Albeit azure is better than GCP considering GCP is heavily leaned towards TF. They also both have Quantum AI teams, which is the cutting edge for AI. This really boils down to the amount of resources, and ease at which they are available to AI scientists which puts them at “better than” smaller tech corporations including Apple. Ladder climbing is also the hardest in these two, as it’s easier for me to go to a small company and be the rockstar AI scientist, whereas at a tech giant, there are many rockstar AI scientists. Source: Me, an AI researcher.
Microsoft AI research is ages behind Facebook and google. Microsoft just fails to attract top talent. There is absolutely no point of comparison. Apple would be good too if they weren’t doing everything behind closed doors. Last good researcher microsoft had was Kaiming He and he got snashed by FB years ago ( i.e. Microsoft sucks at hiring top talent)
I really don’t expect speculative SWE’s to understand the research world of AI. You’re wrong, and I don’t argue with non-AI experts on topics I’m a paid expert in.
Or maybe just show me field in AI where Microsoft is ahead of google and fb. Comparing google brain and fair to MSR is an absolute joke tbh. PS: ease on the kool aid , research papers don’t lie ;)
Oh no this guy again. The new grad with 184 TC
What should I study to get into AI? Also TC or...
My TC is 184 but I’m a new grad. I studied a wide range of courses in school, can you be more specific?
What's wrong with being heavily leaned into Tensorflow? I rate it.
There’s nothing wrong with it. However, realistically when you work as a machine Learning researcher, you use many libraries and not just tensorflow.
Gotcha. Right tool for the job.
Ive always been a big fan of MSR, they seem to do research in way more areas than AI which I find more interesting as well. Quantum AI is an interesting area. UMD had a little conference on quantum AI last September. Seth.lloyd and Scott A were there and a bunch of others. Aren't there some barrier to entries like quantum memory etc.?
Quantum AI is cool in theory, but much like many areas of quantum research, it’s usually highly theoretical with little application except maybe a proof. However MSR & Alphabet both have d-wave machines so they have the resources to actually take it further.
D wave is only an annealer though. A lot of debate as to whether it's even exploiting any quantum phenomena as well.
This isn’t a discussion, but a statement.
People are discussing how it's a wrong statement
Not just wrong but clearly written by a teenager know-it-all