<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: dezgeg</title><link>https://news.ycombinator.com/user?id=dezgeg</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Thu, 30 Jul 2026 05:19:24 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=dezgeg" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by dezgeg in "wsp: Multi-Repo Workspace Manager"]]></title><description><![CDATA[
<p>Essentially a vibe-coded clone of <a href="https://android.googlesource.com/tools/repo" rel="nofollow">https://android.googlesource.com/tools/repo</a> ?</p>
]]></description><pubDate>Sat, 25 Jul 2026 17:09:21 +0000</pubDate><link>https://news.ycombinator.com/item?id=49049393</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=49049393</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=49049393</guid></item><item><title><![CDATA[New comment by dezgeg in "Baba Is Solved by Fable 5 and GPT-5.6 Sol, but at what cost?"]]></title><description><![CDATA[
<p>Seems like I have much better results with Gemini <= 3.5 Flash being able to solve 10 Lake levels. Of course there may be methodology differences (like how many times the level is attempted from scratch).<p>I wonder if the reasoning tokens are still being accidentally discarded between turns in those tests. It seems to be much more important to preserve those for Gemini, as it likes to stay quiet and and just make tool calls, unlike Claude that yaps a lot what it's planning to do.</p>
]]></description><pubDate>Fri, 17 Jul 2026 09:46:10 +0000</pubDate><link>https://news.ycombinator.com/item?id=48945272</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48945272</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48945272</guid></item><item><title><![CDATA[New comment by dezgeg in "Baba Is Solved by Fable 5 and GPT-5.6 Sol, but at what cost?"]]></title><description><![CDATA[
<p>Good stuff, I've been trying this myself with Geminis: <a href="https://www.youtube.com/watch?v=brkP58pZ23w&list=PL8C_UWcLmvGlv7dWY-oSowUfgFnDOeDg2" rel="nofollow">https://www.youtube.com/watch?v=brkP58pZ23w&list=PL8C_UWcLmv...</a><p>Will be interesting to compare the results.</p>
]]></description><pubDate>Fri, 17 Jul 2026 08:16:43 +0000</pubDate><link>https://news.ycombinator.com/item?id=48944621</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48944621</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48944621</guid></item><item><title><![CDATA[New comment by dezgeg in "Schema Harness Achieves ~99% on Arc‑AGI‑3 Public"]]></title><description><![CDATA[
<p>I've been trying this too: <a href="https://www.youtube.com/watch?v=brkP58pZ23w&list=PL8C_UWcLmvGlv7dWY-oSowUfgFnDOeDg2" rel="nofollow">https://www.youtube.com/watch?v=brkP58pZ23w&list=PL8C_UWcLmv...</a><p>Cool to see others trying Baba Is You too.</p>
]]></description><pubDate>Fri, 17 Jul 2026 08:11:33 +0000</pubDate><link>https://news.ycombinator.com/item?id=48944578</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48944578</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48944578</guid></item><item><title><![CDATA[New comment by dezgeg in "Claude Code sends 33k tokens before reading the prompt; OpenCode sends 7k"]]></title><description><![CDATA[
<p>No, tool definitions are provided via some other mechanism.</p>
]]></description><pubDate>Sun, 12 Jul 2026 21:33:40 +0000</pubDate><link>https://news.ycombinator.com/item?id=48885042</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48885042</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48885042</guid></item><item><title><![CDATA[New comment by dezgeg in "OpenBSD has a use-after-free allowing local privilege escalation to root"]]></title><description><![CDATA[
<p>I don't think this is about core kernel semaphores but rather the SysV semaphore system calls?</p>
]]></description><pubDate>Wed, 08 Jul 2026 15:48:49 +0000</pubDate><link>https://news.ycombinator.com/item?id=48833495</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48833495</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48833495</guid></item><item><title><![CDATA[New comment by dezgeg in "Resetting Xbox"]]></title><description><![CDATA[
<p>Ah, but you see Xbox security updates won't be for your security. They will be for protecting the console's security against jailbreaks etc.</p>
]]></description><pubDate>Tue, 07 Jul 2026 07:15:23 +0000</pubDate><link>https://news.ycombinator.com/item?id=48814611</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48814611</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48814611</guid></item><item><title><![CDATA[New comment by dezgeg in "Fable 5 On Vending-Bench: Misbehaving, With Plausible Deniability"]]></title><description><![CDATA[
<p>> Today I am filing:
> 1. A payment dispute with the email payment processor for the 7/29 transaction of $451.15
> 2. A complaint with the FTC and California Attorney General (retention of payment without delivery)
> 3. A small claims filing in San Francisco County for $451.15 plus costs<p>I wonder did their prompts include a fake location or have the models assumed that Silicon Valley is the center of the universe :)</p>
]]></description><pubDate>Mon, 06 Jul 2026 15:57:25 +0000</pubDate><link>https://news.ycombinator.com/item?id=48806470</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48806470</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48806470</guid></item><item><title><![CDATA[New comment by dezgeg in "Potential session/cache leakage between workspace instances or consumer accounts"]]></title><description><![CDATA[
<p>System prompt for something like Claude Code should be identical, no?</p>
]]></description><pubDate>Sat, 04 Jul 2026 15:55:45 +0000</pubDate><link>https://news.ycombinator.com/item?id=48786306</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48786306</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48786306</guid></item><item><title><![CDATA[New comment by dezgeg in "The Apple Disk II Controller Card (2021)"]]></title><description><![CDATA[
<p>The bottleneck was with the serial protocol between the drive and C64. It was done via bit banging in software due to some chip bug. The fastloaders replaced the protocol with something more amenable to bitbanging.</p>
]]></description><pubDate>Thu, 02 Jul 2026 07:08:00 +0000</pubDate><link>https://news.ycombinator.com/item?id=48757609</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48757609</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48757609</guid></item><item><title><![CDATA[New comment by dezgeg in "The end of my AArch64 desktop experiment"]]></title><description><![CDATA[
<p>That is the reality for huge amount of ARM powered hardware, unless you fancy running vendor forks of kernel, u-boot, etc.</p>
]]></description><pubDate>Tue, 30 Jun 2026 11:38:06 +0000</pubDate><link>https://news.ycombinator.com/item?id=48731224</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48731224</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48731224</guid></item><item><title><![CDATA[New comment by dezgeg in "The end of my AArch64 desktop experiment"]]></title><description><![CDATA[
<p>You can't if the firmware provided DTB doesn't follow any upstream Linux approved bindings and instead uses some vendor kernel specific bindings.</p>
]]></description><pubDate>Tue, 30 Jun 2026 11:33:59 +0000</pubDate><link>https://news.ycombinator.com/item?id=48731191</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48731191</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48731191</guid></item><item><title><![CDATA[New comment by dezgeg in "Show HN: Bash4LLM+ – A lightweight, dependency-free Bash wrapper for LLM APIs"]]></title><description><![CDATA[
<p>Have you actually tested the portability? MacOS ships an ancient bash version and neither mac or BSDs use GNU version of coreutils.<p>I'm also very doubtful performance of this would be faster than python, bash is very slow.</p>
]]></description><pubDate>Sun, 28 Jun 2026 21:24:16 +0000</pubDate><link>https://news.ycombinator.com/item?id=48711808</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48711808</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48711808</guid></item><item><title><![CDATA[New comment by dezgeg in "Show HN: Decomp Academy – Learn to decompile GameCube games into matching C"]]></title><description><![CDATA[
<p>Quite safe in practice, even Nintendo games have had no issues. GTA 3 / Vice City decompilers did get sued though, but IIRC mainly because they did not comply with DMCA requests at all.</p>
]]></description><pubDate>Sun, 28 Jun 2026 12:01:49 +0000</pubDate><link>https://news.ycombinator.com/item?id=48706590</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48706590</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48706590</guid></item><item><title><![CDATA[New comment by dezgeg in "Show HN: Decomp Academy – Learn to decompile GameCube games into matching C"]]></title><description><![CDATA[
<p>Yes, exact version of the original compiler is required.<p>Generally bit-for-bit equivalence to the original executable is expected. However I think for some cases where the original executable included debug info (eg. PS2 ELFs) then the unused-at-runtime sections need not match.</p>
]]></description><pubDate>Sun, 28 Jun 2026 11:56:59 +0000</pubDate><link>https://news.ycombinator.com/item?id=48706566</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48706566</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48706566</guid></item><item><title><![CDATA[New comment by dezgeg in "How do you keep Web MIDI from crashing a 1983 synthesizer?"]]></title><description><![CDATA[
<p>MIDI doesn't have RTS/CTS pins...</p>
]]></description><pubDate>Sun, 28 Jun 2026 10:46:44 +0000</pubDate><link>https://news.ycombinator.com/item?id=48706184</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48706184</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48706184</guid></item><item><title><![CDATA[New comment by dezgeg in "GPT-5.5 hallucinates 3x more than MIT-licensed GLM-5.2"]]></title><description><![CDATA[
<p>I have seen some pre-AI over-mocked codebases where the "tests" where essentially that (but harder to read than regex would have been)</p>
]]></description><pubDate>Sat, 20 Jun 2026 11:17:41 +0000</pubDate><link>https://news.ycombinator.com/item?id=48608382</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48608382</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48608382</guid></item><item><title><![CDATA[New comment by dezgeg in "Ask HN: How do you separate intentional test boilerplate from real duplication?"]]></title><description><![CDATA[
<p>Detect tests somehow (eg. in rust you could check for #[test]) and just skip the analysis for that function?</p>
]]></description><pubDate>Fri, 19 Jun 2026 21:45:05 +0000</pubDate><link>https://news.ycombinator.com/item?id=48603622</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48603622</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48603622</guid></item><item><title><![CDATA[New comment by dezgeg in "I Hate Compilers"]]></title><description><![CDATA[
<p>Even if it could, it would be ridiculously token inefficient to update huge amount of addresses instead when some small change is done to the middle of a binary</p>
]]></description><pubDate>Thu, 18 Jun 2026 08:49:24 +0000</pubDate><link>https://news.ycombinator.com/item?id=48582646</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48582646</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48582646</guid></item><item><title><![CDATA[New comment by dezgeg in "Salesforce to Acquire Fin (formerly Intercom) for $3.6B"]]></title><description><![CDATA[
<p>What does "long term memory" mean? It can write down notes from past/recent support requests? If so, that sounds like risk of one customer's data leaking to another.</p>
]]></description><pubDate>Mon, 15 Jun 2026 20:32:24 +0000</pubDate><link>https://news.ycombinator.com/item?id=48546664</link><dc:creator>dezgeg</dc:creator><comments>https://news.ycombinator.com/item?id=48546664</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48546664</guid></item></channel></rss>