>>11969048>What's MLMachine learning
>What's up with Musk and OpenAiThey basically haven't made a single fucking discovery despite millions invested and dozens if not hundreds of talented researchers. Instead, they waste their time releasing deeper and deeper models and throwing a shit ton of computational power to make shit which is EXTREMELY marketable and cool-sounding but which has very limited practical use and doesn't represent a scientific move forward. I suspect that same shit happens in his other companies (the spacex car into space thing sounds suspiciously like a waste of money under the guide of "muh payload test") but, unlike AI, I am not expert in those fields so I usually hold my tongue.
OpenAI, on the other hand, is actually evil. I work in ML research, literally no one I know likes them, their projects, or their marketing tactics (which I will not get into).
Anyway, just take a look at
https://openai.com/Not ONE unique discovery in any of these projects, not ONE. Just a bunch of bored programmers applying old NN architectures and methods to cool-sounding but useless tasks like "Robot hand solves Rubik's cube", except they waste millions of dollars worth of computational power in the process.
