Andrej Karpathy says this AI gap is widening

AI researcher Andrej Karpathy said the gap in AI diffusion that he observed in April is widening.

Michael Macor/The San Francisco Chronicle via Getty Images OpenAI cofounder Andrej Karpathy says most of the world is still barely using large language models.

Karpathy said of those who do use LLMs, most regard AI as a Google-esque tool.

That's not the only gap that is growing amid AI diffusion, Karpathy wrote.

OpenAI cofounder Andrej Karpathy says the world still barely knows AI.

"The general public has mostly not interacted with these systems," Karpathy wrote on X on Sunday.

"When they have, it looks like a derpy chatbot." Karpathy said he was following up on his April observation about a growing gap among AI users.

Just over six months later, the renowned AI researcher who coined the term "vibe coding" sees the widening gap less as "two groups of people speaking past each other," as he phrased it in April, and more as a "sharp funnel." Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*.

It's now less "two groups of people speaking past each other" and more a sharp funnel. - Napkin math somewhere around 6B people (~75% of the population) have barely come… https://t.co/79xGhQdf1n — Andrej Karpathy (@karpathy) October 11, 2026 The situation, Karpathy said, is that based on his estimate, roughly 75% of the world's population "have barely come in contact" with large language models at all.

As for everyone else, Karpathy said around 1 to 2 billion people, "are casual and infrequent users of free-tier ChatGPT-like products." There's a bigger gap between the chatbot users and those who, Karpathy said, are using frontier models to code, a group that he estimates is around 20 million people or 0.2% of the population.

This group, he wrote, sees "large, complex projects that used to take them weeks/months can now be completed by agents with a prompt." The smallest group, which Karpathy estimates is just 5,000 people, consists of those with inside access to frontier AI systems, likely due to their employment at one of the leading companies.

Karpathy said the OpenAI Hugging Face breach showed what's on the horizon.

"It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics," he wrote.

"Things that would have taken top professionals in the industry years of work." The pace of AI advancement is so rapid, Karpathy said, that humans are struggling to review the outputs fast enough.