An Interview with Gregory Allen About Anthropic and the U.S. Government

[AINews] Is Harness Engineering real?

interesting discussion on how much to build context engineering vs. the model being everything

• Central debate emerging: 'Big Model' (minimal harness, model does everything) vs 'Big Harness' (orchestration/framework layer adds value) - mirrors finance debate about trader skill vs institutional position
• Model providers like Anthropic/OpenAI are philosophically minimalist on harness - Claude Code rewrites from scratch every 3-4 weeks, emphasizing 'thinnest possible wrapper' with all secret sauce in the model itself
• Existential threat to AI framework/orchestration companies as reasoning models improve - framework founders questioning their own necessity as models become more capable of self-orchestration

Been using the Claude Excel plugin for a week and I genuinely didn’t expect it to hit this hard

this is an example of where tools are augmenting the hell out of humans, and it s hard to imagine the next step where this becomes entirely agentic including sheet construction. what then?

One Year With Lovable: The Workflow Breakthrough That Changed Everything

This is the operator-builds-with-AI story I keep seeing. Non-technical person automates their actual workflow, not a toy project. The specificity of what they built matters more than the tools they used.