A lot will just move to finance...

However, financial firms are usually pretty against open academic research that would allow publication of papers. Many researchers truly value the ability to contribute to this field openly. Perhaps you are right and they will move to health.

Dodd-Frank isn't repealed and AFAIU, it made taking investor money much harder than it used to be.

We can see the next most lucrative fields where ML can be applied to by Alphabet's portfolio: * Calico, Verily - healthcare * Waymo - self driving cars

Another fields which comes to mind is robotics but Google liquidated their holdings related to this (Boston Dynamics) for some reasons.

Hopefully some find their way into rocketry research to help further commoditise space travel.

I'm trying to. It's not simple, if you're not in the US or Germany :<.

Germany, eh? Interesting. What kind of companies in that field are in Germany?

I wish I knew more about this subject.

Too many to list. DLR - Germany's NASA - is a key player in Europe's space industry, so it draws all kinds of space companies to its neighbourhood.

EDIT: those two links might give you a quick overview:

- https://spaceindividuals.com/space-companies

- https://spaceindividuals.com/space-jobs?country=Germany (over 50% of all listed, but not surprising, since the site's owner is in Germany as well)

Right now, I'm trying to map the industry myself.

There are not that many actual hard data science experts in adtech. Much of it is smoke and mirrors with every vendor claiming "AI" even if it's just some excel tweaking. The industry can't even agree on standard metrics or run a campaign without 20% discrepancies.

Also you can definitely apply ML techniques to data without PII, it's just harder, which actually makes the job more in-demand for high-quality talent.

As someone who has just left the advertising industry, and was doing data stuff, the real issue is that a lot of companies have a lot of data but almost none of it has any actual useful information content in it. It’s a bit of a hard concept to explain to explain to people because they go “but there’s a hundred something columns, and hundreds of millions of rows!”, you can have all the data you want, but if it’s all non-indicative trash you’re not going to be able to usefully predict anything, but hey, you’ve hired a data scientist now and added the cost of a big spark cluster to your AWS bill now, so you may as well continue telling the market you’ve got some real hot stuff.

Yes, agreed. Lots of useless data out there that doesn't really have significant signals. Also a lot of it is duplicated with everyone getting the same logs, so I find it hard to believe that any single company magically derived a big advantage over another.

Not to mention, no amount of AI powered results can beat a slick sales team. $1M spent on quality salesperson is much higher ROI than $1M on fancy AI engineering when it comes to relationship-powered media spend.

In statistics, there's a classic quote by John Tukey "The combination of some data and an aching desire for an answer does not ensure that a reasonable answer can be extracted from a given body of data."

This should be printed in large letters on the wall of anybody claiming to do data science.

They can still use sufficiently anonymized data under GDPR. They also will have data from people who have consented to have their data used. They will also have data from visitors who are not in the Union.

I would not at all be surprised if this is enough for the machine learning and data science people to be effective, especially at very high traffic sites like Facebook and Twitter.

I don't believe for a second that they can make the data "sufficiently anonymized". That is extremely hard and pretty much every time "sufficiently anonymized" data gets leaked it turns out it weasn't sufficiently anonymized after all.

Anonymization won't really help because you can't store unique identifiers. IPs, emails device identifiers, etc and anything derived from them (no hashes of that data either) are all prohibited. So you can't really build a profile of someone if you've got no key to look it up when you need to target someone.

But that still prevents the entire targeted ad industry. If you can't hold any identifying information on a user that didn't opt-in to tracking, you can't show targeted ads to that user.

There's a lot more stuff to do with ML and make money than ad targeting

Adtech is the highest paid application of ML.

Based on what exactly? ML is not that common in adtech companies, many just lie about some basic regression testing. Some even just put up a facade in front of grunts with excel.

Even the most sophisticated ad modeling pales in comparison to fintech and bioscience/pharma where the real money is.

Maybe that’s true, but if the ML ad targeting industry disappears, several of those currently employed ML workers have little to no reason to exist.

Yikes! Why the hyperbole? I'm sure if someone is clever enough to learn and be productive with ML, they'll be able to pivot and be productive in another field and/or using other tools.

He's talking about the positions, not the people in them.

What did you have in mind?

To me, the best applications are going to be those by which the human and the ML'd AI will benefit each other by a greater degree than if each were on their own.

At my particular company in logistics, I can see many places where it can be applied- especially in the dispatching of deliveries and helping to set up a call with customer support.

Predictive maintenance I think will be big as well

Two "might" occurrences means it's unlikely. The chances are these people will be swallowed by whatever the most profitable industry is at the time.

My guess would be similar learning systems that aren't subject to GDPR, such as government, external to EU (US), the ludicrous information of things (smart sensors, cloud, etc) where information doesn't leave the user's control, AI driven products, etc.

Then you'll have quite a few who will just continue regardless, sometimes it's easier to apologize (and pay a small fine) than it is to ask for permission.

Don't get me wrong, I would love for all of these people to end up in the (Scientific) health industry. (It's worth mentioning Scientific, as there is a lot of useless crap that uses some form of learning and offers no real gain - would be disappointing to see them all end up applying deep learning to homeopathy for example.)

Not unless they're willing to take a paycut.

Well advertising and marketing firms don't pay there tech talent very well.

You are begging the question.