Google engineer says self-improving AI is still far off

On October 15, The Decoder reported that Google engineer Haydn Belfield said Google DeepMind has not yet measured any acceleration of AI research driven by AI itself, and that self-improving AI remains far off. The claim concerns AI's effect on the pace of AI research and where self-improving AI stands today. The report gives only Belfield's judgment. It does not disclose the measurement method, the data or the time frame, and it says nothing about follow-up work.

Written by AI from the coverage · updated 2 hours ago

Not enough continuous observations to draw a trend yet.