Yeah this is a perfect use case for LLMs: Very tightly structured context, a well-defined domain of inputs, and a lot of iteration.

Show Post

Unfortunately I feel like these tools are ornerny to setup & build, aren't fun to run. & crucially, if you are not able to self improve your tools, bend them to your usage & will, you are on the losing side of this new age. This is too important to offload purely to platform team and or datadog

Show Post

I really hope we can get off of printf() debugging and start using good observability, tracing. There's a huge ecosystem of tools & systems to get into, but at least having good tools for you & AI both to see your code runs feels hella important.

Show Post

@chadfowler.com 's Phoenix Architecture embraces that dynamic energy, imagining regenerative software where we flip the bit on our disposition, from atemporal forever building to accepting iteration, change. Shear layers, software systems contracts: building in change. bsky.app/profile/chad...

Show Post

Architecting in our heads is both a beautiful thing and sometimes we really can place every piece. But a huge amount of AI strength is the volatility, is dynamic not static energy, of trying things, assessing, learning the shape of things, elaborating them, and gaining new perspectives fast.

Show Post