10/08/2026
The woman helping build AGI said in 2026 that AGI may already be here.
Nobody covered it as a major story.
Daniela Amodei is the president of Anthropic β one of three companies at the centre of the global AI race.
In January 2026 she told CNBC something that should have made every front page.
"By some definitions of Artificial General Intelligence, we've already surpassed that."
Not close.
Not coming soon.
Already surpassed.
She pointed to one specific example.
Claude β the AI her company builds β now writes code at the level of many of Anthropic's own software engineers.
"We probably employ some of the best engineers in the world," she said. "Many of them say Claude can do a lot of what they do."
Then she said something more striking.
"Maybe the concept of AGI itself is now wrong β or just outdated."
The person building it says the goal they are racing toward may no longer be the right goal.
Here is the broader picture.
In January 2025, OpenAI CEO Sam Altman wrote: "We are now confident we know how to build AGI as we have traditionally understood it."
Months later he called AGI "a very sloppy term."
OpenAI's reported internal definition of AGI is reportedly the moment their models generate $100 billion in annual profit.
Scientists including Yoshua Bengio and Max Tegmark proposed measuring AGI across ten cognitive domains in 2025.
No current AI passes all ten.
Geoffrey Hinton won the 2024 Nobel Prize for building the foundations of modern AI.
He now warns that AI agents are beginning to outsmart the tests designed to measure and contain them.
Anthropic CEO Dario Amodei said in January 2025 that AI could surpass almost all humans at almost everything within two to three years.
Community prediction markets put the 50% probability of AGI at the year 2041.
The expert survey median is 2047.
The definition has been moving since 1950.
It is still moving.
Which makes more sense to you β AGI as pure capability (outperform humans across enough tasks and you're there) or AGI as something deeper that requires actual understanding of what it's doing?