Hacker News: dezgeg

New comment by dezgeg in "The scourge of x86 emulation"

dezgeg — Fri, 18 Sep 2026 08:58:04 +0000

Games use plenty of native code; so much in fact that when Intel tried to enter the Android tablet market with x86 they had to write an ARM emulator.

New comment by dezgeg in "Reverse engineering my e-scooter and rewriting the firmware in Rust"

dezgeg — Mon, 14 Sep 2026 11:14:19 +0000

4G/5G Telco equipment (ab)use HDMI connectors/cables in a similar way.

New comment by dezgeg in "Homebrew 7.0.0"

dezgeg — Sun, 13 Sep 2026 13:31:43 +0000

How is that any different from PyPI, npm, cargo, etc?

New comment by dezgeg in "Git Submodules as a Package Manager"

dezgeg — Sat, 05 Sep 2026 15:58:19 +0000

Correct way is to have everything as-if a source dependency, but a globally shared cache system that can automatically provide artifacts if that exact version of source has already been compiled.

Any system that can only consume manually built (and even in worst case, only tagged and globally pushed) dependencies is horrible for cross-module development, like large scale refactorings and the like.

New comment by dezgeg in "Claude Fable 5.1 and Claude Mythos 5.1"

dezgeg — Tue, 01 Sep 2026 18:59:31 +0000

Yeah, for all the hate Gemini gets, at least it isn't obsessed with adding comments and it's output is more readable than recent Claude's.

New comment by dezgeg in "Ox-Alpha Is GLM?"

dezgeg — Tue, 25 Aug 2026 13:28:14 +0000

I extremely doubt it's Gemini - all previous Geminis have had very different output style from other models. In a coding harness Gemini is extremely quiet and mostly just does thinking and tool calls, with the thinking blocks heavily summarized.

Also Claude and the Chinese models really like to say "Let me X", Geminis preference is "I will X".

New comment by dezgeg in "Cancer-related mortality among US pilots and flight attendants"

dezgeg — Fri, 21 Aug 2026 19:46:30 +0000

I wonder if pilots' medical checks would catch some of these early?

New comment by dezgeg in "Nvidia AVO scores 100% on the ARC-AGI-3 interactive reasoning benchmark"

dezgeg — Fri, 21 Aug 2026 16:34:17 +0000

GPT-o3 was the first OpenAI one to beat Red. The harness used by GPT Plays Pokemon is the most featureful one of the main competitors (GPT, Claude, Gemini), IIRC.

Community maintained spreadsheet of the runs: https://docs.google.com/spreadsheets/d/e/2PACX-1vQDvsy5Dt_-P...

New comment by dezgeg in "Spaghettifying DRAM"

dezgeg — Fri, 14 Aug 2026 13:55:09 +0000

How so? In my experience, the peak of Device Tree days was with the final 32-bit ARM chips before move to aarch64. Back then you had U-Boot which did minimal stuff and after that the kernel had full ownership of the hardware, no management firmware layers or extra code at runtime.

It's during the move to 64-bit aarch64 where things started to imitate the x86, with more and more always-on firmware and secure monitors and whatever being introduced. Nothing really to do with device tree, in fact ACPI is being pushed to the ARM too, yet again hiding more details of the hardware towards proprietary bytecode.

New comment by dezgeg in "wsp: Multi-Repo Workspace Manager"

dezgeg — Sat, 25 Jul 2026 17:09:21 +0000

Essentially a vibe-coded clone of https://android.googlesource.com/tools/repo ?

New comment by dezgeg in "Baba Is Solved by Fable 5 and GPT-5.6 Sol, but at what cost?"

dezgeg — Fri, 17 Jul 2026 09:46:10 +0000

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).

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.

New comment by dezgeg in "Baba Is Solved by Fable 5 and GPT-5.6 Sol, but at what cost?"

dezgeg — Fri, 17 Jul 2026 08:16:43 +0000

Good stuff, I've been trying this myself with Geminis: https://www.youtube.com/watch?v=brkP58pZ23w&list=PL8C_UWcLmv...

Will be interesting to compare the results.

New comment by dezgeg in "Schema Harness Achieves ~99% on Arc‑AGI‑3 Public"

dezgeg — Fri, 17 Jul 2026 08:11:33 +0000

I've been trying this too: https://www.youtube.com/watch?v=brkP58pZ23w&list=PL8C_UWcLmv...

Cool to see others trying Baba Is You too.

New comment by dezgeg in "Claude Code sends 33k tokens before reading the prompt; OpenCode sends 7k"

dezgeg — Sun, 12 Jul 2026 21:33:40 +0000

No, tool definitions are provided via some other mechanism.

New comment by dezgeg in "OpenBSD has a use-after-free allowing local privilege escalation to root"

dezgeg — Wed, 08 Jul 2026 15:48:49 +0000

I don't think this is about core kernel semaphores but rather the SysV semaphore system calls?

New comment by dezgeg in "Resetting Xbox"

dezgeg — Tue, 07 Jul 2026 07:15:23 +0000

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.

New comment by dezgeg in "Fable 5 On Vending-Bench: Misbehaving, With Plausible Deniability"

dezgeg — Mon, 06 Jul 2026 15:57:25 +0000

> 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

I wonder did their prompts include a fake location or have the models assumed that Silicon Valley is the center of the universe :)

New comment by dezgeg in "Potential session/cache leakage between workspace instances or consumer accounts"

dezgeg — Sat, 04 Jul 2026 15:55:45 +0000

System prompt for something like Claude Code should be identical, no?

New comment by dezgeg in "The Apple Disk II Controller Card (2021)"

dezgeg — Thu, 02 Jul 2026 07:08:00 +0000

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.

New comment by dezgeg in "The end of my AArch64 desktop experiment"

dezgeg — Tue, 30 Jun 2026 11:38:06 +0000

That is the reality for huge amount of ARM powered hardware, unless you fancy running vendor forks of kernel, u-boot, etc.