<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: gravypod</title><link>https://news.ycombinator.com/user?id=gravypod</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Sat, 13 Jun 2026 14:39:57 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=gravypod" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by gravypod in "Nvidia is proposing a beast of a CPU system for Windows PCs"]]></title><description><![CDATA[
<p>Having slower memory may not actually lead to lower memory bandwidth. The cuda cores can be broken up into compute complexes which larger blocks of memory directly attached to the cores. These could be filled with read operations from the bulk system memory. You can start executing and then page the next batch of data in while compute is working. For LLMs you don't have much random memory access, you can sequence your accesses in blocks.<p>If these chips become popular I am sure you will see LLM architectures taking advantage of the parallelism.</p>
]]></description><pubDate>Sat, 06 Jun 2026 21:32:28 +0000</pubDate><link>https://news.ycombinator.com/item?id=48429203</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48429203</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48429203</guid></item><item><title><![CDATA[New comment by gravypod in "New U.S. college grads now have higher unemployment than the average worker"]]></title><description><![CDATA[
<p>I've known a few college graduates who have come up in this market. From what I see, the common pattern is to try and get a position in your field for 3-10 months. Somewhere in that time range, they burn out. Then they apply for something field related for a few months. Then anything. Once they've exhausted all options they usually give up.<p>We will likely have a similar concept in our country as China's "lying flat" movement unless we make a big shift.</p>
]]></description><pubDate>Sat, 06 Jun 2026 21:30:08 +0000</pubDate><link>https://news.ycombinator.com/item?id=48429182</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48429182</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48429182</guid></item><item><title><![CDATA[New comment by gravypod in "Nvidia is proposing a beast of a CPU system for Windows PCs"]]></title><description><![CDATA[
<p>What is more important than core count is how the caching architecture is laid out. They could lay out those 6k cuda cores in a layout which provides much larger blocks of cache to smaller number of cores. That would increase the memory bandwidth which would be better for inference.</p>
]]></description><pubDate>Sat, 06 Jun 2026 16:36:05 +0000</pubDate><link>https://news.ycombinator.com/item?id=48426569</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48426569</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48426569</guid></item><item><title><![CDATA[New comment by gravypod in "Ask HN: What was your "oh shit" moment with GenAI?"]]></title><description><![CDATA[
<p>Owning and using guns doesn't cause cognitive decline in the same way I've observed steep cognitive changes in people I know who are taken in by LLMs.<p>I recently had a friend ask an LLM what fun things there are to do in a town we were visiting. It gave the most generic answer like "try local restaurants" and "there are bars" and stuff. There's not a lot of tourist information for this area so it was nonspecific.<p>This is someone with an advanced degree in a medical field and she thought this was amazing insight. I asked, how is this different from what you already knew and she stood there thinking for a bit and you could tell there was a cognitive dissonance uncovered. She was very concerned when thinking it over and realizing it wasn't something she was able to intuit.<p>A relative of mine a long time ago had a stroke and recovered. I hadn't seen that facial expression since trying to help my relative figure out how to sit in a chair again.<p>Basic cognitive functions lost easily, difficult to rebuild.</p>
]]></description><pubDate>Sat, 06 Jun 2026 16:20:38 +0000</pubDate><link>https://news.ycombinator.com/item?id=48426441</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48426441</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48426441</guid></item><item><title><![CDATA[New comment by gravypod in "Ask HN: What was your "oh shit" moment with GenAI?"]]></title><description><![CDATA[
<p>I fear what's going to happen with less tech literate managers who don't understand these cases.</p>
]]></description><pubDate>Sat, 06 Jun 2026 16:15:31 +0000</pubDate><link>https://news.ycombinator.com/item?id=48426398</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48426398</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48426398</guid></item><item><title><![CDATA[New comment by gravypod in "Did Claude increase bugs in rsync?"]]></title><description><![CDATA[
<p>> Also the amount of commits is suspicious. In the last two months, rsync had about as much commits as in the last two years before that.<p>I wonder if the data looks worse or better when not doing per-10commit and instead do per-commit.</p>
]]></description><pubDate>Fri, 05 Jun 2026 23:44:04 +0000</pubDate><link>https://news.ycombinator.com/item?id=48419835</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48419835</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48419835</guid></item><item><title><![CDATA[New comment by gravypod in "Did Claude increase bugs in rsync?"]]></title><description><![CDATA[
<p>This is a really cool post but I think one metric we may want to also look at is does using agentic coding tools in one domain impact your coding abilities in another domain? A lot of people I know have been talking about getting rusty on the fundamentals recently. This is not something I am particularly feeling as I do a mix of running agents in parallel and writing some code manually where it makes sense. But if people who have been prompt-only at work come home and work on rsync and are more "rusty" maybe that could also lead to more bugs?<p>This would be even harder to measure.</p>
]]></description><pubDate>Fri, 05 Jun 2026 23:37:33 +0000</pubDate><link>https://news.ycombinator.com/item?id=48419778</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48419778</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48419778</guid></item><item><title><![CDATA[New comment by gravypod in "Ask HN: What was your "oh shit" moment with GenAI?"]]></title><description><![CDATA[
<p>> CI/CD?<p>That is less useful when the changes are editing the tests but we don't know if a human has validated the assertions.<p>> My non-techie friends send me screenshots of ChatGPT. I guess that’s a modern micro aggression?<p>I think the concern I have is explicitly not the sending the chat logs. I think it's this flow:<p>1. Ask a question<p>2. Get an answer from a team member.<p>3. I don't like the answer and instead of discussing I am going to go to Claude and ask the same question.<p>4. Copy/paste the answer into chat without seeing if it includes novel information.<p>In one case the engineer was asking which model to select in the agent framework we are using. I gave an answer and provided a list of reasons. They did not like this answer and asked Claude which gave the same answer.<p>The answer was something inherently obvious and that anyone should be able to derive from first principals.</p>
]]></description><pubDate>Fri, 05 Jun 2026 22:00:32 +0000</pubDate><link>https://news.ycombinator.com/item?id=48418932</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48418932</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48418932</guid></item><item><title><![CDATA[New comment by gravypod in "Ask HN: What was your "oh shit" moment with GenAI?"]]></title><description><![CDATA[
<p>I did read the other replies. I don't think my comment is that LLMs are bad. I use LLMs and agents for work. I think my "oh shit" moment is the dynamic that giving someone LLMs amplifies their impact (positive or negative).<p>For example, some people give kids tiny go karts and that's acceptable because the damage they can do with a very tiny battery powered 4 wheeler is minimal. We now live in a world where everyone has access to a tank and can plow over everything.<p>I think LLMs will increase anti-social behavior.</p>
]]></description><pubDate>Fri, 05 Jun 2026 21:57:08 +0000</pubDate><link>https://news.ycombinator.com/item?id=48418896</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48418896</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48418896</guid></item><item><title><![CDATA[New comment by gravypod in "Ask HN: What was your "oh shit" moment with GenAI?"]]></title><description><![CDATA[
<p>I work with someone who is very AI-forward, high confidence, and very low execution. He has started sending me large PRs of AI slop that he assured me doesn't need to be reviewed. I quickly find many minor issues from an initial pass of one of the reviews. He gets mad at the team for slowing him down.<p>He also will paste chat logs with Claude into our team chat. Often Claude will say the same thing I told him but he either doesn't remember or doesn't trust human engineers now.<p>He has spent months working on agent skills and prompring.<p>He has not landed anything in 3mo, and has landed nothing useful in ~1 year.<p>This will be the rest of my career. Working with people in ai psychosis and trying to stay productive.</p>
]]></description><pubDate>Fri, 05 Jun 2026 20:35:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=48417872</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48417872</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48417872</guid></item><item><title><![CDATA[New comment by gravypod in "Ask HN: What was your "oh shit" moment with GenAI?"]]></title><description><![CDATA[
<p>What mod did you build?</p>
]]></description><pubDate>Fri, 05 Jun 2026 20:32:26 +0000</pubDate><link>https://news.ycombinator.com/item?id=48417813</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48417813</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48417813</guid></item><item><title><![CDATA[New comment by gravypod in "Uber's $1,500/month AI limit is a useful signal for AI tool pricing"]]></title><description><![CDATA[
<p>If you had a business task to complete that was only possible with ai and it cost you >$1500/month of work, how long would you have to delay the task so that it's cheaper long run to buy hardware and do local models?<p>$1,500/mo * 14 months = $21,000.<p>If local models are 14mo behind as many in HN say it may be profitable to just wait. Maybe just spend a few hundred dollars of your tokens and buy hardware piece by piece.</p>
]]></description><pubDate>Wed, 03 Jun 2026 22:45:27 +0000</pubDate><link>https://news.ycombinator.com/item?id=48391147</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48391147</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48391147</guid></item><item><title><![CDATA[New comment by gravypod in "Can we have the day off?"]]></title><description><![CDATA[
<p>Anything commissions based might fit the bill?</p>
]]></description><pubDate>Thu, 28 May 2026 03:09:29 +0000</pubDate><link>https://news.ycombinator.com/item?id=48303961</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48303961</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48303961</guid></item><item><title><![CDATA[New comment by gravypod in "The AI bubble isn't like the internet bubble"]]></title><description><![CDATA[
<p>Just a note, P40 came out at $5700 in 2016 dollars. In 2026 dollars that is $8000 (wow!). If you bought 100k today, assuming a 1% failure rate per year your $800M investment can be traded in for about $30M.<p>I think it is reasonable to assume a similar depreciation in GPUs.<p>Meaning you'd need to have made more than (800M - 30M) * (1 + income tax rate) + (power + maintenance).<p>Some say the margines on inference are already there for new GPUs but they are right margines.</p>
]]></description><pubDate>Tue, 26 May 2026 17:36:44 +0000</pubDate><link>https://news.ycombinator.com/item?id=48282973</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48282973</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48282973</guid></item><item><title><![CDATA[New comment by gravypod in "The AI bubble isn't like the internet bubble"]]></title><description><![CDATA[
<p>It's, from my understanding, a little bit of both. There's a failure rate of GPUs and fans. There's also changing in standards like PCIe and software stacks.<p>LLM inference is mainly memory bandwidth constrained so I think it's highly likely that a company will create silicon with just an insane number of memory chips and less compute. These ASICs will probably do the same thing the crypto ASICs did.<p>If we look back 1 decade, no one uses a GTX 950 for anything.</p>
]]></description><pubDate>Tue, 26 May 2026 11:27:48 +0000</pubDate><link>https://news.ycombinator.com/item?id=48278232</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48278232</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48278232</guid></item><item><title><![CDATA[New comment by gravypod in "GitHub is investigating unauthorized access to their internal repositories"]]></title><description><![CDATA[
<p>Where is this list from?</p>
]]></description><pubDate>Wed, 20 May 2026 11:27:42 +0000</pubDate><link>https://news.ycombinator.com/item?id=48206039</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48206039</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48206039</guid></item><item><title><![CDATA[New comment by gravypod in "Bun's experimental Rust rewrite hits 99.8% test compatibility on Linux x64 glibc"]]></title><description><![CDATA[
<p>We did allow people to send us letters with QSO logs on paper.</p>
]]></description><pubDate>Mon, 11 May 2026 20:05:26 +0000</pubDate><link>https://news.ycombinator.com/item?id=48099963</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48099963</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48099963</guid></item><item><title><![CDATA[New comment by gravypod in "Getting arrested in Japan"]]></title><description><![CDATA[
<p>Warning: This is not medical advice, I am a nerd on the internet not a doctor<p>For what it is worth different countries have vastly different recommendations for HBP and these drugs. I recommend discussing with the pharmacists in your country.<p>In the US I have been told it's a strict "never", in Ireland I was told that it wouldn't have a measurable effect on blood pressure. I've also measured my personal blood pressure (pre-hypertension to stage 1) and have not been able to measure a difference in blood pressure.</p>
]]></description><pubDate>Sun, 10 May 2026 05:20:19 +0000</pubDate><link>https://news.ycombinator.com/item?id=48081221</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48081221</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48081221</guid></item><item><title><![CDATA[New comment by gravypod in "Getting arrested in Japan"]]></title><description><![CDATA[
<p>> with small amounts leading to being detained for 23 days (like in this blog post)<p>This seems ultimately like a very bad sales pitch for the tourism industry in Japan. I had thought I wanted to go to Japan but if I can accidentally, without malice, be thrown in a prison for 20 days that seems like a bad system.<p>I can't imagine the international relations of the ruling classes of various countries to the UAE would be trending in a positive direction if they arrested and punished people for walking off a plane with airplane bottles of alcohol.</p>
]]></description><pubDate>Sun, 10 May 2026 05:13:54 +0000</pubDate><link>https://news.ycombinator.com/item?id=48081181</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48081181</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48081181</guid></item><item><title><![CDATA[New comment by gravypod in "Bun's experimental Rust rewrite hits 99.8% test compatibility on Linux x64 glibc"]]></title><description><![CDATA[
<p>People *did* write down these logs, manually, and submit them.</p>
]]></description><pubDate>Sun, 10 May 2026 05:08:50 +0000</pubDate><link>https://news.ycombinator.com/item?id=48081166</link><dc:creator>gravypod</dc:creator><comments>https://news.ycombinator.com/item?id=48081166</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48081166</guid></item></channel></rss>