🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Apple Speaks the Truth About AI. It’s Not Good. by Ignacio de Gregorio link.medium.com/kyM0F3Ii5Nb

Apple Speaks the Truth About AI. It’s Not Good. by Ignacio de Gregorio link.medium.com/Nb84sVoQaOb

Physical AI Foundation Models Are Here by Ignacio de Gregorio link.medium.com/hauEUi4ViOb LLMs discover the laws of nature!!!!! #AI, #MachineLearning, #DataScience

The secret behind my productivity is that I'm almost exclusively a print-statement debugger

AI Manhattan Project Proposed. The US’ Suicide or Salvation? by Ignacio de Gregorio link.medium.com/YqfyeHvSUOb

“In other words, this report considers the development of AGI of similar strategic importance to the development of nuclear bombs.” — Ignacio de Gregorio link.medium.com/AbLffzCxZOb

The EU AI act is already making Europeans into second class digital citizens. Do not forget the con artists who tried to do this to California.

NVIDIA's stock being 15% down on the news about DeepSeek's R1 model just proves how clueless markets are about AI. If anything, the news should have caused NVIDIA's stock to rise. The panic is caused by DeepSeek v3, R1's base model, having been trained on just under $6 million,

Comparing the training of v3 with GPT-4 is like comparing the learning curve of a student learning by itself (GPT-4) with one that is using a teacher (ironically, GPT-4) to teach the student (v3). V3 is a distillation, not comparable. The funny thing is that the entire crash is

why did ChatGPT suddenly decide it's okay to overwhelm me with emojis

Deep Research, from both Google & OpenAI, marks the first time I am completely 99.9999% convinced that human level intelligence is truly unreachable using current artificial intelligence (AI) technology. Its time to rethink what AI is.

OpenRouter LLM usage leaderboard is here: openrouter.ai/rankings

an AI geek-out extraordinaire for a Saturday morning... "Here’s NVIDIA and AI’s Future in 5 Minutes." by Ignacio de Gregorio medium.com/p/heres-nvidia…

The Second ‘ChatGPT Moment’ is Here by Ignacio de Gregorio medium.com/p/the-second-c…

España tiene las segundas mayores reservas de uranio de la Unión Europea continental, como documenta el Libro Rojo del Uranio de la Agencia Internacional de la Energía. oecd-nea.org/upload/docs/ap… Sin embargo, la extracción de minerales radiactivos como el uranio está prohibida por

MatFormers are very powerful alternatives to transformers. Similar to a regular transformer, but after training, you can split up the model to any size you like and get very strong performance that scales just like a regular transformer. So train once, get models of all sizes!

frontier ai today

Been heavily testing MCP with local and proprietary models, and I'm worried about memory complexity. As long as the cache has high memory complexity, preventing larger context windows, even the simplest of MCP use cases saturate the context window before I can do anything (this