dude is on some generational run. highly recommend reading this anyone into harness design and sourcing evals. and viv is genius in making s…

The referenced tweet (from @himanshustwts) is not publicly accessible without a login, so its exact content cannot be retrieved. However, based on contextual search results, this post appears to be praising work by Viv (@Vtrivedy10) of LangChain on agent harness design and sourcing evals.

Here is the knowledge base summary:

Viv (@Vtrivedy10), an agents and evals researcher at LangChain, has produced a notable body of work on agent harness engineering — the practice of designing the systems around AI models to turn them into reliable work engines. Her writing covers key topics including how to source and curate evaluation datasets (via hand-curation, production trace mining, and external datasets), how to design against overfitting using optimization and holdout splits, and how to iteratively improve agent harnesses using evals as a hill-climbing signal. LangChain has open-sourced related tooling, including the Better-Harness system and the deepagents library, which serve as practical starting points for harness engineering and agent evaluation in production.

Related

• Frontier models will one shot just about anything few years Intelligence is compression Jevons law is more of a suggestion
• Probably the most important book to understand what's going on in the AI race
• Book very much worth reading
• If only we’d trained Grok on just these 2 books, we’d be done already!

Source: industry