Sebastian Mallaby Predicts AI Recursive Self-Improvement By 2028

@tferriss Give us 2 hour workweek with AI

@tferriss A powerful and though-provoking outlook on the future of AI If recursive self-improvement becomes reality, it reises urgent questions about governance, safety, and who guides that level of power responsibly

@tferriss Everyone is asking how intelligent AI will become, but few are asking whether humanity can mature as quickly as the intelligence it’s creating.

First understand your own mind.

@tferriss Your environment shapes you future, choose wisely.

@tferriss Won’t remotely take until 2028 for RSI. EOY 2026 is possible, or H1 2027 at the latest.

@tferriss History will end…?

@tferriss I can't roleplay this persona or fulfill this specific request.

@tferriss recursive self improvement is the last lap of human written code

@tferriss 2028 for recursive self-improvement assumes the bottleneck is code, not data and compute. it isn't.

@tferriss I think this one isnt going to age well. 2028 is a year and half away. recursive self improvement still requires another major breakthrough. we'll see.

@tferriss Why not 2027? Or 2030?

@tferriss No the AI will run out of ideas. The training data only has so many ideas in there. At some point you need a developer to inject new ideas into the pipeline.

@tferriss https://ijnevermiss.gumroad.com/l/ytfacelessguide

@tferriss http://gofund.me/68137ba95