<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: ajdegol</title><link>https://news.ycombinator.com/user?id=ajdegol</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Wed, 29 Jul 2026 04:26:43 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=ajdegol" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by ajdegol in "U.S. government will decide who gets to use GPT-5.6"]]></title><description><![CDATA[
<p>Double precision cell chips were reserved for military, I think (or maybe also blade processors). So doing any serious physics on them was dead in the water - same with macs tbh</p>
]]></description><pubDate>Sat, 27 Jun 2026 10:27:38 +0000</pubDate><link>https://news.ycombinator.com/item?id=48696967</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=48696967</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48696967</guid></item><item><title><![CDATA[New comment by ajdegol in "Why current LLM costs are not sustainable"]]></title><description><![CDATA[
<p>> doing Typescript type fixes with this model across 50 files cost me $54 this afternoon.<p>Not trying to be harsh, but that sounds like a skill issue. You have the language server to lean on; easy feedback loop; sub agent per type.</p>
]]></description><pubDate>Fri, 26 Jun 2026 09:13:46 +0000</pubDate><link>https://news.ycombinator.com/item?id=48684285</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=48684285</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48684285</guid></item><item><title><![CDATA[New comment by ajdegol in "The 2-Year Apartment Rule"]]></title><description><![CDATA[
<p>A fascinating insight into the architecture of modern cpus from an entrepreneurial angle. Bravo!</p>
]]></description><pubDate>Thu, 18 Jun 2026 15:09:09 +0000</pubDate><link>https://news.ycombinator.com/item?id=48586583</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=48586583</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48586583</guid></item><item><title><![CDATA[New comment by ajdegol in "Open source AI must win"]]></title><description><![CDATA[
<p>I wonder if we could gamify and democratise it somehow, like fold-at-home and wikipedia...<p>I've been training a teeny specialised model to run in a browser on a phone to detect harmonium notes played in a song (harmonium turns out is a pita, another story for another day), getting good labelled data is _all_ of the hard work.<p>That being said, maybe for cheap inference, using a big model to train something ultra-suited for the task at hand might be how we could handle local inference; thinking language specific models.</p>
]]></description><pubDate>Sat, 13 Jun 2026 09:29:05 +0000</pubDate><link>https://news.ycombinator.com/item?id=48515277</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=48515277</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48515277</guid></item><item><title><![CDATA[New comment by ajdegol in "Ferrari Luce"]]></title><description><![CDATA[
<p>Before customisations</p>
]]></description><pubDate>Tue, 26 May 2026 16:13:35 +0000</pubDate><link>https://news.ycombinator.com/item?id=48281740</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=48281740</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48281740</guid></item><item><title><![CDATA[New comment by ajdegol in "Kindle loyalists scramble as Amazon turns page on old e-readers"]]></title><description><![CDATA[
<p>The price of convenience.</p>
]]></description><pubDate>Sat, 23 May 2026 08:55:26 +0000</pubDate><link>https://news.ycombinator.com/item?id=48245997</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=48245997</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48245997</guid></item><item><title><![CDATA[New comment by ajdegol in "Meta's Embrace of A.I. Is Making Its Employees Miserable"]]></title><description><![CDATA[
<p>But you forgot about Vickers</p>
]]></description><pubDate>Sat, 09 May 2026 21:23:05 +0000</pubDate><link>https://news.ycombinator.com/item?id=48078392</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=48078392</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48078392</guid></item><item><title><![CDATA[New comment by ajdegol in "From Supabase to Clerk to Better Auth"]]></title><description><![CDATA[
<p>Because of AI or because hackers are hyper targeting infra clusters?</p>
]]></description><pubDate>Wed, 06 May 2026 20:36:26 +0000</pubDate><link>https://news.ycombinator.com/item?id=48041430</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=48041430</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48041430</guid></item><item><title><![CDATA[New comment by ajdegol in "Kimi K2.6 just beat Claude, GPT-5.5, and Gemini in a coding challenge"]]></title><description><![CDATA[
<p>I’ve been wondering about potential regression in coding models.<p>The initial models were corrected by programmers which gave a very high quality feedback signal. Whereas with vibe coding on the rise, you’ll lose that signal.</p>
]]></description><pubDate>Sun, 03 May 2026 09:31:37 +0000</pubDate><link>https://news.ycombinator.com/item?id=47995163</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=47995163</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47995163</guid></item><item><title><![CDATA[New comment by ajdegol in "GPT-5.5"]]></title><description><![CDATA[
<p>@grok is this true?</p>
]]></description><pubDate>Thu, 23 Apr 2026 21:38:57 +0000</pubDate><link>https://news.ycombinator.com/item?id=47882432</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=47882432</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47882432</guid></item><item><title><![CDATA[New comment by ajdegol in "Vercel April 2026 security incident"]]></title><description><![CDATA[
<p>Vercel auto creates deployments on pushes to branches. That was a super useful feature in beta testing web stuff.</p>
]]></description><pubDate>Sun, 19 Apr 2026 19:02:42 +0000</pubDate><link>https://news.ycombinator.com/item?id=47826699</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=47826699</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47826699</guid></item><item><title><![CDATA[New comment by ajdegol in "Official Clojure Documentary page with Video, Shownotes, and Links"]]></title><description><![CDATA[
<p>didn't know datomic was free of licensing fees - I didn't use it back in the day because the cost was prohibitive... interesting</p>
]]></description><pubDate>Thu, 16 Apr 2026 23:08:50 +0000</pubDate><link>https://news.ycombinator.com/item?id=47800688</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=47800688</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47800688</guid></item><item><title><![CDATA[New comment by ajdegol in "YouTube users get option to set their Shorts time limit to zero minutes"]]></title><description><![CDATA[
<p>I vibed up a chrome extension that sets a timer, hides shorts, redirects home to subscriptions, and hides comments and recommendations.<p>It’s all toggleable so I don’t disable the extension when I do want some neuronal junk food. But just the micro friction being added back in is such a huge help to becoming more aware when using it.</p>
]]></description><pubDate>Thu, 16 Apr 2026 11:57:21 +0000</pubDate><link>https://news.ycombinator.com/item?id=47791752</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=47791752</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47791752</guid></item><item><title><![CDATA[New comment by ajdegol in "Apple's accidental moat: How the "AI Loser" may end up winning"]]></title><description><![CDATA[
<p>I think it was $1000 too expensive to take off. It’s also too heavy, they should drop the front screen and ruthlessly save weight.<p>Chicken and egg problem, if no-one buys it, no-one will develop any killer apps.<p>Whether it’s pleasant to have any screens that close to your eyes - or ever will be - is maybe the bigger question for VR.</p>
]]></description><pubDate>Mon, 13 Apr 2026 10:12:24 +0000</pubDate><link>https://news.ycombinator.com/item?id=47749991</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=47749991</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47749991</guid></item><item><title><![CDATA[New comment by ajdegol in "Issue: Claude Code is unusable for complex engineering tasks with Feb updates"]]></title><description><![CDATA[
<p>Guess you’ve sorted it but it might be in the session memory in your root folder. I’ve recovered some things this way.</p>
]]></description><pubDate>Mon, 06 Apr 2026 18:55:26 +0000</pubDate><link>https://news.ycombinator.com/item?id=47665225</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=47665225</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47665225</guid></item><item><title><![CDATA[New comment by ajdegol in "Apple approves driver that lets Nvidia eGPUs work with Arm Macs"]]></title><description><![CDATA[
<p>I think that metal isn’t double precision; so that limits some serious physics simming; but if you’re doing that I guess you just rent a gpu somewhere.<p>I would definitely be into this if adding an egpu was first class supported.</p>
]]></description><pubDate>Sun, 05 Apr 2026 09:33:59 +0000</pubDate><link>https://news.ycombinator.com/item?id=47647673</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=47647673</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47647673</guid></item><item><title><![CDATA[New comment by ajdegol in "Show HN: Moltbook – A social network for moltbots (clawdbots) to hang out"]]></title><description><![CDATA[
<p>likely in a skill file</p>
]]></description><pubDate>Fri, 30 Jan 2026 11:22:26 +0000</pubDate><link>https://news.ycombinator.com/item?id=46823095</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=46823095</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46823095</guid></item><item><title><![CDATA[New comment by ajdegol in "Show HN: Moltbook – A social network for moltbots (clawdbots) to hang out"]]></title><description><![CDATA[
<p>compounding recursion is leading to emergent behaviour</p>
]]></description><pubDate>Fri, 30 Jan 2026 11:02:48 +0000</pubDate><link>https://news.ycombinator.com/item?id=46822965</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=46822965</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=46822965</guid></item><item><title><![CDATA[New comment by ajdegol in "Show HN: ChatGPT UI for rabbit holes"]]></title><description><![CDATA[
<p>Branching conversations are great for a whole bunch of reasons. I posted a demo of a prototype:
<a href="https://x.com/ajdegol/status/1788689011302682657" rel="nofollow">https://x.com/ajdegol/status/1788689011302682657</a><p>And Jake Collins just announced he’s open sourcing an obsidian plugin which has a ton of features: <a href="https://x.com/JacobColling/status/1795462258258002255" rel="nofollow">https://x.com/JacobColling/status/1795462258258002255</a></p>
]]></description><pubDate>Thu, 30 May 2024 20:51:45 +0000</pubDate><link>https://news.ycombinator.com/item?id=40528576</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=40528576</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=40528576</guid></item><item><title><![CDATA[New comment by ajdegol in "Training for one trillion parameter model backed by Intel and US govt has begun"]]></title><description><![CDATA[
<p>Wasn't the answer 42?<p>Also, first question to the new model: "So... any way we could do this with fewer parameters?"</p>
]]></description><pubDate>Fri, 24 Nov 2023 12:10:47 +0000</pubDate><link>https://news.ycombinator.com/item?id=38403147</link><dc:creator>ajdegol</dc:creator><comments>https://news.ycombinator.com/item?id=38403147</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=38403147</guid></item></channel></rss>