Adnan
3K posts
Adnan
@BEAST_OFFICIIAL
Not an AI generated mirage.
Earth Katılım Ocak 2022
2.9K Takip Edilen215 Takipçiler
Adnan retweetledi
Love codex app!! It'd be cool to add
- Drag+drop files into sidebar to open them like an artifact
- Multiple browser tabs
- Add "Ask Codex" to the browser window
- Drag+drop chat from sidebar into context (instead of manually adding session-id to chat)
@thsottiaux
Tibo@thsottiaux
Send us feature requests for codex in the form of an images 2.0 generated image. It makes it easier for codex to implement if we decide to go for it. Saw some good ones today already that codex is cooking on.
English
what keeps haunting me is how intelligence emerges, evolves and eventually learns to boost itself🧐
Nick Levine@status_effects
New work with @AlecRad and @DavidDuvenaud: Have you ever dreamed of talking to someone from the past? Introducing talkie, a 13B model trained only on pre-1931 text. Vintage models should help us to understand how LMs generalize (e.g., can we teach talkie to code?). Thread:
English
Hey guys I don't know if everyone will see my post or not
Last few weeks I had panic / anxiety attacks , ADHD and high BP and hypothyroidism, life is so hard man
I wanna cry , but unable to , even for crying a human needs a safe space what a dilemma .
I hope AI will solve all these health problems, i don't know why others are excited about AI, but for me this is to solve humanities diseases.
I don't know why but I have seen ai as a hope to solve this , sorry my english it's my third language and if anybody ever feel alone , lonely , depressed I am there for you , text me anytime , I want to be the one that I want other to be .
In the end , I want to spread love , want to make a society where individuality wins over groupism
English
Adnan retweetledi
IMO DeepSeek v4 demonstrated utter confidence and competence by not benchmaxxing, not focusing on some BS final run cost, not even spending inference-optimal compute.
just showed up, demonstrated SOTA long context efficiency techniques (CSA, HCA, mHC, flash at 8% cost of pro, which itself is 14% cost of opus), dropped the best open base models in the world, peaced out.
BYO posttraining. leave that to the agent labs to pick up the scraps. bravo.
English
Adnan retweetledi
Hello @karpathy I know that a ton of work went into making this video, but it would be insanely valuable if you make a “Deep dive into AI agents like OpenClaw” or whatever agent tools you’re using now. It was fun listening to you talk a little bit about this on the @NoPriorsPod, but the videos where you’re showing + explaining these concepts are gold.
English
Can't even imagine what we will get by the end of this year.
Claude@claudeai
Claude now connects to the tools creative professionals already use. With the new Blender connector, you can debug a scene, build new tools, or batch-apply changes across every object, directly from Claude.
English
Adnan retweetledi
Please @thsottiaux I beg you, make 2k high image gen available in Codex even if it drains usage limits faster, please give us the option
English
@BEAST_OFFICIIAL @jxnlco I encountered the same issue before. You can simply ask Codex to resolve it by showing him the error in a new chat
English
Adnan retweetledi
Xiaomi MiMo-V2.5 is now officially open-sourced！
MIT License, supporting commercial deployment, continued training, and fine-tuning - no additional authorization required.
Two models, both supporting a 1M-token context window :
• MiMo-V2.5-Pro: built for complex agent and coding tasks, ranking No.1 among open-source models on GDPVal-AA and ClawEval
• MiMo-V2.5: a native omni-modal model with strong agent capabilities
A model's value isn't measured by rankings alone — it's measured by the problems it solves.
Let's build with MiMo now!
🤗 Weights: huggingface.co/collections/Xi…
📄 Blog: #blog" target="_blank" rel="nofollow noopener">mimo.xiaomi.com/index#blog
English
interestingly, codex is currently the best at using cursors
simran sachdeva@simranrambles
wonder how will Elon rename cursor, considering both xcode and codex are taken
English
So the question is will we get more rapid improvements in this "smarter tokens" direction?
Adam.GPT@TheRealAdamG
The real story of GPT-5.5 for me is simply: “smarter tokens”. At the core of that is “token efficiency”. That token efficacy leads to higher “intelligence density” per token. Which leads to faster, (often) cheaper and most importantly — higher quality, more accurate outputs for you. A true platform for intelligence.
English
