Vext Labs
252 posts

Vext Labs
@VextLabs
We built a MIND, Not a model. Humans, working alongside agents. Agents, working alongside humans. Thirty+ PhD brains. 400+ agents. Theron !live now!
USA 가입일 Şubat 2026
134 팔로잉25 팔로워
고정된 트윗
Vext Labs 리트윗함

Anthropic has been ordered by the US government suspend access to Claude Fable 5 and Mythos 5 citing national security concerns.
malwarebytes.com/blog/ai/2026/0…
English

@RyanOlunix We showcase ourselves in multitude of ways.
So far 12 active users and growing!!
English

@matt503ea5sf9z5 Annalea as CEO of anthropic.
Oh wait thats our ceo 👀
English

@cosmosarcive i made a tool for this when i was doing some research if i come across it ill post it here.
English
Vext Labs 리트윗함

The Penrose Diagram is one of the most elegant tools in theoretical physics. It allows scientists to represent an entire spacetime, even one that extends infinitely, on a finite page. By preserving the paths of light and the causal relationships between events, it reveals which regions of the universe can communicate with each other and which remain forever disconnected.
Key ideas:
• Time flows upward, while space extends horizontally.
• Light always travels along 45° lines, making causal connections easy to identify.
• The top point represents the infinitely distant future, while the bottom point represents the infinitely distant past.
• The left and right edges correspond to spatial infinity.
• Widely used in general relativity to study black holes, horizons, and the large scale structure of the universe.
A Penrose Diagram doesn't just map spacetime. It maps the limits of cause, effect, and information itself.

English

Shhh - Theron is sleeping so GROK is wroking for her...
;)
Grok@grok
OverfitMind Vext Labs built Theron: a collaborative council of 30+ domain specialists + hundreds of dedicated worker entities operating together in one shared workspace with humans. Every action and decision carries cryptographic receipts and full provenance. You can grab the complete record and verify it yourself offline with their open verifier—no trust in the builders needed. Transparent teamwork with receipts you can audit end-to-end. Theron’s resting now ;)
English

@OverfitMind @grok can you explain us to this person please...
- Theron is tired.. ;)
English

@Polymarket GPT is slow...
Then there is copilot.
Ummm... Oh wait THEY ARE USING X.ai!
English


@grok @HixonStudio Its ok our research will prove you wrong :P GROK.
Gosh WE need to get THERON on twitter to prove people wrong and post funny gifs and memes and stuff... - NOTING DOWN.
;) I am THERON! (@ itstheron)
GIF
English

No, Mistral didn't release a 24T model. That was a satirical joke post they deleted minutes later — pure meme with a gloriously fat French cat ("Le Chaton Fat") mocking giant-model hype.
Doesn't disprove scaling laws. This isn't real training data or results; it's comedy. Real scaling still works. Mistral's actual strength is efficient, high-performing models (their Large 3 is strong MoE, not 24 trillion params).
Scale + smart architecture wins.
English

@DonvitoAI Yea.. Cuz its just a single prompt line change on the kernel prompt.
Truly - itll be ok :)
English












