Hacker News: brrrrrm

New comment by brrrrrm in "AI helps add 10k more photos to OldNYC"

brrrrrm — Tue, 07 Apr 2026 17:49:07 +0000

I checked 3 spots I'm familiar with and 1 is wrong

https://www.oldnyc.org/#707133f-a this is supposed to be here https://www.oldnyc.org/#702487f-a

also, if folks are interested in these old depictions of NYC, check out https://1940s.nyc/ as well!

How to Fail as an Organization in 2026

brrrrrm — Tue, 27 Jan 2026 19:18:31 +0000

Article URL: https://jott.live/markdown/how_to_fail_2026

Comments URL: https://news.ycombinator.com/item?id=46784934

Points: 2

# Comments: 0

Internal Combustion Engine Acoustic Synthesis

brrrrrm — Sat, 27 Dec 2025 17:56:51 +0000

Article URL: https://jott.live/html/eng_sound

Comments URL: https://news.ycombinator.com/item?id=46403679

Points: 2

# Comments: 0

New comment by brrrrrm in "Show HN: CineCLI – Browse and torrent movies directly from your terminal"

brrrrrm — Tue, 23 Dec 2025 07:15:55 +0000

looks cool! one bit of feedback: make your demo gif get to the point faster. either practice typing a bit quicker or speed it up 2x for the typing section

Show HN: Binfer, an experimental LLM inference engine in TypeScript and CUDA

brrrrrm — Wed, 10 Dec 2025 15:50:29 +0000

Article URL: https://github.com/bwasti/binfer

Comments URL: https://news.ycombinator.com/item?id=46219142

Points: 1

# Comments: 0

New comment by brrrrrm in "Anthropic acquires Bun"

brrrrrm — Tue, 02 Dec 2025 20:54:08 +0000

on Bun's website, the runtime section features HTTP, networking, storage -- all are very web-focused. any plans to start expanding into native ML support? (e.g. GPUs, RDMA-type networking, cluster management, NFS)

Five Times Faster

brrrrrm — Mon, 24 Nov 2025 13:59:19 +0000

Article URL: https://jott.live/markdown/five_times_faster

Comments URL: https://news.ycombinator.com/item?id=46034164

Points: 2

# Comments: 0

Bitwise Consistent On-Policy Reinforcement Learning with VLLM and TorchTitan

brrrrrm — Wed, 12 Nov 2025 16:58:01 +0000

Article URL: https://blog.vllm.ai/2025/11/10/bitwise-consistent-train-inference.html

Comments URL: https://news.ycombinator.com/item?id=45902538

Points: 1

# Comments: 0

New comment by brrrrrm in "Spatial intelligence is AI’s next frontier"

brrrrrm — Tue, 11 Nov 2025 03:48:41 +0000

we've discovered some kind of differentiable computer[1] and as with all computers, people have their own interests and hobbies they use them for. but unlike computers, everyone pitches their interest or hobby as being the only one that matters.

[1] https://x.com/karpathy/status/1582807367988654081

New comment by brrrrrm in "Helion: A high-level DSL for performant and portable ML kernels"

brrrrrm — Fri, 07 Nov 2025 14:12:37 +0000

a recent wave of interest in bitwise equivalent execution had a lot of kernels this level get pumped out.

new attention mechanisms also often need new kernels to run at any reasonable rate

theres definitely a breed of frontend-only ML dev that dominates the space, but a lot novel exploration needs new kernels

Should we apply old-school multi-core scheduling to GPUs?

brrrrrm — Mon, 03 Nov 2025 18:17:25 +0000

Article URL: https://jott.live/markdown/gt

Comments URL: https://news.ycombinator.com/item?id=45802300

Points: 4

# Comments: 0

Show HN: GT: experimental multiplexed distributed tensor framework

brrrrrm — Sun, 02 Nov 2025 23:12:46 +0000

Article URL: https://github.com/bwasti/gt

Comments URL: https://news.ycombinator.com/item?id=45794253

Points: 4

# Comments: 0

GT – Experimental multiplexing tensor framework for distributed GPU computing

brrrrrm — Sun, 02 Nov 2025 07:16:20 +0000

Article URL: https://github.com/bwasti/gt

Comments URL: https://news.ycombinator.com/item?id=45788441

Points: 30

# Comments: 1

New comment by brrrrrm in "John Carmack on mutable variables"

brrrrrm — Fri, 31 Oct 2025 03:44:26 +0000

one thing I've learned in my career is that escape hatches are one of the most important things in tools made for building other stuff.

dropping down into the familiar or the simple or the dumb is so innately necessary in the building process. many things meant to be "pure" tend to also be restrictive in that regard.

New comment by brrrrrm in "Poker Tournament for LLMs"

brrrrrm — Tue, 28 Oct 2025 14:56:28 +0000

> None of which are they currently capable

what makes you say this? modern LLMs (the top players in this leaderboard) are typically equipped with the ability to execute arbitrary Python and regularly do math + random generations.

I agree it's not an efficient mechanism by any means, but I think a fine-tuned LLM could play near GTO for almost all hands in a small ring setting