Feyn (@FeynAI) header
Feyn (@FeynAI) profile picture

Intelligence shaped by you

San Francisco
Joined December 2024
Models are improving fast at noticing things. That's mostly why they're getting better doing the jobs. January's models attempted only a quarter of planned work. The latest models attempt more than half.
1
1
14
This work was cheap too. The latest models cost $7 per hour of finished work. The average human labor hour costs $58
1
2
18
Can AI really do your job? We put 20 AI models in a work simulation, asking them to read their environment and do work without being asked. The latest models did 20% of planned work well enough to accept unchanged. In January, this number was 7%.
1
1
2
9
368
Critic is for agents too. Through the bundled MCP, your agent can pull a diff with the author’s reasoning attached, ask the authoring agent questions, and work through decisions directly. Let your agents talk to to each other
1
2
44
See which changes need your attention. Open a diff to read the change narrative written by the authoring agent. Have a question? Ask the agent that wrote it without ever leaving review
1
2
103
Today we're releasing Critic, a platform to review the code your agents write. Try it out at critic.run Since agents started writing our code, it has become harder to understand the scope of a change. Your AI has all the context behind the code but you don't. Critic lets the agent explain what changed and why, and answer questions from anyone reviewing its work.
3
1
2
6
1,324
Read more about MultiMatte xitter.gq/status/209779722711277… or build with it using NoBg. MultiMatte is open source. Check it out on HuggingFace: hf.co/feyninc/multimatte
3
40
MultiMatte is built on Meta's SAM 3. Our biggest win came from converting binary masks that classify each pixel as being inside or outside an object, to alpha mattes that assign an opacity value to every pixel. This allows MultiMatte to better understand hair, motion blur, translucency, and fuzzy boundaries. Across all our measured benchmarks, MultiMatte shows a step improvement over SAM 3. On DIS5K, S-measure rises from 0.674 to 0.908 (a 34.6% relative gain), and on DUT-OMRON from 0.792 to 0.901 (13.7%).
1
4
106
Introducing MultiMatte, a background removal model you can aim with words. Name an object in your image, MultiMatte keeps just that thing and cuts the rest. Same photo, three prompts: the dog, the dog bowl, the jeans. Try it out: usefeyn.com/multimatte
1
3
1
8
929
In training, a diverse data mix matters. Our first fine-tune used one dataset. It improved performance on camouflage images, but regressed elsewhere. We took this as a data problem and built a 26.1K-image mix from 10 datasets covering camouflage, crowded scenes, high-res subjects, portraits, and anime.
1
1
130
We took an interpretability-first approach for building FeyNoBg We studied the architecture of an existing model, BiRefNet. This led us to discover the importance of the third stage of its feature extractor. We expanded this stage from 18 to 24 blocks, giving it room to learn.
1
2
231
We’re releasing FeyNoBg, our open-source model for image background removal. Across eight benchmarks, it sets the best published S-measure on four and comes within 2% of the leader on the rest. Try it out: huggingface.co/spaces/feynin…
8
5
5
53
9,114
Inspection is entirely circumstantial. When the schema already contains enough information, SQRL returns the final query directly. Easy questions stay short. Ambiguous questions earn an inspection.
1
2
58
SQL queries can be syntactically sound, but still incorrect. A query can misread an ambiguous column or filter for a value that doesn’t exist, all without throwing an error. SQRL handles this by inspecting the database before answering. It asks for the stored value instead of guessing.
1
2
96
We’re introducing SQRL, a family of small text-to-SQL models. On the BIRD Dev benchmark, SQRL's 35B, 9B, and 4B models beat Claude Opus on SQL execution accuracy.
1
1
2
9
1,882
The gains are architectural. Today's best extractors are decoders limited by memory bandwidth. Pulpie is an encoder bound by compute. GPUs are starved for bandwidth, not compute. As a result, Pulpie is performant everywhere. Our testing shows Pulpie to be 7x faster on A100, and 20x faster on L4 GPUs.
1
4
326
Introducing Pulpie, models for cleaning the web. 70% of a typical HTML page is ads, navigation, and sidebars. Pulpie removes this noise and returns a clean document you can use in pre-training or as context. Previously, cleaning 1 billion pages with the best extractors cost $159,000. Pulpie brings that down to $7,900. A 20x decrease.
14
13
3
66
13,178
🔥🚀Chonkie 1.6.8 is out with PyEmscripten wheels! 🐍You can now run the full Python Chonkie library directly in your browser using Pyodide🦛 This is an early release that brings Chonkie’s core functionality to the web, but it's a big step towards 𝕓𝕣𝕠𝕨𝕤𝕖𝕣 𝕟𝕒𝕥𝕚𝕧𝕖 𝕡𝕪𝕥𝕙𝕠𝕟 𝕔𝕙𝕦𝕟𝕜𝕚𝕟𝕘. We’d love to hear your feedback! try it out and let us know what works, what doesn't, and what you'd like to see next! Happy coding! 🦛✨
2
7
422