[compactmani]

I stopped here: The residency program is similar to a top Master’s or PhD program in deep learning.

If you accept that a Master's program is contained within a PhD program, and this residency is similar to a PhD program in deep learning from a top school, then this would be the fastest PhD-like (revised from equivalent which is too strong of a claim) program ever. Only 12 months and a google badge!

[sleet]

The very next sentence is:

> The residents are expected to read papers, work on research projects, and publish their work in top-tier venues.

My reading of that they're saying that the program has similar activities not that it provides a comparable outcome.

[niels_olson]

I submit the connections at Google are unbelievable. It seems reasonable that you could in fact learn substantially faster in that environment. I don't know if you can cram 6-7 years into one, but maybe in 2-3.

[jofer]

A PhD isn't so much about learning the material as it is about learning to manage and complete projects. Your job is to identify unsolved problems, come up with ideas to solve them, and carry it through to a few completed research projects. Depending on your field, a lot of this can become more managing logistics and people than "pure" research.

In my experience (which, admittedly, has only involved much less flexible corporations than Google), you don't get to the level of responsibility you have as a PhD student or postdoc until you've been with a large corporation for >20 years.

I admit I have a slightly skewed view, but I've very skeptical that this is anywhere close to a complete PhD program.

It looks a lot more like a 1-year internship on an R&D team. Still a very good and useful thing, but you're not the one responsible for completion of projects.

[grayclhn]

The connections at a top PhD program are unbelievable too.

[ska]

I think what they really meant was that this program would be similar to (one year of) a top masters or Ph.D program, which would be plausible. Your inference of 2x doesn't meet the giggle test.

It's going to vary person to person of course, but I'd expect the learning rate to be at best about the same as a top program.

[choppaface]

Perhaps the idea is they want an APM-like program but for deep learning researchers. The APM program has been pretty successful, and a lot of deep learning progress boils down to tuning architectures and dealing with overfitting and other domain problems. This sounds like a good opportunity for a practitioner, and all the recruiting spin is in place to attract adventurous candidates who might be open to deviating from the traditional grad program grind.

[kylebgorman]

There's no such thing as a PhD in deep learning. Your best bet is going to be doing machine learning at a top-tier CS school that has someone who is interested in it. But you're still going to have to take architecture, algorithms, OSes, TA, etc.

[sjg007]

Which makes it a great deal!

[compactmani]

I know right.

Google: Now offering PhD-like programs in only 12 short months!

[UmDieWelt]

It's no match for my 12 week PhD bootcamp.

[nandreev]

Sounds tiring! fourhoursemester.com anyone?