on ai psychosis:
as i watch engineers fall deeper into the belief that an amalgamation of mathematical probabilities somehow understands their codebase better than they do, i find myself thinking back to a time when software wasn’t built for hypergrowth, but simply to do x without inventing y.
ai seems almost fundamentally opposed to this philosophy. ask it to do x and it will eagerly invent y and z before it has even tried to understand x.
the danger isn’t that ai writes bad code- it’s that it makes writing unnecessary code 100% free
and the human condition is such that some of us will always prefer the fast, steep gains of ai, even when it does a bajillion unrelated things to accomplish something that could have been done without changing anything else.
so, somewhat paradoxically, the quality of software may keep declining for as long as ai keeps getting better. the cheaper complexity becomes to create, the less incentive there is to understand or avoid it.
try to preserve this craft created by our ancestors
write a LOC by hand today
AI engineering after 13 years of backend:
- Your agent is a wrapper over a loop
- The loop is a wrapper over LLM calls
- Each LLM call is a wrapper over an HTTP request
- The HTTP request needs a timeout, a retry and an idempotency key
- The retries need a budget
- The budget needs a rate limiter
- The rate limiter needs observability
- The observability needs evals, because a 200 OK can still be a wrong answer
And all of it is a wrapper over one simple truth: An agent is a distributed system that can hallucinate.
building a coding agent? you can add herdr support yourself. no waiting on us.
new docs: https://t.co/wXOe9oMlEq
thanks to these folks for already shipping it 🫶
Figma doesn’t want you to use MCP
Amazon doesn’t want you to shop with agents
Reddit won’t let Claude access threads
X won’t let ChatGPT read tweets
It’s starting.
This happened with APIs 8 years ago.
Junior AI engineer: "Let's build an agent."
Mid-level: "Let's build a multi-agent system with LangGraph, a vector DB and 3 MCP servers."
Senior: "We need evals, guardrails and a model router before launch."
Principal: "Can this be one LLM call and a SQL query?"
The hardest skill in AI engineering is knowing when you don't need more AI.
Not every seed of service you plant will sprout and flourish, but if you keep planting, eventually some of them will. That's just the math of good deeds and good faith.