3 ms·
One would think one potential solution is to start working on problems we don't actually understand. General AI, algebraic frameworks, high efficiency cross-cut
by Retra 7y ago
One would think one potential solution is to start working on problems we don't actually understand. General AI, algebraic frameworks, high efficiency cross-cutting models, etc. Pie in the sky stuff.
- tachyonbeam 7y agoI work in AI and I find it pretty discouraging at times. The biggest conference in the field had about 9000 attendees last time. There are thousands of publications coming out every year. It's basically impossible to keep up, and your chances of getting scooped (someone publishing your idea before you) are pretty high. Many of the more obvious research directions have already been tried. Many ideas don't work so well. It's tough. I personally have moments where I struggle to believe I can have any impact in this field.