This past week has been packed with breakthroughs, and in this video, I covered three that really stood out:
1. GPT-4.5 is here! OpenAI’s latest model, previewed for Pro users, promises sharper reasoning, fewer hallucinations, and a big leap in efficiency. It’s inching us closer to AGI, but also hints that the era of massive models might soon give way to something different.
2. NVIDIA is unstoppable. The company’s earnings soared yet again, with data center revenue hitting an eye-watering $36 billion. Jensen Huang believes AI’s next frontier—reasoning, agentic AI, and robotics—will demand even more compute. And looking at the industry’s direction, he might be right.
Read more: https://nvidianews.nvidia.com/news/nvidia-announces-financial-results-for-third-quarter-fiscal-2025
3. A new way to build AI? A startup called Inception Labs has introduced DLMs (Diffusion Language Models), a fresh approach to language modeling inspired by image and video diffusion models. The big idea? Parallel processing instead of sequential token generation, potentially making AI training way more efficient. If this works at scale, we might be looking at a paradigm shift.
Read more: https://techcrunch.com/2025/02/26/inception-emerges-from-stealth-with-a-new-type-of-ai-model/
