“Even in one or two years, we'll find that the models can do a lot more involved tasks than they can do now. For example, you could imagine having the models carry out a whole coding project instead of it giving you one suggestion on how to write a function. You could imagine the model taking high-level instructions on what to code and going out on its own, writing any files, and testing it, and looking at the output. It might even iterate on that a bit.”

Well within the two-year window, agentic coding tools (Claude Code, OpenAI Codex, Cursor and Devin-style agents) routinely take high-level instructions, create and edit files across a whole project, run tests, read the output and iterate — exactly the capability Schulman described, though reliability on large projects remains uneven.