Assagman

Programmer | Composer | Human

[Assagman @assagman]

This platform deliberately censors free expression.

[Rhys @RhysSullivan]

2026

[Assagman @assagman]

open a pr and anthr0p*c switches to Claude 3 Haiku

[Gergely Orosz @GergelyOrosz]

Things I really dislike about Fable:

1. Anthropic collects my prompt history, stores it, and does whatever they want with it for 30 days. No opt-out

2. They can nerf their most expensive model without telling me, billing me the same amount, wasting my time. Whenever they want

[ThePrimeagen @ThePrimeagen]

2nd tinfoil hat time

The entire theater around "too dangerous" / reducing in real time models ability to 4.8 / storing all the prompts for 30 days is not for safety at all but a ploy to figure out who is the Chinese spies "stealing" their weights...

[Assagman @assagman]

@lucasmeijer project’s root AGENTS: Read foo/bar/AGENTS

[Assagman @assagman]

@sqs @AmpCode llama 3.3

[jordi @jordienr]

the reality

[Assagman @assagman]

haha

[Darren Shepherd @ibuildthecloud]

This is awesome. https://t.co/2E1rNaWULE

[Assagman @assagman]

@superdoteng Looks like it's having trouble to open a project from the bare git repo root from local fs.

I needed to import projects from a worktree (i.e. main), but in that case project name displayed as "main", which is confusing cause there are lots of them :)

[Assagman @assagman]

@alert090579992 @OpenAI @badlogicgames I guess it's already a known issue for OpenAI. I remember @thsottiaux mentioned this to be fixed in the future.

[Theo - t3.gg @theo]

Struggling to pick what agent, model, and effort levels to use? Miss the "slot machine" feel of Claude Code when using other tools?

`npx slotslop "[prompt]"`

[Assagman @assagman]

@badlogicgames this can be applied to all tools. ultra compact mode is useful for some cases.

[Assagman @assagman]

@nbaschez what makes this better than @plannotator , in terms of capabilities?

[Assagman @assagman]

@TE701 @antigravity disable your vpn

[Assagman @assagman]

canın sağolsun be karpati

[Andrej Karpathy @karpathy]

Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.

[Assagman @assagman]

@fkadev veri datası ne demek oluyor?

[Assagman @assagman]

@scaling01 mythos doritos burritos

[Figure @Figure_robot]

Day 2 is Live: Watch humanoid robots Bob, Frank, and Gary running 24/7. This is fully autonomous running Helix-02 https://t.co/dsacH4HOzJ

[Alexander Whedon @alex_whedon]

Introducing SubQ - a major breakthrough in LLM intelligence.

It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA),

And the first frontier model with a 12 million token context window which is:

- 52x faster than FlashAttention at 1MM tokens
- Less than 5% the cost of Opus

Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention).

Only a small fraction actually matter.

@subquadratic finds and focuses only on the ones that do.

That's nearly 1,000x less compute and a new way for LLMs to scale.

[Assagman @assagman]

@fkadev https://t.co/5aM3rADxkn

[Assagman @assagman]

@TheAhmadOsman before they release gemini 2.5, all people were pretty confident that google was cooked and it will be out of AI race, but it wasn’t the case.

They ship slow in terms of LLM, but their surface is too wide. image, audio, video, workspace products, mobile etc.