<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: ilc</title><link>https://news.ycombinator.com/user?id=ilc</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Wed, 19 Aug 2026 01:52:52 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=ilc" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by ilc in "Cursor launches Origin, GitHub alternative"]]></title><description><![CDATA[
<p>A quick question: Where can I read your TOS, and other legal documents?<p>Knowing things like: Will my code end up in Grok?  What ownership do I have of my content, etc... matter.</p>
]]></description><pubDate>Tue, 18 Aug 2026 22:50:03 +0000</pubDate><link>https://news.ycombinator.com/item?id=49353926</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=49353926</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49353926</guid></item><item><title><![CDATA[New comment by ilc in "Qwen3.8-2.4T"]]></title><description><![CDATA[
<p>Any one weight, but all of them.  And also crushing the architecture itself?<p>I wouldn't pick up 400gb of hardware to run in that mode.  I might try it for fun, but even then you are looking at handling a 95GB active parameter set.<p>This is NOT a model for most home labs.  I'm sure some can and will use it.  But most, should steer clear.</p>
]]></description><pubDate>Wed, 12 Aug 2026 15:49:53 +0000</pubDate><link>https://news.ycombinator.com/item?id=49274320</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=49274320</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49274320</guid></item><item><title><![CDATA[New comment by ilc in "Qwen3.8-2.4T"]]></title><description><![CDATA[
<p>To compare a 1 bit quant to the full fat model is misleading.<p>Honestly this model people at home can tinker with, if you have a big enough Mac.  Maybe 4 Strix Halo/DGX Spark, and then at 1 bit quant?  Nah.<p>Use the right sized model, for your hardware.  You'll get better results.</p>
]]></description><pubDate>Wed, 12 Aug 2026 15:40:04 +0000</pubDate><link>https://news.ycombinator.com/item?id=49274172</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=49274172</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49274172</guid></item><item><title><![CDATA[New comment by ilc in "U of Michigan drops first-semester grades to ‘curb mental health crisis’"]]></title><description><![CDATA[
<p>To those saying this is soft: <a href="https://registrar.mit.edu/classes-grades-evaluations/grades/grading-policies" rel="nofollow">https://registrar.mit.edu/classes-grades-evaluations/grades/...</a><p>If you want to say MIT is soft.  Feel free.</p>
]]></description><pubDate>Tue, 11 Aug 2026 16:28:49 +0000</pubDate><link>https://news.ycombinator.com/item?id=49260716</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=49260716</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49260716</guid></item><item><title><![CDATA[New comment by ilc in "Muse Code and Muse Spark 1.2"]]></title><description><![CDATA[
<p>Having audited one code harness (Qwen Code) in the past on telemetry.  Let's say I'm sus of the Zuck offering free candy.</p>
]]></description><pubDate>Thu, 06 Aug 2026 14:00:07 +0000</pubDate><link>https://news.ycombinator.com/item?id=49196838</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=49196838</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49196838</guid></item><item><title><![CDATA[New comment by ilc in "Humans missed 1 in 3 threats approving AI agent commands across 40k game runs"]]></title><description><![CDATA[
<p>Sandbox and use Local AI.  This is the real answer.</p>
]]></description><pubDate>Thu, 06 Aug 2026 13:55:43 +0000</pubDate><link>https://news.ycombinator.com/item?id=49196783</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=49196783</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49196783</guid></item><item><title><![CDATA[New comment by ilc in "Rune 1.1: adds Python, an Emacs editor, a symbol index and is now free"]]></title><description><![CDATA[
<p>As soon as it said "non-commercial" I noped.<p>That is so vague with some of what I do, that it isn't worth even looking further.</p>
]]></description><pubDate>Fri, 31 Jul 2026 02:37:10 +0000</pubDate><link>https://news.ycombinator.com/item?id=49118389</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=49118389</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49118389</guid></item><item><title><![CDATA[New comment by ilc in "MCP 2026-07-28 Specification: transport going stateless"]]></title><description><![CDATA[
<p>I'd made the shift to HTTP/Stateless from MCP a few months ago.  It's the right thing to do IMHO.  Reliability up, problems down.  TOON support is natural, if desired, etc.<p>My only question is how do you handle channels in the architecture now.  From what I saw in Claude Code, shifting to a totally http world has some timeout issues if a server drops out and comes back.  Because of that I'm stuck writing stubs for my internal use MCP, this is fine for me, but if you are cleaning up semantics:  Understanding how we expect clients to act around failure would really help, the story.</p>
]]></description><pubDate>Tue, 28 Jul 2026 19:36:52 +0000</pubDate><link>https://news.ycombinator.com/item?id=49088837</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=49088837</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49088837</guid></item><item><title><![CDATA[New comment by ilc in "Laguna S 2.1"]]></title><description><![CDATA[
<p>This is definitely one where wait 2-3 weeks and look again, feels like the right strategy.  It could be good, it could not.  It is too volatile to say.</p>
]]></description><pubDate>Fri, 24 Jul 2026 16:21:57 +0000</pubDate><link>https://news.ycombinator.com/item?id=49037883</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=49037883</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49037883</guid></item><item><title><![CDATA[New comment by ilc in "Laguna S 2.1"]]></title><description><![CDATA[
<p>What harness/quant did you use for testing?</p>
]]></description><pubDate>Tue, 21 Jul 2026 20:06:29 +0000</pubDate><link>https://news.ycombinator.com/item?id=48997500</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=48997500</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48997500</guid></item><item><title><![CDATA[New comment by ilc in "China’s open-weights AI strategy is winning"]]></title><description><![CDATA[
<p>Not the parts which matter to trust.  Which is my point.<p>You can state the math, but not why it won't discuss various topics, etc.  Once you see the models waffling on subject with objective truths.  You wonder what else is wrong.<p>I do not exempt US models from this.  They do it too, ask anything about politics, elections etc.  And they can get... weird.<p>It doesn't take much to create a systemic error class in a model at these scales.  And history has shown nation states are willing to do these things.<p>Just be wary.</p>
]]></description><pubDate>Mon, 20 Jul 2026 22:12:41 +0000</pubDate><link>https://news.ycombinator.com/item?id=48985530</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=48985530</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48985530</guid></item><item><title><![CDATA[New comment by ilc in "China’s open-weights AI strategy is winning"]]></title><description><![CDATA[
<p>Training data, training methodology.  All NOT OPEN.<p>Until we know what a model is trained on, and how it is trained in high detail, I hesitate to call them "Open Source" in any way.  They are free.  But, we don't know what their priorities are etc.  Witness the censorship we see in all models in one form or another.  I'm not absolving any side of this.<p>Just saying: Don't be blind.</p>
]]></description><pubDate>Mon, 20 Jul 2026 17:03:12 +0000</pubDate><link>https://news.ycombinator.com/item?id=48981553</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=48981553</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48981553</guid></item><item><title><![CDATA[New comment by ilc in "OpenAI reduces Codex Model Context Size from 372k to 272k"]]></title><description><![CDATA[
<p>Opus: Yeah, it can be a little fuzzy at 600-700k, I prefer it 400K and under in general.  But, I've had usable sessions to 850k.<p>It really depends on the session, and what you are doing how far you can push it.</p>
]]></description><pubDate>Sun, 19 Jul 2026 23:58:08 +0000</pubDate><link>https://news.ycombinator.com/item?id=48972857</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=48972857</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48972857</guid></item><item><title><![CDATA[New comment by ilc in "Sick leave: Germany rising but not the worst in Europe"]]></title><description><![CDATA[
<p>There will always be abusers.  The issue I see is day 1 asks.  Not day N.  If I'm out a week asking why is reasonable.  If I'm out a day, it's more annoying for everyone.<p>And I'm not advocating for no limits with no responsibility to the employer.  That'd be stupid.  Clearly there's a "more sane" middle ground.<p>Even on a menial job, they are dealing with more value than they make daily, by definition.<p>A middle road exists, and it is solid, why it isn't being taken, I don't know.</p>
]]></description><pubDate>Sat, 04 Jul 2026 19:33:19 +0000</pubDate><link>https://news.ycombinator.com/item?id=48788218</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=48788218</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48788218</guid></item><item><title><![CDATA[New comment by ilc in "Only 1 of the Top 5 AI Coding Models on WebDev Arena Isn't Chinese"]]></title><description><![CDATA[
<p>Misinformation.  Incorrect sorting selected to prove a point.<p>Nice try OP.</p>
]]></description><pubDate>Sat, 04 Jul 2026 19:22:02 +0000</pubDate><link>https://news.ycombinator.com/item?id=48788119</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=48788119</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48788119</guid></item><item><title><![CDATA[New comment by ilc in "Sick leave: Germany rising but not the worst in Europe"]]></title><description><![CDATA[
<p>Disclaimer: I'm a US Citizen / Worker, white collar.<p>My first job had: Unlimited sick time.  Take it when you need it.  No doctor's note.  I think you may have needed one at the week mark.  Never hit at that employer.  Most employers have been limited sick, with a note at a week.<p>I can't imagine note on day 1.  That's just... nuts.  The lack of trust shown there is massive.  I'm making calls on stuff far larger than my sick days.  If you can't trust me on sick days, how can you trust me to do my job?</p>
]]></description><pubDate>Sat, 04 Jul 2026 19:14:53 +0000</pubDate><link>https://news.ycombinator.com/item?id=48788055</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=48788055</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48788055</guid></item><item><title><![CDATA[New comment by ilc in "Leanstral 1.5: Proof abundance for all"]]></title><description><![CDATA[
<p>Given that they directly compare to GPT-5.5 in their documentation.   This comes off as puppy kicking to me.  They state it is not SOTA, even IN its domain!<p>Honestly: Think twice before dragging your firm into what you say.<p>Disclaimer: I speak for myself.  Not any firm I am associated with.</p>
]]></description><pubDate>Sat, 04 Jul 2026 01:55:14 +0000</pubDate><link>https://news.ycombinator.com/item?id=48781990</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=48781990</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48781990</guid></item><item><title><![CDATA[New comment by ilc in "Show HN: Oak – Git alternative designed for agents"]]></title><description><![CDATA[
<p>I could, I hit it in some situations.  But I'm running some fair size repos where I do multi-worktree work.<p>That said, it hasn't been enough of an issue for me to want to fix.</p>
]]></description><pubDate>Tue, 23 Jun 2026 00:01:02 +0000</pubDate><link>https://news.ycombinator.com/item?id=48638287</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=48638287</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48638287</guid></item><item><title><![CDATA[New comment by ilc in "Ask HN: How do you get into a flow state when using AI to code?"]]></title><description><![CDATA[
<p>Built my own orchestration layer, to allow me to optimize the one thing that is most precious when working with AI.  My attention and my time.</p>
]]></description><pubDate>Thu, 11 Jun 2026 17:31:52 +0000</pubDate><link>https://news.ycombinator.com/item?id=48493501</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=48493501</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48493501</guid></item><item><title><![CDATA[New comment by ilc in "The tech jobs bust is real. Don't blame AI (yet)"]]></title><description><![CDATA[
<p>It's dead Jim.</p>
]]></description><pubDate>Mon, 13 Apr 2026 23:25:57 +0000</pubDate><link>https://news.ycombinator.com/item?id=47759288</link><dc:creator>ilc</dc:creator><comments>https://news.ycombinator.com/item?id=47759288</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=47759288</guid></item></channel></rss>