The visual model runs up to three times faster.
Most of the web will never get APIs for AI agents, argues @Yutori_ai co-founder @DhruvBatra_ (ex-Meta FAIR).
Yutori's bet: agents use the web like people do. Pixels in, clicks out - their N1.5 browser agent beats Opus 4.7 and GPT-5.5 at 4-5x lower cost👀
https://youtu.be/m7yTwkzscWk
Users expressed gratitude for the insightful conversation about Yutori's N1.5 Browser Agent beating Opus and GPT models at lower cost.
No Digg Deeper questions have been answered for this story yet.
Thanks for having me on @chain_ofthought!
Yutori's new Navigator N1.5 model beats Opus 4.7 and GPT-5.5 on browser tasks while running 2-3x faster and 4-5x cheaper 👀📊
@DhruvBatra_ Co-founder and Chief Scientist of @yutori_ai (ex-head of embodied AI at Meta FAIR), came on @chain_ofthought to explain the bet: most of the web won't add APIs, so agents have to use it like people do. Pixels in, clicks out - he thinks we'll be sharing the digital roadways for the foreseeable future.
Chapters: (00:00) Pixels in, clicks out (01:37) Why most of the web will never get APIs (08:47) Aggregation, specialization, and human friction (11:39) Digital niches and specialized models (16:41) The web's heavy tail and where browser agents win (20:40) Inside Yutori's Navigator and Scouts (24:08) N1.5: writing JavaScript to cut trajectory length (27:45) Training on live websites (33:29) Open source: FAIR's legacy and the Chinese frontier (37:22) Agent frameworks: OpenClaw, Hermes, heartbeats (40:57) How non-technical users adopt agents (44:25) Smart glasses, robotics, and embodied AI (50:57) Compute demand and smaller on-device models (53:12) Why the company is called Yutori
Thesis: computer-use models clicking buttons IS how we'll agentify the long tail of the web.
Most of the web will never get APIs for AI agents, argues @Yutori_ai co-founder @DhruvBatra_ (ex-Meta FAIR).
Yutori's bet: agents use the web like people do. Pixels in, clicks out - their N1.5 browser agent beats Opus 4.7 and GPT-5.5 at 4-5x lower cost👀
https://youtu.be/m7yTwkzscWk
Yutori's new Navigator N1.5 model beats Opus 4.7 and GPT-5.5 on browser tasks while running 2-3x faster and 4-5x cheaper 👀📊
@DhruvBatra_ Co-founder and Chief Scientist of @yutori_ai (ex-head of embodied AI at Meta FAIR), came on @chain_ofthought to explain the bet: most of the web won't add APIs, so agents have to use it like people do. Pixels in, clicks out - he thinks we'll be sharing the digital roadways for the foreseeable future.
Chapters: (00:00) Pixels in, clicks out (01:37) Why most of the web will never get APIs (08:47) Aggregation, specialization, and human friction (11:39) Digital niches and specialized models (16:41) The web's heavy tail and where browser agents win (20:40) Inside Yutori's Navigator and Scouts (24:08) N1.5: writing JavaScript to cut trajectory length (27:45) Training on live websites (33:29) Open source: FAIR's legacy and the Chinese frontier (37:22) Agent frameworks: OpenClaw, Hermes, heartbeats (40:57) How non-technical users adopt agents (44:25) Smart glasses, robotics, and embodied AI (50:57) Compute demand and smaller on-device models (53:12) Why the company is called Yutori
@DhruvBatra_ 🤔💡
.. we should have had you write the episode title!
Full conversation + transcript: https://www.chainofthought.show/podcast/63-most-of-the-web-will-never-get-apis-for-ai-agents-dhruv-batra/
Watch our full @chain_ofthought with @DhruvBatra_ on the models that let agents act across the messy, human-built web + our wide ranging conversation on what's coming in AI:
YouTube: https://youtu.be/m7yTwkzscWk Episode page (listen, takeaways, transcript): https://www.chainofthought.show/podcast/63-most-of-the-web-will-never-get-apis-for-ai-agents-dhruv-batra/
@DhruvBatra_ @chain_ofthought Thanks for a fantastic conversation Dhruv!
For anyone interested, check out the full conversation, transcript, takeaways & more @chain_ofthought.show - https://www.chainofthought.show/podcast/63-most-of-the-web-will-never-get-apis-for-ai-agents-dhruv-batra/
@DhruvBatra_ It was our pleasure 🤝
@chain_ofthought We can ask our agents which ones they prefer 😀
