Open Source Fights Back: Llama 4, Qwen 3, and the Democratization of AI
Meta releases Llama 4 with 400B parameters fully open, Alibaba counters with Qwen 3. We explore how open source is closing the gap with proprietary models and what it means for developers.
Yann LeCunThomas Wolf
Llama 4MetaQwen 3
GPT-5 First Impressions: What Changed and What Didn't
OpenAI finally ships GPT-5 and the AI world reacts. We break down the benchmarks, the new reasoning capabilities, the multimodal upgrades, and whether it lives up to the hype. Plus: what this means fo
Sam AltmanAndrej Karpathy
GPT-5OpenAIbenchmarks
2025 Yearly Recap - Someone Trained an LLM in Space This Year (And 50 Other Things You Missed)
The ultimate 2025 AI recap. From LLMs trained in space to the open-source revolution, we cover the 50 most important AI developments of 2025. Nearly 2 hours of pure AI history.
Nisten TahirajYam PelegWolfram RavenwolfLDJ
+1 more
yearly recap2025open source
ThursdAI - Dec 11 - GPT 5.2 Is HERE! Plus LLMs in Space, MCP Donated, Devstral Surprises
GPT 5.2 launches! Also: the first LLM inference run in space via satellite. Anthropic donates MCP to open governance. Mistral's Devstral coding model surprises everyone.
Nisten TahirajYam PelegLDJArthur Mensch
OpenAIGPT-5space
ThursdAI - Dec 4 - DeepSeek V3.2 Goes Gold Medal, Mistral Returns to Apache 2.0
DeepSeek V3.2 achieves gold medal on competitive programming benchmarks. Mistral announces return to Apache 2.0 licensing for their models. The open-source AI movement accelerates.
Nisten TahirajYam PelegWolfram RavenwolfClem Delangue
DeepSeekMistralopen source