AI agents are not amazing, they are slop: says OpenAI cofounder Andrej Karpathy as he strongly disagrees with CEO Sam Altman on AGI timeline

OpenAI cofounder and AI researcher Andrej Karpathy has strongly pushed back against the “agentic AI” hype, characterizing the current generation of autonomous AI systems as “slop.” Karpathy has directly challenged the notion that autonomous AI agents are on the cusp of delivering as promised, positioning himself in contrast to the optimistic timelines often suggested by figures like OpenAI CEO Sam Altman. “I feel like the industry is making too big of a jump and is trying to pretend like this is amazing, and it’s not. It’s slop," Karpathy said on The Dwarkesh Podcast.“They’re not coming to terms with it, and maybe they’re trying to fundraise or something like that. We’re at this intermediate stage,” he added.

When asked why ‘Agentic AI’ success will take a decade and why not one year or 50 years, Karpathy emphasised that current AI agents need ‘a lot of work’ as the ‘models are not there’.“Overall, the models are not there. I feel like the industry is making too big of a jump and is trying to pretend like this is amazing, and it's not. It's slop. They're not coming to terms with it, and maybe they're trying to fundraise or something like that. I'm not sure what's going on, but we're at this intermediate stage,” he explained.“The models are amazing. They still need a lot of work. For now, autocomplete is my sweet spot.

But sometimes, for some types of code, I will go to an LLM agent,” the OpenAI co-founder added.This timeline directly contrasts with the optimistic projections suggested by Sam Altman – another OpenAI cofounder – earlier this year. He said that AI agents will join the workforce in 2025.“We are now confident we know how to build AGI as we have traditionally understood it. We believe that, in 2025, we may see the first AI agents ‘join the workforce’ and materially change the output of companies,” Altman said in his blogpost.

Karpathy advocated for collaboration between humans and AI, rather than aiming for AI to replace human roles. He urged developers to focus on building systems that enhance user reasoning and learning, rather than models that merely generate automatic, potentially low-value results.