[Bjorkbat]

AI/ML is a bit of a wildcard for me.

On the one hand, I’m certain many companies will be eager cut costs through automation. On the other, investors will likely grow cold if the results just can’t live up to the hype fast enough.

[bart_spoon]

I’m in ML and looking to pivot to web dev due to this. My personal view is that ML is in the midst of its own dot-com bubble. There are plenty of valid use cases for ML currently, but on the whole it’s a solution in search of a problem, and a bit of a figurehead at most companies who are at the very least wanting to use ML/AI/data driven in marketing materials. If a serious recession hits, I think a lot of these companies will make the realization these initiatives aren’t worth much to them and cut their losses.

Eventually the field will stabilize in a much healthier place, but only after some dark days.

[rafiki6]

My guess is you haven't worked anywhere where ML is a part of the core product. think e.g. search and recommendations.

[bart_spoon]

I have. But at that job we got the ML system to a very solid state where it simply became a matter of maintenance. That didn’t require a whole lot of time, which meant people started looking for other ways to use ML, unsuccessfully, to justify keeping themselves around.

[whiplash451]

And the hype won’t deliver fast enough. What has not been automated today will not be automated (robustly) in two years time.

[mountainriver]

There is hype and then there is practical ML which is integrated and working in lots and lots of companies today

[mountainriver]

The results are living up. ML is integrated into the core of so many companies at this point it’s not like it was 10 years ago. ML is here to stay and it will be almost the entire economy in the not to distant future

[bart_spoon]

I think the results are mixed. I don’t think anyone is claiming ML will die off. But in my own experience, for every company with a healthy ML environment, there are dozens that are kind of just throwing money and data around without much success, or without even the type of problems ML can solve. And at many companies that do have a solid ML use case, you often have teams of data scientists looking to find new projects that justify their existence because the maintenance of existing ML products doesn’t take up capacity.

ML will always be around, but it’s also definitely in a bubble right now in my opinion and there will likely be a contraction before the profession matures and gets to a healthy place.

[mountainriver]

Thats fair, everything is in a bit of a free money bubble right now. And yeah I've seen team looking for ML solutions that didn't really have any.