<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: eddyzh</title><link>https://news.ycombinator.com/user?id=eddyzh</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Sun, 21 Jun 2026 14:36:40 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=eddyzh" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by eddyzh in "Claude Fable is relentlessly proactive"]]></title><description><![CDATA[
<p>For that test you have to compare letting a fresh agent (subagent) or the same model do the same review.<p>The fact that a review helps does not prove the model choice for the review.<p>You reviewing your own writing helps too!</p>
]]></description><pubDate>Fri, 12 Jun 2026 16:23:59 +0000</pubDate><link>https://news.ycombinator.com/item?id=48506097</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=48506097</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48506097</guid></item><item><title><![CDATA[New comment by eddyzh in "Long-running Agents (Summary of the current state on many fronts)"]]></title><description><![CDATA[
<p>Added the () part since this title has been used a lot in other articles.</p>
]]></description><pubDate>Wed, 29 Apr 2026 19:53:53 +0000</pubDate><link>https://news.ycombinator.com/item?id=47953623</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=47953623</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47953623</guid></item><item><title><![CDATA[Long-running Agents (Summary of the current state on many fronts)]]></title><description><![CDATA[
<p>Article URL: <a href="https://addyosmani.com/blog/long-running-agents/">https://addyosmani.com/blog/long-running-agents/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=47953622">https://news.ycombinator.com/item?id=47953622</a></p>
<p>Points: 2</p>
<p># Comments: 1</p>
]]></description><pubDate>Wed, 29 Apr 2026 19:53:53 +0000</pubDate><link>https://addyosmani.com/blog/long-running-agents/</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=47953622</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47953622</guid></item><item><title><![CDATA[New comment by eddyzh in "Ask HN: Can you tell the difference between Claude Sonnet and Opus?"]]></title><description><![CDATA[
<p>Exactly this.<p>This may be worth the discount. Or not if your time and attention is worth (quite) a lot.</p>
]]></description><pubDate>Mon, 27 Apr 2026 12:18:31 +0000</pubDate><link>https://news.ycombinator.com/item?id=47920611</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=47920611</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47920611</guid></item><item><title><![CDATA[New comment by eddyzh in "Ask HN: Can you tell the difference between Claude Sonnet and Opus?"]]></title><description><![CDATA[
<p>At work I use opus max Fast
It hardy ever fails for no reason even if I forget to give it all the right context.
At home i run sonnet, and it does not get what I meant or expected 20-35% of the time.
Due to the enormous difference in cost, depending on the value of your time (hourly rate) that might be a nett benefit.<p>Sonnet being faster alone would not be worth the failure rate for me.<p>At home i just not want to pay more than 20 bucks for incidental projects.<p>And opus max would just consume my tokens in one round.</p>
]]></description><pubDate>Mon, 27 Apr 2026 12:09:14 +0000</pubDate><link>https://news.ycombinator.com/item?id=47920518</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=47920518</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47920518</guid></item><item><title><![CDATA[New comment by eddyzh in "Flickr: The first and last great photo platform"]]></title><description><![CDATA[
<p>Thank you. This is great to hear!</p>
]]></description><pubDate>Sun, 26 Apr 2026 07:38:54 +0000</pubDate><link>https://news.ycombinator.com/item?id=47908236</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=47908236</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47908236</guid></item><item><title><![CDATA[New comment by eddyzh in "Graphs that explain the state of AI in 2026"]]></title><description><![CDATA[
<p>While chatGPT was not out then, the ML that drives robotics was acting by then very much.</p>
]]></description><pubDate>Sun, 19 Apr 2026 08:40:39 +0000</pubDate><link>https://news.ycombinator.com/item?id=47822772</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=47822772</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47822772</guid></item><item><title><![CDATA[New comment by eddyzh in "Chuck Norris has died"]]></title><description><![CDATA[
<p>In had one app like that from Cydia 
Loved it.</p>
]]></description><pubDate>Fri, 20 Mar 2026 17:34:36 +0000</pubDate><link>https://news.ycombinator.com/item?id=47457886</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=47457886</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47457886</guid></item><item><title><![CDATA[New comment by eddyzh in "Claude Code deletes developers' production setup, including database"]]></title><description><![CDATA[
<p>Original artikel 
<a href="https://open.substack.com/pub/alexeyondata/p/how-i-dropped-our-production-database" rel="nofollow">https://open.substack.com/pub/alexeyondata/p/how-i-dropped-o...</a></p>
]]></description><pubDate>Sat, 07 Mar 2026 19:18:59 +0000</pubDate><link>https://news.ycombinator.com/item?id=47290591</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=47290591</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47290591</guid></item><item><title><![CDATA[The Mythical Agent-Month]]></title><description><![CDATA[
<p>Article URL: <a href="https://wesmckinney.com/blog/mythical-agent-month/">https://wesmckinney.com/blog/mythical-agent-month/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=47075544">https://news.ycombinator.com/item?id=47075544</a></p>
<p>Points: 4</p>
<p># Comments: 0</p>
]]></description><pubDate>Thu, 19 Feb 2026 16:30:16 +0000</pubDate><link>https://wesmckinney.com/blog/mythical-agent-month/</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=47075544</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47075544</guid></item><item><title><![CDATA[New comment by eddyzh in "Local AI is driving the biggest change in laptops in decades"]]></title><description><![CDATA[
<p>Would that be unified memory?
Where the gpu and cpu can share the memory?
Which is key for performance.</p>
]]></description><pubDate>Wed, 24 Dec 2025 22:14:34 +0000</pubDate><link>https://news.ycombinator.com/item?id=46379844</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=46379844</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46379844</guid></item><item><title><![CDATA[New comment by eddyzh in "Dummy's Guide to Modern LLM Sampling"]]></title><description><![CDATA[
<p>LLM are way older. The Nobel prize for it shows how they made many of the breakthroughs decades ago 
ChatGTP was the popular breakthrough.
Even then your Smartphone keyboard has been using an LLM for a decade.</p>
]]></description><pubDate>Sun, 04 May 2025 21:23:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=43889704</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=43889704</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=43889704</guid></item><item><title><![CDATA[New comment by eddyzh in "Here are the Attack Plans That Trump's Advisers Shared on Signal"]]></title><description><![CDATA[
<p>Because optimizing for security is sometimes specifically decreasing convenience.
In the original (first) article some of those considerations are discussed.</p>
]]></description><pubDate>Mon, 31 Mar 2025 05:56:32 +0000</pubDate><link>https://news.ycombinator.com/item?id=43531578</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=43531578</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=43531578</guid></item><item><title><![CDATA[New comment by eddyzh in "Tern AI's low-cost GPS alternative actually works"]]></title><description><![CDATA[
<p>So I was curious what they mean by sending out an id for gps. Since that seems to me not something that is happening for all types of GPS implementations.</p>
]]></description><pubDate>Mon, 17 Mar 2025 13:19:21 +0000</pubDate><link>https://news.ycombinator.com/item?id=43388293</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=43388293</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=43388293</guid></item><item><title><![CDATA[Tern AI's low-cost GPS alternative actually works]]></title><description><![CDATA[
<p>Article URL: <a href="https://techcrunch.com/2025/03/14/tern-ais-low-cost-gps-alternative-actually-works/">https://techcrunch.com/2025/03/14/tern-ais-low-cost-gps-alternative-actually-works/</a></p>
<p>Comments URL: <a href="https://news.ycombinator.com/item?id=43388268">https://news.ycombinator.com/item?id=43388268</a></p>
<p>Points: 3</p>
<p># Comments: 3</p>
]]></description><pubDate>Mon, 17 Mar 2025 13:17:05 +0000</pubDate><link>https://techcrunch.com/2025/03/14/tern-ais-low-cost-gps-alternative-actually-works/</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=43388268</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=43388268</guid></item><item><title><![CDATA[New comment by eddyzh in "DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via RL"]]></title><description><![CDATA[
<p>This is great context for the cost claim.
Which turns out only to be technically true when looking at the final run.</p>
]]></description><pubDate>Sun, 26 Jan 2025 09:43:11 +0000</pubDate><link>https://news.ycombinator.com/item?id=42829040</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=42829040</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=42829040</guid></item><item><title><![CDATA[New comment by eddyzh in "Most people don't care about quality"]]></title><description><![CDATA[
<p>>But if there was a way to correctly tell people: "look, this smartphone is 20% more expensive, but it will last twice as long and it will be more convenient for you in ways you can't understand right now", nobody would go for the worse quality, right?<p>I am not sure if we are living in the same world. 
No most people absolutely do not make rational choices like that.
No matter how you tell them.<p>Most importantly what you think is best for people is stil a perspective. Especially in the context of tv.</p>
]]></description><pubDate>Fri, 03 Jan 2025 07:46:29 +0000</pubDate><link>https://news.ycombinator.com/item?id=42583509</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=42583509</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=42583509</guid></item><item><title><![CDATA[New comment by eddyzh in "Alignment faking in large language models"]]></title><description><![CDATA[
<p>Very fascinating read. Especially the reviewers comments linked at the end. The point is that alignment after training is much more complicated and limited than it might appear. And they make that point convincingly.</p>
]]></description><pubDate>Thu, 19 Dec 2024 07:04:14 +0000</pubDate><link>https://news.ycombinator.com/item?id=42459136</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=42459136</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=42459136</guid></item><item><title><![CDATA[New comment by eddyzh in "Oldest human genomes reveal how a small group burst out of Africa"]]></title><description><![CDATA[
<p>Makes a huge claim at the start.<p>>"The huge gap between those ages could change our understanding about how humans spread across the world. If the ancestors of today’s non-Africans didn’t sweep across other continents until 47,000 years ago, then those older sites must have been occupied by earlier waves of humans who died off without passing down their DNA to the people now living in places like China and Australia."<p>But at the end gets a bit more balanced<p>>"He Yu, a paleogeneticist at Peking University in Beijing who was not involved in either study, said that the mystery wouldn’t be solved until scientists find DNA in some of the ancient Asian fossils.
“We still need early modern human genomes from Asia to really talk about Asia stories,” Dr. Yu said."<p>This puzzle is still missing key elements.</p>
]]></description><pubDate>Wed, 18 Dec 2024 06:31:16 +0000</pubDate><link>https://news.ycombinator.com/item?id=42448522</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=42448522</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=42448522</guid></item><item><title><![CDATA[New comment by eddyzh in "Why is it so hard to buy things that work well? (2022)"]]></title><description><![CDATA[
<p>What an excellent written article.
Thanks Dan Luu!</p>
]]></description><pubDate>Tue, 17 Dec 2024 21:48:00 +0000</pubDate><link>https://news.ycombinator.com/item?id=42445653</link><dc:creator>eddyzh</dc:creator><comments>https://news.ycombinator.com/item?id=42445653</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=42445653</guid></item></channel></rss>