much trivial to do. At Anthropic/Google/OpenAI scale they can easily buy VMs in data centers spread all over the world with IP shuffling. There is no way to tell who is accessing the data.</p>
]]></description><pubDate>Sun, 11 Jan 2026 18:58:52 +0000</pubDate><link>https://news.ycombinator.com/item?id=46578609</link><dc:creator>stanfordkid</dc:creator><comments>https://news.ycombinator.com/item?id=46578609</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46578609</guid></item></channel></rss>