First impression of Fable: a lot more concise out of the box. Example of asking it to prototype optimizer updates https://github.com/yaroslavvb/learning-rotations
5:54 PM · Jun 9, 2026 · 4K Views
Sentiment
Users praise Method 3 from Fable AI as a good example of non-gradient-descent learning that relies on multiplicative updates.
Pos
100.0%
Neg
0.0%
1 comments with sentiment.
Cluster Engagement
Posts from X
Most Activity
Most Activity
VIEWS778
Yaroslav Bulatov@yaroslavvb
Method 3 is a good example of a non-"gradient descent"-based learning, a multiplicative update
Yaroslav Bulatov@yaroslavvb
First impression of Fable: a lot more concise out of the box. Example of asking it to prototype optimizer updates https://github.com/yaroslavvb/learning-rotations
18hViews 778Likes 2Bookmarks 1
BOOKMARKS3LIKES7RETWEETS1
Yaroslav Bulatov@yaroslavvb
The design sense of Fable is really nice, colors come out consistently on first try https://yaroslavvb.github.io/learning-rotations/
Yaroslav Bulatov@yaroslavvb
First impression of Fable: a lot more concise out of the box. Example of asking it to prototype optimizer updates https://github.com/yaroslavvb/learning-rotations
17hViews 654Likes 7Bookmarks 3