Isenberg says humans now err more than frontier agents
One line, no benchmark, and a whole lot of founders quietly staffing as if it were already true.
Greg Isenberg posted a single line with no chart attached. He said he does not know if it is AGI, but he thinks we have reached the point where the average human makes more mistakes than the average frontier model AI agent.
That is the whole post. No error rates, no task list, no side by side comparison, no definition of what counts as a mistake or which agents he has in mind. It is a feeling from someone who spends his days watching founders wire these things into real work, offered as a feeling.
What is actually being claimed
Read it carefully and it is narrower than the headline sounds. Isenberg is not saying agents are smarter than people. He is saying the error rate is now comparable and, in his view, tipped. Those are different arguments, and the second one is the one that changes how you hire. If you believe a human doing a repetitive judgment task will misfire more often than a model doing the same task, the case for a first hire in that seat gets thin fast.
The case against is that nobody has shown the numbers. The average human and the average frontier agent are both doing an enormous amount of load bearing work in that sentence. Humans and agents also fail differently. A person who does not know something usually knows they do not know it. A model producing a confident wrong answer at scale is a different shape of problem, and cheaper failures are not the same as fewer failures.
The read
Treat it as a vibe check on where the sentiment sits, not as evidence. It is worth a minute anyway, because it is the assumption a lot of founders are already building their staffing plans around, whether or not anyone has run the test. The useful version of this claim is the one you run yourself, on your own workflow, with a count of how many outputs came back wrong.
Until somebody publishes that count, this stays a hunch from an operator with a good view of the room.
I don't know if it's AGI but I think we've gotten to the point where the average human make more mistakes than the average frontier model AI agent