4 Emerging Practices for Building with Taste in the Age of AI

As AI makes building easier, taste becomes the differentiator. But how do you actually express that taste? What are the decision points, skills, and practices that let you channel judgment into the things you build with agents? Brandon Kase and Christine Yip discuss emerging practices that help develop and channel taste into our work. 00:00 Intro 01:54 #1: Leverage shifting left inside agent harnesses 13:11 #2: Compress intent with expert language 26:06 #3: Match the tool interface to the task 40:48 #4: Use oracles (when you can) X: https://x.com/bkase_ https://x.com/christinetyip @_lopopolo’s talk at AI Native DevCon and OpenAI’s post on harness engineering    • Ryan Lopopolo: OpenAI's Framework for Ship...   and https://openai.com/nl-NL/index/harnes... Brigitta Bockeler’s keynote on harness engineering, guides, and sensors https://martinfowler.com/articles/har... @emilkowalski’s motion vocabulary for AI-generated animation https://x.com/emilkowalski/status/206... @rauchg on human language as an API to the world https://x.com/rauchg/status/206186213... @trq212 on Anthropic’s dynamic workflows https://x.com/trq212/status/206190733... Cloudflare’s code-mode-style search interface https://blog.cloudflare.com/code-mode... EffectTS https://effect.website/ @bkase’s talk and blog on using algebra for software engineering    • Beyond Types in Swift – Brandon Kase   and https://www.bkase.io/posts/semigroups... @bkase building a Game Boy debugger using code mode https://github.com/bkase/gbllm/blob/m... @justincormack on using oracles while building an S3-like system    • Justin Cormack - When Tests Lie: Using Obs...   Music licensed through Soundstripe. Code: TRTIUYVAKR6LCSPB