Speed or quality? Yes.
— Groq Inc (@GroqInc) 26 septembre 2025
Compound on Groq: GPT5: pic.twitter.com/yg6pLe9Nu3
Speed or quality? Yes. Compound on Groq: GPT5:
By
–
Speed or quality? Yes.
— Groq Inc (@GroqInc) 26 septembre 2025
Compound on Groq: GPT5: pic.twitter.com/yg6pLe9Nu3
Speed or quality? Yes. Compound on Groq: GPT5:

By
–
This is going to be a fun panel. We'll be talking about both the positive and negative impacts of AI on content creation. If you're at VidSummit in a couple weeks, definitely join us for this panel!
By
–
In terms of how it was made – nothing complicated, just ask Codex to make a visualisaiton in HTML, or you can ask for React too, but a bit more complex to run. In this case it is using D3 js libary. I usually ask for light colours (since LLMs like dark neon ones for some reason).
By
–
they didnt train new base model for 5 right what makes you so sure they will for 6 regardless my claim wasnt about new pretrains nor was it about timeline of 6 just enumerating what they clearly see as valuable for next steps to solve gdpval which is in my mind synonymous
By
–
multi model is a big part of the Agent Labs thesis
— swyx 🐣 (@swyx) 26 septembre 2025
havent published findings but yeah you kinda want to argmax harness(capability)
so if you believe that capability is a high dimensional space (as in, “good at coding” is not enough, it breaks down to a hundred subtasks of… https://t.co/8yO4HdtbjF
multi model is a big part of the Agent Labs thesis havent published findings but yeah you kinda want to argmax harness(capability) so if you believe that capability is a high dimensional space (as in, “good at coding” is not enough, it breaks down to a hundred subtasks of
By
–
step-by-step LLM Engineering Projects each project = one concept learned the hard (i.e. real) way Tokenization & Embeddings > build byte-pair encoder + train your own subword vocab
> write a “token visualizer” to map words/chunks to IDs
> one-hot vs learned-embedding: plot
By
–
As Jensen mentioned with @altcap @BG2Pod @bgurley, something that few people know is that @nvidia is becoming the American open-source leader in AI, with over 300 contributions of models, datasets and apps on @huggingface in the past year.
— clem 🤗 (@ClementDelangue) 26 septembre 2025
And I have a feeling they're just… pic.twitter.com/ko9oTeGiZ8
As Jensen mentioned with @altcap @BG2Pod @bgurley
, something that few people know is that @nvidia is becoming the American open-source leader in AI, with over 300 contributions of models, datasets and apps on @huggingface in the past year. And I have a feeling they're just
By
–
Where do those training TFLOP figure come from? I didn't know OpenAI shared numbers like that
By
–
BREAKING 🚨: Perplexity is testing a new "Sonar Testing" reasoning model internally. Potentially, it will arrive as a reasoning upgrade to the existing Sonar model.
— 🚨 AI News | TestingCatalog (@testingcatalog) 26 septembre 2025
Imo, "Sonar Testing" is the best name ever 👀 pic.twitter.com/a1ChGPDGLh
BREAKING : Perplexity is testing a new "Sonar Testing" reasoning model internally. Potentially, it will arrive as a reasoning upgrade to the existing Sonar model. Imo, "Sonar Testing" is the best name ever

By
–
NVIDIA's Vice President of Generative AI Software, Kari Briski, joined Canada's Minister of Artificial Intelligence and Digital Innovation, Evan Solomon, and Cohere's Cofounder and CEO, Aiden Gomez, for a special address moderated by SiriusXM host Amber Mac at All In Canada AI