[edit: they're not that bad in the sense that hiring is a inherently lossy process of projecting something incredibly complicated, like skills, personality, motivation, and situation into a 45 minute interview where only 1 or 2 dimension can be measured. If you increase the time/cost and do hire fast fire fast, then fine, you can get a better interview process, but it's not free. Other industries use stamps and certs to do that sorting, also not cost free. Coding interviews, yes we all hate it, but it's all a tradeoff.]

https://archive.ph/gGqvX

In 2010, if you were in the right places and looking at the right trends (Theano, ImageNet, CUDA, the explosion of available datasets), you could reasonably predict that machine learning would be a good investment. (and ironically probably one of the worst thing to get into right now due to potential bull-whip effect)

Now I fully understand that hindsight is 20/20, and that any new and untested technology will be highly uncertain. But if you're working on something exciting, could you make a pitch as to why your specific field would take off in the next 10 years and is a good time to invest in right now? Rust? WASM? Maybe even robotics?

(For my own prediction: I think something Application-specific integrated circuits are super interesting to look into for the next 10 years. The slowing of Moores Law and Dennard scaling, and the ever increasing focus on specialized hardware all points to that ASICS would become more interesting. Personally I know very little about this field, but it feels like the age of homogenous computing is over and the next 10 years could be an exciting time for be able to go `full-stack`)

Comments URL: https://news.ycombinator.com/item?id=25502388

Points: 43

# Comments: 39