Skeptic Calls Elite Startup's Infinite Context Claims Post-Trained Model Slop
Reactions
2 postsAD 
Alexander Doria@Dorialexander"this hidden elite startup has solved infinite context/continual learning/tokenizers/pdf" => a glm post-train or bizarro merge on mid data.
- likes: 25
- replies: 1
- bookmarks: 2
- reposts: 1
AD Alexander Doria@Dorialexanderso today is 5 trillion "tokens" length.
- likes: 25
- replies: 2
- bookmarks: 5
- reposts: 1