Hacker News: villgax

New comment by villgax in "Muse – Meta’s personal AI agent"

villgax — Wed, 09 Sep 2026 06:04:24 +0000

Cant wait for Snitch Bench results of this

New comment by villgax in "Muse – Meta’s personal AI agent"

villgax — Wed, 09 Sep 2026 03:15:32 +0000

This gon snitch so much

New comment by villgax in "Cognition (Devin) raises $2B at $48B valuation"

villgax — Wed, 09 Sep 2026 02:11:23 +0000

Lol who even uses this

New comment by villgax in "Nvidia agrees to acquire Hugging Face for $13B"

villgax — Thu, 27 Aug 2026 06:57:53 +0000

As opposed to whom else? Do you think any of the other labs or acccelerators are clamoring for open source hardware?

New comment by villgax in "OpenAI Jalapeño: Better than Nvidia Blackwell"

villgax — Wed, 26 Aug 2026 15:01:29 +0000

obviously, Anthropic is gonna do their own chips as well just like openai, all hyperscalers already have their own chip stacks & neoclouds wouldnt even carry this as just as they are not carrying AMD stack lol

New comment by villgax in "OpenAI Jalapeño: Better than Nvidia Blackwell"

villgax — Wed, 26 Aug 2026 06:54:38 +0000

Basically the EvLeaks of AI lol

New comment by villgax in "Show HN: Make your logo extra bright on HDR screens"

villgax — Sun, 23 Aug 2026 07:59:07 +0000

Assaults the senses takes a new example now.

New comment by villgax in "Codex in ChatGPT desktop app for Linux is now in preview"

villgax — Thu, 13 Aug 2026 07:01:50 +0000

people good w/ linux would not have been holding their breath for companies which keep intelligence for a premium & h0st@ge depending on policy

New comment by villgax in "Running Kimi K3 on MI355X at Better Performance per Dollar Than B300"

villgax — Sun, 02 Aug 2026 14:09:49 +0000

No mention of Infiniband or SPX lol

New comment by villgax in "Running Kimi K3 on MI355X at Better Performance per Dollar Than B300"

villgax — Sun, 02 Aug 2026 07:52:15 +0000

I went the gpus.io website & it’s $2.95/hr right now, this is like comparing MSRP to actual market price. B300s are in demand & hence cost more, but these lazy editors at wafer.ai can't be bothered to do TCO on actual ownership nor share code to replicate their setups. Instead just relying on current market prices to win one row, which isnt even about per/$ on actual MSRPs.

New comment by villgax in "Running Kimi K3 on MI355X at Better Performance per Dollar Than B300"

villgax — Sun, 02 Aug 2026 06:50:31 +0000

Lol, such a lazily written article by wafer.ai

GPUs. 8× MI355X (TP8) B300 (TP8+DCP8)

Decode tok/s per stream 118 tok/s 172 tok/s

Peak aggregate. 952 tok/s 1,568 tok/s

Peak aggregate per GPU 119 tok/s 196 tok/s

On every row the B300 beat the MI355X

The B200 is being forcefully compared against something which is not gonna fit within it's memory in a single node & not much details about multi-node interconnectivity, disagg or not. As expected of a shoddy slop.

The only point it won is of cost per hour is one aggregation website for rentals, the premium a B300 commands against the $3/hr AMD chip which no provider has in abundance. Never bothered to do TCO of owning the hardware either.

New comment by villgax in "India's first privately-developed rocket reaches orbit on debut launch"

villgax — Sun, 02 Aug 2026 04:29:37 +0000

Checks notes, the same exact line of commentary that exists for Spacex themselves

New comment by villgax in "RipGrep musl binaries occasionally segfault during very-large searches"

villgax — Sun, 02 Aug 2026 04:28:47 +0000

Actually read the post before ignoramusing here

New comment by villgax in "RipGrep musl binaries occasionally segfault during very-large searches"

villgax — Sat, 01 Aug 2026 14:02:37 +0000

No wonder search in codex is so a$$

New comment by villgax in "Postgres LISTEN/NOTIFY actually scales"

villgax — Sat, 25 Jul 2026 04:00:52 +0000

Even AI generated art needs taste, looks so bland & off-point.

New comment by villgax in "Show HN: Make senders work to get into your inbox"

villgax — Sat, 25 Jul 2026 03:26:49 +0000

I would not even go to the extent of pay to call, just pay to be let through, regardless of whether I pick up or call. Don’t even want to get into chargebacks coz I didnt pickup lol.

New comment by villgax in "India's first privately-developed rocket reaches orbit on debut launch"

villgax — Fri, 24 Jul 2026 18:08:52 +0000

you can imagine if they bother cracking landing back to re-use, it's KO for spacex

New comment by villgax in "Make people pay to get into your inbox"

villgax — Wed, 15 Jul 2026 12:30:15 +0000

This should be the case even for phones as well

New comment by villgax in "Alternative(s) to run CUDA on non-Nvidia hardware"

villgax — Tue, 14 Jul 2026 13:09:51 +0000

@claude add this to the graveyard of wannabees

New comment by villgax in "Flash-MSA: Accelerating Million-Token Training with Sparse Attention Kernels"

villgax — Mon, 13 Jul 2026 02:55:37 +0000

World’s first?

Such lazy, much farming

https://github.com/fla-org/native-sparse-attention?utm_sourc...