simulacra8

7 ms·

1. 1.▲
   Text Diffusion – Brendan O'Donoghue, Google DeepMind [video](youtube.com)

   3 points by simulacra8 24d ago|0 comments
2. 2.▲
   by simulacra8 24d ago

   dLLMs are still promising. It's a counterintuitive hardware limitation. You should watch this: https://www.youtube.com/watch?v=r305-aQTaU0
3. 3.▲
   by simulacra8 1y ago

   I doubt, old project https://github.com/uclaacm/gradient-descent-visualiser
4. 4.▲
   by simulacra8 1y ago

   what Anthropic did right with MCP that Google didn't with A2A?