This is such a brilliant and succinct explanation of unsupervised learning

ilya shares the same thing in an interview with jensen Everyones betting on the Strong belief that compression will lead to unsupervised learning, no complex algo needed..just near ~lossless compression!

Can Reinforcement Learning beat the market? In this free Substack post I explore a trading strategy based on Q-learning that delivers market-beating returns on the Visa stock with much smaller drawdown. Full code and backtest is available for free. Check it out!

Sounds like quantum mechanics

Hutter was on this. See also

@pawtrammell weigh in

it's middle out after all

fully bayesian 🤠

Wow I finally learned what researchers are doing.

hallmark of great and deep understanding is the capacity to explain smth in a surprisingly simple way

So, its analogous to the principle of least time for light to travel, also known as Fermat's Principle, except for information as opposed to light. BASED.