OpenClaw creator makes a case for 'specialized intelligence' over superintelligence

• OpenClaw creator Peter Steinberger isn't a fan of AGI or superintelligence.
• He says he believes in "specialized" AI instead.
• "What can one human being actually achieve?" he asked. The same goes for AI, he said.

In Silicon Valley, many investors, founders, and tech giants often frame AI as an all-powerful force — something approaching omniscience.

They call it artificial general intelligence, or AGI, a hypothetical system capable of matching or exceeding human reasoning across virtually all tasks and, beyond that, superintelligence, which would surpass human cognition altogether.

AGI is the thing OpenAI and its competitors are racing to be the first to bring to life.

The industry isn't a monolith, however, nor are its builders.

Peter Steinberger, the creator of OpenClaw, the now-viral personal AI assistant powering the agent-only social network, Moltbook, told the Y Combinator podcast on Saturday that he thinks the best AI is specialized, not generalized.

"What can one human being actually achieve? Do you think one human being could make an iPhone or one human being could go to space?" he said. "As a group we specialize, as a larger society we specialize even more."

He said the same holds true for AI. Current AI systems are labeled as "general." In practice, however, they are already specialized for particular tasks — like startups building models to solve Erdos math problems or identify gene mutations.

While the hype around AI has largely focused on the types of models being built — from large language models to computer vision systems or even world models — some startups and tech giants are also experimenting with more focused, subject-specific forms of intelligence.

Axiom, a startup founded by former Meta researcher Carina Hong and backed by $64 million in seed funding, for instance, is building specialized intelligence to tackle advanced mathematics.

Google DeepMind has developed AlphaGenome, a system designed to predict how variants or mutations in human DNA affect a wide range of gene-regulating biological processes.

Companies are also developing smaller models, which could make it easier to build AI systems tailored to specific subjects or domains. Aidan Gomez, the CEO of Cohere, which builds AI technology for enterprises, told Business Insider in 2024 that there was increasing pressure to build "smaller, more efficient models" and to make them smarter with the right data and algorithms, rather than just scaling them.

Steinberger is not the only one to push back against the idea of an all-powerful general intelligence.

Timnit Gebru, a computer scientist who founded and leads the Distributed AI Research Institute, said AGI is a "fictional thing" in a November video published by Nature.

She said the backbone of engineering lies in building well-scoped, testable systems. The pursuit of a "fictional, undefined 'machine god,'" she said, is pushing the industry toward deeper labor exploitation and environmental damage.

Steinberger, previously best known for launching the PDF processing company PSPDFKit, came out of retirement to build with AI. He said the technology has come a long way since he started, when his goal was simply to build "something that could type stuff so my computer could do things."