Commentator tszzl argues recursive self-improvement is a continuous historical process active since the Renaissance, not a future AI milestone

@tszzl .

@tszzl old post on HN i sometimes think about

@tszzl *watching the mushroom cloud consume downtown Hiroshima* "ya know, if you think about it this is really just more sunshine"

@tszzl

@tszzl so recently?

@tszzl consciousness is the original recursively self improving process

incredible, thanks for shaing it

this climb almost feels like thermodynamic destiny that has been locked in since we evolved social brains that could do compound learning of (agro)culture.

this essay from 1987 similarly meditates on agriculture's downstream effects -- a mind-bending read: https://web.cs.ucdavis.edu/~rogaway/classes/188/materials/Diamond-TheWorstMistakeInTheHistoryOfTheHumanRace.pdf

@tszzl bruh i read that as princess

@tszzl I've been saying we will look back on the industrial revolution as the first 0.1% of the artificial intelligence revolution but I think you're right it actually started earlier.

@tszzl give me a crane large enough and l it’ll come up with plans to a better crane

A huge mistake of Eliezer et al was to think of RSI and the singularity as being led by a rogue singleton rather than an ensemble of (thousands? millions? billions? trillions?) of intelligences - the global economy is the real ASI, has been around for a long time, and is also partially but not wholly aligned with values that benefit humans / conscious beings

@tszzl I'm so stoned I read this as RSI is a princess that's been trapped in a other castle

@tszzl What is artificial intelligence if not the corporation

@tszzl id say the agricultural revolution was the intelligible beginning

@tszzl @8bit5_0 RSI is the friends we made along the way.

@tszzl renaissance rationalitization and oceanic navigation lock into blah blah blah

@tszzl https://en.wikipedia.org/wiki/San_Francisco_Renaissance

@tszzl lots of 'long rsi' responses in the comments but the real answer is probably the tipping point in the pleistocene when social learning/imitation became reliable enough for culture to start compounding

@xlr8harder @tszzl

@tszzl sadly the ancient method of discovery was recovered by newton reading pappus’ collection of the treasury of analysis. should’ve stayed dead