Hacker News: snvzz

New comment by snvzz in "Hax – a minimalist, terminal-native coding agent written in C"

snvzz — Fri, 14 Aug 2026 15:15:16 +0000

The LLM can run elsewhere, such as the cloud.

For the part that runs locally, efficiency might matter.

New comment by snvzz in "Zapscape (CVE-2026-64561): Guest-to-Host Escape in KVM/x86"

snvzz — Fri, 07 Aug 2026 11:34:31 +0000

In seL4, a VMM escape, hard as that'd be, yields nothing.

VMM handles all VM exceptions, and is just another user program. It has no higher capabilities than the VM itself.

New comment by snvzz in "How to Make a Nintendo 64 Game in 2026"

snvzz — Thu, 06 Aug 2026 23:12:12 +0000

Eh, you'd have to pay me to use anything from Analogue.

ModRetro put their effort where it matters[0], whereas Analogue aren't even honest about their hardware, with baseless claims of 100% compatibility, recurring shipping delays or screwing preorders out of picking case shells. Which can't even be readily replaced.

0. https://www.youtube.com/watch?v=9gOJUEcKYRk

New comment by snvzz in "Amiga clarifies Commodore rights and the future of AmigaOS"

snvzz — Thu, 06 Aug 2026 05:04:19 +0000

"open source" is the string that's missing in that announcement.

The historical source code (AmigaOS git history up to at least when Commodore went bankrupt) should be MIT and in e.g. github.

But that doesn't seem to be what they have in mind.

The shitshow will likely continue.

New comment by snvzz in "Point AI at a "Proven" Microkernel?"

snvzz — Thu, 06 Aug 2026 05:00:11 +0000

seL4 is open source and pretty small.

What's stopping you?

Why Were the Dabaos Late?

snvzz — Mon, 03 Aug 2026 15:05:07 +0000

Article URL: https://www.crowdsupply.com/baochip/dabao/updates/why-were-the-dabaos-late

Comments URL: https://news.ycombinator.com/item?id=49156757

Points: 1

# Comments: 0

New comment by snvzz in "Show HN: I worked on a new browser for 2 years, today it passed Acid 3"

snvzz — Sat, 01 Aug 2026 05:27:26 +0000

No source code, no binary, just a page with some claims about code that might even exist.

Eh.

New comment by snvzz in "Nano Banana 2 removed from Google Earth"

snvzz — Sat, 01 Aug 2026 05:14:10 +0000

Censorship is now being called `guardrails`.

New comment by snvzz in "ESP32-C6 Power Consumption: Arduino vs. Zephyr vs. ESP-IDF Comparison"

snvzz — Sat, 01 Aug 2026 05:12:22 +0000

RISC-V conveniently provides a `wfi` (wait for interrupts) instruction or hint -- it can act as a nop when not implemented.

New comment by snvzz in "ESP32-C6 Power Consumption: Arduino vs. Zephyr vs. ESP-IDF Comparison"

snvzz — Thu, 30 Jul 2026 09:05:26 +0000

As of the recent launch of the ESP32-S31, all the ESP32 product lines have been migrated to RISC-V.

Unsurprisingly, the old, non-standard, encumbered architecture is not missed.

ESP32-S31 now in mass production and available for purchase

snvzz — Tue, 28 Jul 2026 12:35:35 +0000

Article URL: https://blog.adafruit.com/2026/07/27/esp32-s31-now-in-mass-production-and-available-for-purchase/

Comments URL: https://news.ycombinator.com/item?id=49082879

Points: 3

# Comments: 0

New comment by snvzz in "About the security content of macOS Tahoe 26.6"

snvzz — Tue, 28 Jul 2026 12:16:47 +0000

If anything, there's a strong argument to switch to seL4.

New comment by snvzz in "The story behind APC-2"

snvzz — Fri, 24 Jul 2026 16:26:30 +0000

To record audio, sample it with an ADC. Then store digitally.

New comment by snvzz in "Amiga 1000: Ten years ahead of its time"

snvzz — Fri, 24 Jul 2026 04:32:13 +0000

>For one, it had a forbid() system call which would stop the scheduler from pre-empting the current running process.

This is, by the way, equivalent to SCHED_FIFO on Linux.

There are very legitimate reasons a hard realtime task may need to be non-interruptable.

New comment by snvzz in "Geekbench 7"

snvzz — Fri, 24 Jul 2026 04:29:41 +0000

The CUDA addition (a vendor-specific API) is definitely a mistake.

New comment by snvzz in "New Framework Desktop Option with AMD Ryzen AI Max+ Pro 495 and 192GB Memory"

snvzz — Thu, 23 Jul 2026 14:09:23 +0000

Sounds very Whataboutism.

This is about the latter, and not just the correctable ones; Detecting them also has worth.

New comment by snvzz in "New Framework Desktop Option with AMD Ryzen AI Max+ Pro 495 and 192GB Memory"

snvzz — Thu, 23 Jul 2026 14:07:32 +0000

>the thing about ECC memory is that it doesn't support reporting of ECC errors

Wait that. It does, and that's the point. Although the CPU and motherboard do of course also have to support it, but with AMD it is possible. I do it on a 9800x3d.

One bit errors are always both detected and corrected. Multibit errors are mostly at least detected.

New comment by snvzz in "New Framework Desktop Option with AMD Ryzen AI Max+ Pro 495 and 192GB Memory"

snvzz — Thu, 23 Jul 2026 12:31:47 +0000

Non-ECC memory.

And into the trash it goes.

Refer to https://news.ycombinator.com/item?id=48969530

New comment by snvzz in "Amiga 1000: Ten years ahead of its time"

snvzz — Thu, 23 Jul 2026 12:29:49 +0000

The keyword being latency.

Even a powerful PC, such as running WinNT or Linux on my 9800x3d, doesn't feel as tight as my pretty basic Amiga 500 does.

It's the difference between a RTOS proper (AmigaOS, QNX, BeOS) and a toy (Windows, Linux, MacOS).

New comment by snvzz in "Everyone should know SIMD"

snvzz — Thu, 23 Jul 2026 01:10:17 +0000

Easier than ever, with RISC-V Vector (RVV), which is part of RVA23.