Hacker News

points by bigyabai 2 days ago | 0 comments

> How can people still not see this?

Perhaps the 30+ years of non-LLM generated Linux support for UEFI/ACPI and ARM device tree chips has something to do with it.

[–] bitwize 2 days ago

Developed with techniques that were state-of-the-art at the time.

The state of the art has changed.

[–] bigyabai 2 days ago

You still need documentation. Asahi on M1 is 7 years old, with no ACPI substitute.

It may shock you to hear that coding was never the bottleneck for emulation or reverse-engineering.

[–] bitwize 2 days ago

LLMs are arguably even more superhuman for the bits that do matter for reverse engineering, than they are at coding!

Again, any RE project that forbids AI use will get its milkshake drunk by a similar project that embraces AI.

[–] bigyabai 2 days ago

I'll believe it when I see it. As I said, there are still unsolved problems that AI can't fix taking shots in the dark.

I have no particular love for Asahi, but they've made 10x more progress than the vibe coded alternatives.

[–] orangecat 2 days ago

It may shock you to hear that coding was never the bottleneck for emulation or reverse-engineering.

Yes, and LLMs work well for the non-coding parts too. Pointing Claude at Gravity's repos has gotten Linux running well on my M4 Pro mini, with GPU acceleration as of a few days ago. So either I'm a genius or LLMs are pretty good at this, and I'm not a genius.

[–] bigyabai 2 days ago

The Honeykrisp driver and upstream M4 Linux support was written by Asahi, not Gravity: https://www.phoronix.com/news/Linux-7.4-SoCs

You're downstream of all of those kernel and Mesa changes on Gravity.

[–] cromka 2 days ago

> The state of the art has changed.

I keep repeating myself, but make an effort to check the "state of the art" of the crap they added to their kernel tree.