az226's comments

On: Muse Wonder

[az226]

Yes

On: Beam: Reflection's 501B open-weight model

[az226]

Seems rather sloppy to not validate it wasn’t in their training or RL data.

On: Why Is Sam Altman a Free Man?

[az226]

Techno feudalism.

On: GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price

[az226]

6x

On: GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price

[az226]

Mature training pipelines, plus ever expanding RL datasets of increased quality, and mega GPU clusters to finish training in a few weeks. Automated safety and reliability testing.

On: Show HN: Air-gapped file encryption as self-decrypting HTML page

[az226]

Offline would be a better word.

On: Virtio-nvgpu: Near-native Nvidia GPU access inside a KVM guest

[az226]

GeForce GPUs don’t support MIG. Only workstation and data center cards do that.

On: Virtio-nvgpu: Near-native Nvidia GPU access inside a KVM guest

[az226]

If memory serves, there is a way to hack the drivers to unlock MIG for GeForce GPUs.

On: Claude Opus 5.5 Intelligence, Performance and Price Analysis (Max)

[az226]

How did you get the reasoning trace? Is it the actual one or the summarized one?

[simonw → az226]

It's the summarized one returned by their API.

Piping the visible reasoning trace through their token counter API (I use https://tools.simonwillison.net/claude-token-counter for that) counts 27,888 tokens, so it's definitely a summary of the 128,000 actual token trace.

On: We got admin access to Baseten's production GitHub

[az226]

And yet we see no prosecution for OpenAI’s felonies on HF.