- GPT-5.5, Claude 4.6, Gemini 3.0 all made AGI a reality simultaneously
- ChatGPT broke the Turing test years ago when it launched
- Physical AI is the next step – AGI could power real-world robots
In a recent interview on The Joe Rogan Experience, Marc Andreessen said that the latest cutting-edge models have finally enabled artificial general intelligence (AGI), which he describes as AI comparable to humans.
Andreessen claims that models like GPT-5.5, Claude 4.6 and Gemini 3.0 are now “as smart as a person”, concluding that AGI will no longer happen soon, but will already be here from 2026.
Speaking to Joe Rogan, the frontman and former software engineer explained that “99% of the time, the answer I get from AI… is better than what I would get from talking to almost any expert I have access to.” »
AGI limits continue to move, but likely reached in 2026
Instead of crediting a single model provider for reaching this milestone, he said several industry leaders reached the same point at the same time, a few months before the interview (in the first half of 2026).
The result is a system that behaves like a doctor, a lawyer, a world-class programmer and much more, combining the fluidity of reasoning and problem-solving with the ability to recall enormous amounts of information in an instant.
Stating that the definition of AGI has evolved in recent years, he cited the Turing Test, which has served as the benchmark for achieving AI for about six decades. With the launch of ChatGPT at the end of 2022, “we have passed the… test”.
But despite continued progress, Andreessen says society has failed to stop and understand the importance of this phenomenon, leading the public to significantly underestimate how quickly AI is advancing.
Sam Altman, CEO of OpenAI, has also been an influential speaker in the field of AGI, although he has not yet acknowledged that GPT models have officially achieved this status. The closest the OpenAI CEO comes to this milestone occasion is by calling GPT-5.5 “autistic general intelligence” in an X-rated article – a mockery that shares the same acronym as “artificial general intelligence.”
In a January 2025 blog post, he said: “We are now confident that we know how to construct AGI as we have traditionally understood it. »
In other words, while Andreessen thinks AGI is here, Altman now proposes that AGI, a fluid step, is no longer as important on the AI timeline.
The future of AGI and the “AI vampire”
In the interview, Andreessen also dismisses any fears that AI will replace human jobs, comparing the revolution to electricity, computers and, later, the Internet. Overall demand will continue to grow as productivity continues to rise, but workers could be affected in other ways.
He specifically mentioned the “AI vampire,” where many workers could end up working more to produce more, to the point that many become addicted to producing more work and sleeping less.
But it’s important to note that Andreessen sees AGI as fueling the real implementation of AI: robotics. Referring specifically to Tesla’s explicit and near-exclusive use of cameras for autonomous driving, he explains how Musk’s companies are on track to introduce physical AI through Tesla’s Autonomy, xAI’s Conversational Intelligence, and Optimus’ humanoid incarnation.
Just a few weeks ago, in an article
While Marc Andreessen says AGI is already here, industry figures like Sam Altman maintain a more cautious stance on this milestone. One thing is clear, however: attention is now shifting to the practical implications of computational intelligence as it transcends the physical world.
Follow TechRadar on Google News And add us as your favorite source to get our news, reviews and expert opinions in your feeds.
