The Bitter Lesson

Summary

Summarized wtih ChatGPT

The main lesson from 70 years of AI research is that methods relying on computation, like search and learning, are more effective than those based on human knowledge. Researchers often focus on incorporating their understanding, but this approach limits long-term progress. Ultimately, embracing general methods that scale with computation leads to breakthroughs in AI.

Key Takeaways:

1. Prioritize computational methods over human knowledge in AI research.
2. Embrace search and learning techniques for better outcomes.
3. Focus on developing methods that can adapt to complex problems rather than relying on simplified human concepts.

Highlights from Article

Most AI research has been conducted as if the computation available to the agent were constant (in which case leveraging human knowledge would be one of the only ways to improve performance) but, over a slightly longer time than a typical research project, massively more computation inevitably becomes available. Seeking an improvement that makes a difference in the shorter term, researchers seek to leverage their human knowledge of the domain, but the only thing that matters in the long run is the leveraging of computation.

• AI is smarter the build for the expectation that more and more computation will be available (but is this true when the AI companies start coming under the gun to be more revenue driven?)

A similar pattern of research progress was seen in computer Go, only delayed by a further 20 years. Enormous initial efforts went into avoiding search by taking advantage of human knowledge, or of the special features of the game, but all those efforts proved irrelevant, or worse, once search was applied effectively at scale. Also important was the use of learning by self play to learn a value function (as it was in many other games and even in chess, although learning did not play a big role in the 1997 program that first beat a world champion). Learning by self play, and learning in general, is like search in that it enables massive computation to be brought to bear. Search and learning are the two most important classes of techniques for utilizing massive amounts of computation in AI research. In computer Go, as in computer chess, researchers' initial effort was directed towards utilizing human understanding (so that less search was needed) and only much later was much greater success had by embracing search and learning.

• AI doesn’t need to think like us - we need to point it at a problem and let it try to figure out how to solve it on it’s own merit.

As in the games, researchers always tried to make systems that worked the way the researchers thought their own minds worked—they tried to put that knowledge in their systems—but it proved ultimately counterproductive, and a colossal waste of researcher’s time, when, through Moore’s law, massive computation became available and a means was found to put it to good use.

We have to learn the bitter lesson that building in how we think we think does not work in the long run. The bitter lesson is based on the historical observations that 1) AI researchers have often tried to build knowledge into their agents, 2) this always helps in the short term, and is personally satisfying to the researcher, but 3) in the long run it plateaus and even inhibits further progress, and 4) breakthrough progress eventually arrives by an opposing approach based on scaling computation by search and learning.

the actual contents of minds are tremendously, irredeemably complex; we should stop trying to find simple ways to think about the contents of minds, such as simple ways to think about space, objects, multiple agents, or symmetries.

• One saving grace is that we are complex beasts. We can think in a way that, for now, we can’t articulate, and we should take comfort in that while AI is learning however it likes to learn.

All material owns to the authors, of course. If I’m highlighting or writing notes on this, I mostly likely recommend reading the original article, of course.

See other recent things I’ve read here.