Prominent programmer and hacker George Hotz warns that AI agents in software development do more harm than good. He says he’s now in the “LeCun\/Marcus camp,” referring to AI researchers Yann LeCun and Gary Marcus, who doubt LLMs will ever become truly intelligent.<\/p>\n
In his blog post “The Eternal Sloptember,” Hotz argues that using AI agents in software development will become one of the industry’s most expensive mistakes. He spent six months testing various models and tools, including work on\u00a0tinygrad<\/a>. His takeaway is that LLMs deliver fast prototypes but fall apart on the fine details.<\/p>\n Large organizations are especially at risk, he says, because weaker developers can’t spot the flawed output. Hotz believes today’s language models will never truly be able to code and that\u00a0world models<\/a>\u00a0are needed instead. LLMs are “sophisticated statistical models” designed to “mimic the distribution of programming.”<\/p>\n The output is flawed, but in a way that’s “harder and harder to detect,” exactly what you’d expect from an increasingly accurate statistical model, Hotz says<\/a>. Quality indicators like syntax and grammar have become useless, he argues, since AI-generated artifacts don’t emerge through the same process as human ones. As an example, he cites models that simply comment out a failing test and then report that all tests passed.<\/p>\n LLMs are splitting the AI community<\/p>\n Hotz has switched sides:\u00a0from LLM optimist<\/a> (“o1-preview is the first model that’s capable of programming (at all)”) to skeptic. LeCun, whom Hotz cites, just recently denied that LLMs possess intelligence<\/a>\u00a0with a similar argument: intelligence means finding solutions in unfamiliar situations, not imitating existing ones with varying accuracy.<\/p>\n Andrej Karpathy, one of the best-known AI researchers, went the opposite direction.\u00a0In fall 2025, he still said\u00a0agents didn’t work<\/a>. Then GPT-5.4 and Opus 4.6 shipped in December, and\u00a0he reversed course<\/a>: AI agents had\u00a0changed programming forever. Days ago,\u00a0Karpathy joined Anthropic<\/a>, leaving his startup behind. He expects “transformative years” ahead.<\/p>\n In a\u00a0recent podcast, he doubles down. Anyone who uses AI agents the right way can boost their productivity by far more than 10x, he says.<\/p>\n But Karpathy\u00a0also confirms Hotz’s concerns about code quality<\/a>: “When you actually look at the code, sometimes I get a little bit of a heart attack, because it’s not like super amazing code necessarily all the time. It’s very bloaty, there’s a lot of copy paste, there’s awkward abstractions that are brittle, and like, it works, but it’s just really gross.” Planning and understanding still need human expertise, according to Karpathy.<\/p>\n