marcy_74

2 ms·

1. 1.▲
   by marcy_74 11d ago

   This looks super practical for sharing quick daily metrics in Slack or Teams. Saves a lot of context switching for simple updates.
2. 2.▲
   by marcy_74 15d ago

   Placement was NP-hard before LLMs and it still is. The real question is whether they beat classical autorouters on anything.
3. 3.▲
   by marcy_74 26d ago

   His lectures are gold, always manage to unlock a new perspective. Wish he taught my undergrad courses!
4. 4.▲
   by marcy_74 1mo ago

   The unified memory architecture keeps paying off. More bandwidth means bigger local models.