3 ms·
Robots learn to evolve and improve
- fragsworth 11y agoI think they just took a concept figured out in 1994 (https://www.youtube.com/watch?v=JBgG_VSP7f8 https://www.youtube.com/watch?v=JBgG_VSP7f8) and applied it to robots, in an attempt to get an easy publication written up. Really, though, we know we can make robot parts. The hard part is the software. What these people did by making it robotic was completely pointless (because making the robots is a huge amount of unnecessary added work) that doesn't add to our knowledge until we can solve all the software problems in simulations first.
- mangeletti 11y agoThere's a thing that's concerning to me about AI and robotics articles in mainstream media; nearly every such article has some sort of "but there's no need to worry just yet" qualifying statement before the end. It's almost as if there is this obvious future we're all subconsciously aware of, that we fear greatly, but that we're somehow afraid to confront in meaningful way.
- p1esk 11y agoYes, that's pretty much the case. You can think about it like this: the alien intelligence will arrive on this planet in a few decades, and we have no clue how to "confront in meaningful way". The question is not "if", it's "when".
- Houshalter 11y agoThe AI Risk has recently gotten quite a bit of media attention from the book _Superintelligence_. As well as some comments from notable people who read the book, like Elon Musk, Stephen Hawking, Bill Gates, and Sam Altman. There is also a lot of recent progress in AI that's bringing attention to the issue. Computer vision most notably. It seemed a basically impossible problem with very slow progress. Then within a few years we went past human accuracy on some tasks. Very open ended tasks like identifying 1000 different objects from real world images scraped from the internet.