|
|
|
|
|
|
by hacknat
1399 days ago
|
|
|
I'm not saying that there aren't dystopian problems in our future, but my bias is that we're actually probably not doing enough, not that we're doing too much. Measuring technology progress is actually very difficult to do, most of the actors have massive incentives to lie and exaggerate. > a VR metaverse dominated by super sophisticated ad tech No one is joining Meta (in fact they're losing people now), it's going to fail. Actually VR in general has been unable to gain market traction for years. More and more signs point to people exiting social media (TikTok is still a concern though). > all controlled by super intelligent AGI Don't believe the AI hype. We've actually made very little progress on theoretical AI, the breakthroughs have all been in engineering (mostly just adding stacks to NNs that hyperscalar cloud architectures have made easy). We don't actually know what General Intelligence is yet or how to implement it. The engineering progress is making people think there is crazy progress in AI when there isn't. > think DALL E and where that's headed. DALL E is just cobbling together two AI innovations, semantic association and image creation. AI has been able to "combine" two types of image styles together for a while. Now there's a semantic addon that does the heavy lifting of looking up the styles for you based on the words you said/wrote. |
|
|
The graduating president of the Caltech student body just joined Meta. They may be losing people, but they are hiring, and although their work may not be benign, it does qualify as interesting, challenging, and a potential "growth" opportunity.