hn.today

Steve Yegge on AI Agents and the Future of Software Engineering

newsletter.pragmaticengineer.com5 points1 comments
Screenshot of Steve Yegge on AI Agents and the Future of Software Engineering

Steve Yegge, a veteran engineer and creator of the AI-agent orchestrator Gas Town, argues that coding agents have reached a turning point. Initially skeptical, he became convinced after using Claude Code and now sees AI capability improving on a steep curve, with model advances arriving faster and mistakes feeding into future training. He believes writing code by hand is becoming obsolete, even as demand for software grows. His eight-level adoption scale runs from avoiding AI, through using agents in an IDE, to delegating work to multiple agents and building an orchestrator. He warns that capable engineers who remain at the early stages risk falling behind.

Yegge predicts severe disruption at large companies: he estimates that employers may cut roughly half their engineering staff as they fund AI tools and rely on smaller teams. In his view, big organizations are poorly suited to the coming wave of innovation, which will come from small teams and non-programmers using AI to build software. He also says intensive agent use is mentally and physically draining, limiting sustainable AI-assisted work to about three hours a day. Traditional coding skills, debugging tools, and programming-language choices matter less than before, but building software has become more enjoyable. Yegge describes accepting the loss of hand-coding as a grieving process and suggests his non-coder wife could become a leading contributor to their game.

Read on newsletter.pragmaticengineer.com1 comments on Hacker News

Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.

More in AI

Mapping every neuron in Qwen3-8B

Mapping every neuron in Qwen3-8B

Researchers found individual neurons in Qwen3-8B that reliably steer its replies into distinct personas or styles, including pirate, German-speaking, and childlike voices. Nudging these neurons changed outputs across many questions without prompting or fine-tuning, revealing how specific behaviors can be localized and edited inside a language model. (fermionresearch.com)

ChatGPT for SNES

ChatGPT for SNES

ChatGPT for SNES runs a 2.2-million-parameter language model on the Super Nintendo’s Super FX 2 chip, with an optional Wikipedia fact book. The unofficial experiment can run in an emulator or on a flash-cart-equipped console, but its tiny model often gives incorrect answers. (bxfoundry.com)

The daily digest

Today's best Hacker News stories, summarized and screenshotted, one email a day.