Ask HN: What jobs have been taken over by AI/ML? | Hacker News Reader

Ask HN: What jobs have been taken over by AI/ML?

8 pointsn_t9 comments

The news cycle of jobs getting taken over by AI/ML/DL keeps repeating, but can someone provide few concrete examples of which jobs have been already taken over or are in the process of getting taken over? Machine automation took over many industrial/manufacturing jobs but that happened many years ago already. I am looking for examples where specifically AI/ML/DL technology was reason for job vaporization.

View on HN

AI/ML has not matured enough to take over any jobs. However "computing" in general has replaced a lot of jobs such as flight scheduling, CRM voice calls, analysts who used to pore over crime data have been replaced with excel statistic functions etc.

OCR Basics: "Reading and Sorting Mail Automatically" 1970 US Postal Service

https://www.youtube.com/watch?v=cdfYrpRHqJ0

Right now AI is displacing man hours, not people. Police are able to eliminate costly evidence gathering with video surveillance and LPR. More capabilities are coming available now.

When the economy blows up again, that’s where AI will start eating jobs. You’ll be able to afford the capital investment in technology by shedding people.

My guess is that the most initial impact will be for supporting the operations of traditional process based computing. First they’ll come for the sysadmins and low level BA types, then the armies of CRUD developers slinging java.

Honestly, “AI” can mean many things. My employer used to have many people managing old school SANs. Cloud storage makes them moot. Is that AI? Probably not. But I would certainly pitch it as AI related to claim my share of the open savings.

Spying on citizens.

You mean something like Sensetime. The political aspect aside, here no jobs were replaced but actually created.

I guess credit approval was taken over somewhat by credit scores but that was quite a while ago.

There’s some basic ml in generating credit scores. Mapping credit information to default rates. Logistic Regressions I believe.

As you said there might be some "basic ML" involved in credit rating but that can probably replaced by mean/median/mode kind of computation easily. Although this is one area which ML can completely take over, i.e. instead of maintaining a static score, ML can predict the likelihood of an individual returning the loan based on gender/age/income/zip code and tonnes of other info.

Facebook's powerful lookalike ML means lots of account ad op managers who used to be necessary to get good online targeting are obsolete.

Translation services

Reply on news.ycombinator.com