“Bots Manage Bots”: a16z Co-Founder Declares Arrival of AGI

AI has crossed into AGI, says a16z co-founder Marc Andreessen.

Artificial intelligence has already crossed the threshold into AGI. This was stated by Andreessen Horowitz (a16z) co-founder Marc Andreessen in a three-hour episode of The Joe Rogan Experience.

According to the investor, the turning point occurred “about three months ago” following the release of the latest generations of models from OpenAI, Anthropic, Google, and xAI — GPT-5.5, Claude 4.6, Gemini 3, and Grok 4.3.

The investor did not reference industry consensus or the scientific definition of AGI. In the podcast, it was presented as his personal assessment of the situation.

Andreessen claims that modern models “in 99% of cases” provide better answers than almost any available expert. He describes the new versions of LLM as a “qualitative leap” and a constant cognitive enhancer — a tool for writing texts, constructing arguments, analyzing complex topics, and finding solutions.

His method of working with AI involves a multi-step simplification of the answer: first “explain like I’m ten,” then “like I’m five,” and finally “like I’m two.”

Medicine

The a16z co-founder shared that during an illness, he used chatbots and called such systems “the best doctor in the history of the world” due to their round-the-clock availability and ability to guide a patient step-by-step.

He also stated that doctors are already widely using ChatGPT. “Everyone” asks the LLM, “what’s going on with this person” after seeing a patient. Some input data during the consultation.

Previously, the American Medical Association reported that 81% of doctors in the US use AI in their professional activities. The organization emphasized that such systems should remain an auxiliary tool and be used with transparency and safety requirements.

Andreessen also described multimodal medical scenarios: a user uploads blood test results, after which the neural network “tells you what’s wrong.” He cited an acquaintance who assembled a personal AI health panel “in the spirit of Star Trek.” The investor also mentioned the drop in the cost of full genome sequencing to about $200.

Programming and AI Agents

According to Andreessen, the current “cutting-edge level of development” in Silicon Valley involves working with about 20 AI bots simultaneously. They perform tasks in parallel 24/7, while a person checks the results “every 10 minutes” and provides feedback.

The next stage is agent hierarchies. Each bot will have its own “sub-bots,” and above them, bot managers. In a year, Andreessen predicts, a model with “10–20 virtual assistants, each with another 10–20 similar helpers,” will become routine.

Later, this will evolve into multi-level structures where “bots manage bots that manage other bots.” In such a configuration, one developer will be able to oversee up to 1,000 agents and become “1,000 times more productive.”

Science

Andreessen stated that AI is progressing particularly rapidly in areas with “provably correct answers” — mathematics, physics, chemistry, biology, and programming. According to him, systems are already solving problems that humans could not.

Next comes the creation of new drugs, cancer treatment, acceleration of space research, and the discovery of “new physics.”

Back in May, an AI model from OpenAI disproved an 80-year-old hypothesis by Paul Erdős on unit distances.