Ajeya Cotra

So there was relatively constant growth in effective flops, amount of computation available adjusted by ability to use that computation efficiently.

There was no obvious way to know how many flops AGI would take, but there were some intuitively compelling guesses.

For example, an AGI that was as smart as humans might need a similar level of computing capacity as the human brain.

Kortra picked five intuitively compelling guesses, the namesake bio-anchors, and turned them into a weighted average.

Then she calculated, given the rate at which available flops were increasing and the number of flops needed for AGI, how long until we closed the distance and got AGI?

At the time, I found this deeply unintuitive, but it's held up.

Improvement in AI since 2020 really has come from compute, the construction of giant data centers.

Improvement in the underlying technology really has been measurable in effective flops, that is, the multiple it provides to compute, rather than some totally different, incommensurable paradigm.

And Kotra's anchors, the intuitively compelling guesses about where AGI might be, match nicely with how far AI has improved since 2020, and how far it subjectively feels like it still has to go.

All of the weird hard parts went as well as possible.

So, again, what went wrong?

3.

In 2023, Tom Davidson published an updated version of BioAnchors that added a term representing the possibility of recursive self-improvement.

the new calculations shifted the median date of AGI from 2053 to 2043.

This doesn't explain why our own timeline seems to be going faster than BioAnchor's.

Even 2043 now feels like on the late side.

And anyway, recursive self-improvement has barely begun to have effects.

But in 2025, John Crocs published a thorough report card on Davidson's model.

He took his numbers from Epic, who used real data from the 2020-2025 period that earlier forecasters didn't have access to, as well as the latest projections for what AR companies plan to do over the next few years, to come up with more formal projections.

Most of his critiques apply to BioAnchors too.