AlphaGo Zero: ~3e23 flops
AlphaGo Zero was trained on far more compute than any other AI model of its era — roughly 3e23 flops, comparable in order of magnitude to frontier LLMs.
Similar snapshots
Data point
16
The European Commission is weighing an EU-wide ban on social media access for children under 16, following a panel report due Monday.
Data point
70%
Daniel Kokotajlo puts a 70% probability on AI development leading to catastrophic outcomes such as AI takeover or something approaching human extinction.
Data point
2029
Daniel Kokotajlo's 50% probability estimate for the arrival of superintelligence is currently 2029, with a strong possibility it happens by end of the decade.
Regulation must come before 2030
If governments wait until mass unemployment sets in before regulating AI, it will already be too late — the companies' strategy is to achieve superintelligence first, then automate jobs.
Data point
100x
From 175 billion parameters in 2020 to approximately 10 trillion today, AI models have grown by roughly two orders of magnitude in six years.
Data point
90th %ile
AI 2027 vastly exceeded Kokotajlo's own pre-publication readership forecasts, reaching what he estimates as a 90th percentile outcome in views.
Data point
20%
In the AI 2040 Plan A scenario, one fifth of all cognitive labor is projected to be performed by AI by 2031.
Data point
100x
In the AI 2040 scenario, by 2033 there are 60 million AIs operating at 100 times human cognitive speed, dramatically transforming the economy.
OpenAI coached resignations
Apple's lawsuit alleges OpenAI advised departing employees not to disclose their next employer and coached them on avoiding immediate 'walkout' termination to preserve access to Apple's confidential data.
