7 ms·
What if A.I. gets 100x better in a matter of months?
- ipsum2 5y agoArticle was written in mid 2020, and we haven't seen a (public) 100x gpt-3 AI yet, so this prediction can safely be marked incorrect.
- fragsworth 5y agoIt was posed as a question, not a prediction, and is not any less worth considering today
- ipsum2 5y ago"Will the Olympics take place in 2021?" is both a question and a prediction.
- gremlinsinc 5y ago> There are numbers in the post, but the argument goes that a 100x more effective A.I. will cost in the range of only $1bn, which is a relatively small fraction of Big Tech R&D. If this were the case then I'm afraid china probably is already there which is almost as frightening as a galactic alien civ that's colonizing planets. Because there's no way other civs could keep up they'd always be a step ahead.. Or 500... Considering at some point so could start doubling knowledge monthly then weekly the daily then every second....
- gmuslera 5y agoBetter at what? Computers outdated a lot of employments already, but opened a lot of new fields and job opportunities. Making advancements is not the problem. Still worrying about the weapon instead of the hand improving and holding it?
- 2342dfasdf 5y ago| but in the meantime? Chaos
- mikewarot 5y agoThis could happen, if someone came up with a better learning algorithm that took 1/10th of the time for the same accuracy, then others improved it while computing hardware kept flowing inward because it was now worth doing things that were previously seen as infeasible. The hardware isn't likely to get faster at more than Moore's law, but the way we use it, and connect it, might.
- stmfreak 5y agoWhy do people assume that an Artificial Intelligence will want to do all the crappy human jobs?