TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

What Do Tesla Autopilot and Grandmaster Chess Have in Common?

2 pointsby pongogogoabout 1 year ago

1 comment

advaelabout 1 year ago
Analogy is a useful and perhaps essential human reasoning tool, but it has some pretty key failure modes<p>For example, here we talk about a more broad general computational approach to chess using neural networks. This is an exciting development in many ways! But also, note how it&#x27;s building on already-superhuman performance of algorithms for playing chess. Part of the semi-supervised process described involves using an existing chess system (note that chess computers have for a long time been considered superhuman chess players) to assess scraped data and extrapolate synthetic data<p>We then want this analogy to apply to self-driving cars. So uh... let&#x27;s be real for a sec. I&#x27;m not a self-driving car expert, but I am a machine learning researcher and engineer and I&#x27;ve worked on self-driving car agents. I&#x27;ve kept up with the field on and off. I&#x27;ve heard people like Musk claim &quot;oh man we&#x27;re gonna be there in like a year, tops&quot; approximately every six months since 2012 or so.<p>This is like hundreds of binary orders of magnitude a harder problem than chess. The size of the search space, the incredible amount of precision required to be safe, and the overwhelmingly vast space of just &quot;edge cases&quot; that could in aggregate cause tragic fatalities and untold property damage is astronomical. Chess is a discrete state space and while not a solved game in the strict mathematical sense, certainly one we already know how to make good agents for. There is no working rule-based autopilot to simply bootstrap a neural autopilot from. I get that this is just making the same &quot;data wins&quot; talking point people have been saying for about a decade (both a useful insight in some domains and a convenient conclusion if you wanna sell GPUs and cloud servers), but I&#x27;m really failing to see how google throwing a bunch of compute at a more general solution to an ultra-feasible problem implies anything about a much harder and as of yet still unclear how to solve problem, realistically.<p>You can&#x27;t solve everything by throwing more dakka at it. Attention mechanisms are the big thing right now, and they like many other breakthroughs were not just &quot;throw more compute at it&quot;. Every big breakthrough in machine learning only does something cool when you throw a bunch of data at it. That&#x27;s how machine learning works. This might mean that the well-resourced org that can afford the compute and scrape the data necessary to get something useful may not even have made the relevant breakthrough, which I guess is why they so often conclude inane things like &quot;wow all it really took was zillions of datapoints. Scale really is everything!&quot; No, fool, scale is necessary but not sufficient to solve problems this way. Maybe the thing that&#x27;s too hard will get better with scale! But often it won&#x27;t. Often we need to actually invent new types of models that capture the structure of the problem better, fiddly little details that people don&#x27;t read about in soaring thinkpieces by businessmen. Sometimes we get the breakthroughs in the form of training protocols, dataset curation, or even building some boring old business logic around the &quot;AI&quot; in some novel way with a boring old tool that the press isn&#x27;t breathlessly anthropomorphizing and hand-wringing about, like a programming language. Until someone actually solves the problem in a way that actually works, it&#x27;s hard to say exactly what the solution will look like<p>I really don&#x27;t get why people are so eager to believe, at so many times, in so many domains, that we&#x27;ve learned everything we need to know, and all the problems will just disappear if you just pump enough energy into it. But I do know there&#x27;s a strong monetary incentive to both believe and make others believe it if you&#x27;re one of the people whose stock price hinges on that belief, whose fortunes come from duping investors and governments both into funding your R&amp;D by sheer bravado
评论 #39814290 未加载