<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: aix1</title><link>https://news.ycombinator.com/user?id=aix1</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Sat, 29 Aug 2026 12:13:19 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=aix1" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by aix1 in "M5Stack Launches PaperMono"]]></title><description><![CDATA[
<p>Also firmware updates (on the non-locked variants), no?<p>What other potential uses are there?  Uploading books without WiFi is the only one that comes to mind.</p>
]]></description><pubDate>Fri, 28 Aug 2026 07:12:27 +0000</pubDate><link>https://news.ycombinator.com/item?id=49475352</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49475352</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49475352</guid></item><item><title><![CDATA[New comment by aix1 in "Anthropic's best AI model struggles to attract users as cheaper tools thrive"]]></title><description><![CDATA[
<p>One thing that jumped out at me is a slower-than-I'd-expect rate of adoption of Opus 5.  If you're running Opus 4.8 and choosing not to migrate to 5, what's behind that decision?</p>
]]></description><pubDate>Sun, 23 Aug 2026 10:29:56 +0000</pubDate><link>https://news.ycombinator.com/item?id=49407658</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49407658</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49407658</guid></item><item><title><![CDATA[New comment by aix1 in "Nvidia's new financial strategy does not compute"]]></title><description><![CDATA[
<p>These are GPU rental price futures, so you won't be investing in the hardware.<p>The pitch is that, if you know you'll need a certain amount of compute at a certain point in the future, you can lock in the rental price today.<p><a href="https://www.cmegroup.com/media-room/press-releases/2026/8/11/cme_group_and_silicondatatolaunchcomputefuturesonoctober5tounloc.html" rel="nofollow">https://www.cmegroup.com/media-room/press-releases/2026/8/11...</a></p>
]]></description><pubDate>Wed, 19 Aug 2026 17:22:20 +0000</pubDate><link>https://news.ycombinator.com/item?id=49364436</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49364436</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49364436</guid></item><item><title><![CDATA[New comment by aix1 in "Claude: System Prompts"]]></title><description><![CDATA[
<p>What have you been observing?  Genuinely curious; I use Fable as my main model and haven't noticed any regression.</p>
]]></description><pubDate>Sun, 16 Aug 2026 15:56:48 +0000</pubDate><link>https://news.ycombinator.com/item?id=49321190</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49321190</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49321190</guid></item><item><title><![CDATA[New comment by aix1 in "Anthropic revenue reportedly jumps to more than $11.5B in second quarter"]]></title><description><![CDATA[
<p>Yes they do.  They have to file an S-1 with the SEC, which will be made public about a month before the IPO.<p>The S-1 has to include, among other things, three years of audited financial statements, plus interim statements (unaudited).  It will cover both revenue and expenses, the latter breaking out things like cost of revenue, R&D, sales and marketing etc.<p>Based on the (unofficial but reported) IPO target date of late Sep to early Oct, the S-1 will have to be made public in a few weeks from now.</p>
]]></description><pubDate>Sun, 16 Aug 2026 15:16:17 +0000</pubDate><link>https://news.ycombinator.com/item?id=49320861</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49320861</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49320861</guid></item><item><title><![CDATA[New comment by aix1 in "Research papers using "kidney disappointment" instead of "kidney failure""]]></title><description><![CDATA[
<p>Here is one hypothesis: <a href="https://theconversation.com/problematic-paper-screener-trawling-for-fraud-in-the-scientific-literature-246317" rel="nofollow">https://theconversation.com/problematic-paper-screener-trawl...</a><p><quote>
Have you ever heard of the Joined Together States? Or bosom peril? Kidney disappointment? Fake neural organizations? Lactose bigotry? These nonsensical, and sometimes amusing, word sequences are among thousands of “tortured phrases” that sleuths have found littered throughout reputable scientific journals.<p>They typically result from using paraphrasing tools to evade plagiarism-detection software when stealing someone else’s text. The phrases above are real examples of bungled synonyms for the United States, breast cancer, kidney failure, artificial neural networks, and lactose intolerance, respectively.
</quote></p>
]]></description><pubDate>Sun, 16 Aug 2026 15:08:47 +0000</pubDate><link>https://news.ycombinator.com/item?id=49320804</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49320804</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49320804</guid></item><item><title><![CDATA[New comment by aix1 in "Ant teams beat gravity-based puzzle solvers"]]></title><description><![CDATA[
<p>Amazing, thanks for sharing.</p>
]]></description><pubDate>Sat, 15 Aug 2026 14:36:49 +0000</pubDate><link>https://news.ycombinator.com/item?id=49310950</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49310950</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49310950</guid></item><item><title><![CDATA[New comment by aix1 in "GLM-5.3: Frontier coding with emergent cyber capabilities"]]></title><description><![CDATA[
<p>It still <i>knows</i> how to speak English.  When I tell it to explain something in plain language, it generally does a very good job.  The weird thing is that those instructions don't persist: it lapses back into Claude-speak pretty much every turn no matter how hard I try to instruct it not to.<p>(In my case "it"=Fable; I assume Opus is similar.)</p>
]]></description><pubDate>Fri, 14 Aug 2026 07:26:15 +0000</pubDate><link>https://news.ycombinator.com/item?id=49295628</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49295628</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49295628</guid></item><item><title><![CDATA[New comment by aix1 in "Gemini 3.7 Flash"]]></title><description><![CDATA[
<p>With all due respect, did you read my comment beyond the first paragraph?  It addresses both points, TPU economics/pivot to sales + internal shortages making it hard to train models, to the extent they can be addressed based on public sources.<p>There are other factors at play, but they're more recent/second-order.</p>
]]></description><pubDate>Thu, 13 Aug 2026 18:37:54 +0000</pubDate><link>https://news.ycombinator.com/item?id=49290177</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49290177</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49290177</guid></item><item><title><![CDATA[New comment by aix1 in "Gemini 3.7 Flash"]]></title><description><![CDATA[
<p>The "let's make money by selling/renting out TPUs" faction has won and the "let's make money by training and selling a frontier model" faction has lost.<p>And it's arguably not crazy, at least if SemiAnalysis's estimates are to be believed:<p><pre><code>  * 20% of all TPU shipments from Q3 2026 through Q4 2027 are sold to SPVs serving Anthropic ($150B of contracted revenue); vs
  * ~$12B ARR for Gemini.
</code></pre>
<a href="https://newsletter.semianalysis.com/p/gemini-is-cooked-but-gcp-is-cooking" rel="nofollow">https://newsletter.semianalysis.com/p/gemini-is-cooked-but-g...</a><p>Because they compete for the same scarce resource, the result is a resource crunch for the group that's lost: <a href="https://www.latimes.com/business/story/2026-05-18/inside-ai-compute-crunch-driving-google-researchers-to-quit" rel="nofollow">https://www.latimes.com/business/story/2026-05-18/inside-ai-...</a></p>
]]></description><pubDate>Thu, 13 Aug 2026 18:00:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=49289688</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49289688</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49289688</guid></item><item><title><![CDATA[New comment by aix1 in "Gemini 3.7 Flash"]]></title><description><![CDATA[
<p>No, relabelling a Pro model as Flash would make no economic sense (the Pro series is larger than Flash and more expensive to serve).</p>
]]></description><pubDate>Thu, 13 Aug 2026 17:50:32 +0000</pubDate><link>https://news.ycombinator.com/item?id=49289552</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49289552</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49289552</guid></item><item><title><![CDATA[New comment by aix1 in "Gemini becomes Google's fastest-growing product ever as it hits 1B users"]]></title><description><![CDATA[
<p>Google did not sell those TPUs because they had excess capacity.  Google sold them because it can make lots more money that way then using them for Gemini inference (the shift to sales is also indicative of some power dynamics within the company).</p>
]]></description><pubDate>Wed, 12 Aug 2026 05:31:37 +0000</pubDate><link>https://news.ycombinator.com/item?id=49268145</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49268145</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49268145</guid></item><item><title><![CDATA[New comment by aix1 in "Stealing Reasoning Traces from Proprietary LLM APIs"]]></title><description><![CDATA[
<p>Good point, thanks.</p>
]]></description><pubDate>Tue, 11 Aug 2026 17:04:31 +0000</pubDate><link>https://news.ycombinator.com/item?id=49261244</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49261244</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49261244</guid></item><item><title><![CDATA[New comment by aix1 in "Stealing Reasoning Traces from Proprietary LLM APIs"]]></title><description><![CDATA[
<p>Having thought about this a little more, it's clear that server-side storage is not compatible with Zero Data Retention (ZDR).  However, in non-ZDR settings, it seems likely that the providers are capturing all that data anyway?<p>> a per user key would have solved this issue for sure<p>It would have helped with PII leakage, but not with plain-text trace extraction attacks, right?</p>
]]></description><pubDate>Tue, 11 Aug 2026 16:09:35 +0000</pubDate><link>https://news.ycombinator.com/item?id=49260463</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49260463</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49260463</guid></item><item><title><![CDATA[New comment by aix1 in "Stealing Reasoning Traces from Proprietary LLM APIs"]]></title><description><![CDATA[
<p>I really don't understand why server-side storage of the trace isn't a viable approach here, with only a unique key flowing to the client and back.  Does it have something to do with how backend load-balancing works?</p>
]]></description><pubDate>Tue, 11 Aug 2026 15:49:16 +0000</pubDate><link>https://news.ycombinator.com/item?id=49260167</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49260167</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49260167</guid></item><item><title><![CDATA[New comment by aix1 in "Changes at Google DeepMind: Demis Hassabis from CEO to Chair, Jeff Dean departs"]]></title><description><![CDATA[
<p>One detail I missed is that Koray's new title is "SVP of Google DeepMind"[1], i.e. identical to his old title.<p>So, basically, the third change is even more of a no-op than I thought.<p>It seems pretty clear that GDM will no longer have a CEO.<p>[1] <a href="https://blog.google/company-news/inside-google/message-ceo/next-chapter-ai-momentum/" rel="nofollow">https://blog.google/company-news/inside-google/message-ceo/n...</a></p>
]]></description><pubDate>Fri, 07 Aug 2026 05:46:31 +0000</pubDate><link>https://news.ycombinator.com/item?id=49206305</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49206305</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49206305</guid></item><item><title><![CDATA[New comment by aix1 in "Changes at Google DeepMind: Demis Hassabis from CEO to Chair, Jeff Dean departs"]]></title><description><![CDATA[
<p>So there are two events here:<p>Event 1 - Jeff & co leaving.<p>Event 2 - Demis stepping down. DeepMind doesn't need a "chair" and "Alphabet's chief scientist" is a bullshit title that Jeff invented for himself when he moved from Google Research to GDM, to make it look like we wasn't abandoning the former for the latter.<p>The third change is basically ratifying the status quo, since Koray has been de factor running GDM for a while now (and directly reporting to Sundar in addition to reporting to Demis, who in turn also reported to Sundar - quite some triangle there).</p>
]]></description><pubDate>Thu, 06 Aug 2026 18:34:01 +0000</pubDate><link>https://news.ycombinator.com/item?id=49200520</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49200520</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49200520</guid></item><item><title><![CDATA[New comment by aix1 in "Changes at Google DeepMind: Demis Hassabis from CEO to Chair, Jeff Dean departs"]]></title><description><![CDATA[
<p>I actually think there's a low probability of that.<p>The four have so much influence within the company that they could have trivially set this up as part of Alphabet, if they wanted to.  They are also ridiculously wealthy.</p>
]]></description><pubDate>Wed, 05 Aug 2026 17:22:13 +0000</pubDate><link>https://news.ycombinator.com/item?id=49185961</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49185961</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49185961</guid></item><item><title><![CDATA[New comment by aix1 in "Changes at Google DeepMind: Demis Hassabis from CEO to Chair, Jeff Dean departs"]]></title><description><![CDATA[
<p>> this could be a path for Demis to be CEO in the future in the AI era<p>His current title is CEO of Google DeepMind.  Becoming Chief Scientist of Alphabet seems to be a step away from the path to replacing Sundar, no?<p>P.S. I think his real interest is Isomorphic, and his new role will offer fewer distractions.</p>
]]></description><pubDate>Wed, 05 Aug 2026 16:45:16 +0000</pubDate><link>https://news.ycombinator.com/item?id=49185366</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49185366</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49185366</guid></item><item><title><![CDATA[New comment by aix1 in "Changes at Google DeepMind: Demis Hassabis from CEO to Chair, Jeff Dean departs"]]></title><description><![CDATA[
<p>I think there are multiple separate events here, timed to coincide in order to minimise disruption.  It's hard to be sure whether they share overlapping causes, but I'd guess that there's at least an element of that.</p>
]]></description><pubDate>Wed, 05 Aug 2026 16:39:26 +0000</pubDate><link>https://news.ycombinator.com/item?id=49185276</link><dc:creator>aix1</dc:creator><comments>https://news.ycombinator.com/item?id=49185276</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49185276</guid></item></channel></rss>