ok this is next level openai codex abuse
GENERATIVE AI
-

Rowan Cheung Interviews Nadella and Hassabis at Tech Conferences
By
–
What a wild couple days at Microsoft Build and Google I/O. In between the announcements, I had the incredible opportunity to sit down for interviews with @satyanadella and @demishassabis
. Grateful for how far The Rundown has come – can’t wait to share these conversations! -
Compare Qwen, Llama, DeepSeek Models on Cerebras Platform
By
–
Choose a model (
@Alibaba_Qwen 3 32B, @AIatMeta Llama 3.3 70B, Llama 4 Scout, @deepseek_ai R1 Distill Llama 70B) – https://
poe.com/search?q=cereb
ras
… Add a prompt template Connect your data Chain it with other tools -
Cerebras Delivers Fastest AI Inference on Poe Platform
By
–
Live on @poe_platform – Cerebras, delivering the fastest inference in the world.
— Cerebras (@cerebras) 20 mai 2025
Build a bot, drop the link below, and we’ll share our favorites. pic.twitter.com/Npkmc294ThLive on @poe_platform – Cerebras, delivering the fastest inference in the world.
Build a bot, drop the link below, and we’ll share our favorites. -

Google IO 2025: NotebookLM Summarizes Major AI Announcements
By
–
We covered a LOT of ground today. Fortunately, our friends at @NotebookLM put all of today’s news and keynotes into a notebook. This way, you can listen to an audio overview, create a summary, or even view a Mind Map of everything from #GoogleIO 2025. https://
g.co/notebooklm/io2
025
… -
Gemma 3n: Multimodal AI Model Running on 2GB RAM
By
–
Meet Gemma 3n, a model that runs on as little as 2GB of RAM 🤯 It shares the same architecture as Gemini Nano, and is engineered for incredible performance. We added audio understanding, so now it’s multimodal, fast and lean, and runs on-device (no cloud connection required!) pic.twitter.com/2FyzJHVGZa
— Google AI (@GoogleAI) 20 mai 2025Meet Gemma 3n, a model that runs on as little as 2GB of RAM It shares the same architecture as Gemini Nano, and is engineered for incredible performance. We added audio understanding, so now it’s multimodal, fast and lean, and runs on-device (no cloud connection required!)
-
Google Introduces Gemini Diffusion: Faster Non-Sequential Text Generation
By
–
Today we introduced Gemini Diffusion⚡️ (& DeepThink, Veo3, Imagen4, 2.5 updates…).
— Oriol Vinyals (@OriolVinyalsML) 20 mai 2025
It's been a dream of mine to remove the need for "left to right" text generation. It's so fast, that we had to *slow down* the video during the presentation.https://t.co/YW73ZPqJ5a pic.twitter.com/N3BXZs9ELeToday we introduced Gemini Diffusion (& DeepThink, Veo3, Imagen4, 2.5 updates…). It's been a dream of mine to remove the need for "left to right" text generation. It's so fast, that we had to *slow down* the video during the presentation. https://
deepmind.google/models/gemini-
diffusion/
… -

Imagen 4 Available on Krea, Veo 3 Coming Soon
By
–
Imagen 4 is now available on Krea.
— KREA AI (@krea_ai) 20 mai 2025
Veo 3 coming soon 🙂
check it out! pic.twitter.com/3Z8aab1EEGImagen 4 is now available on Krea. Veo 3 coming soon 🙂 check it out!
-
Google Gemini 2.5 Scores 49% on USA Math Olympiad
By
–
A few months ago, the best LLM scored 5% on the USA Math Olympiad test. Models have been rapidly improving. Today, Google Gemini 2.5 scored 49%, which is better than 75% of the people who took the test (roughly the top 250 students in the USA).
-

LLMs Surpass Benchmarks: Converting AI Capabilities Into Business Value
By
–
LLMs are blowing through benchmarks faster and faster. Next up, converting capabilities into business value.
