Robots will soon make bankers and many more jobs totally redundant

[derspiess]

And I thought ATMs were supposed to totally replace bank tellers by the 90s.

My bank already did this.  No tellers.

This will not happen 'soon' unless by soon you mean in 50 years or more.

'Soon' in this case was four years ago.  Seems to be working fine.

Yeah, I have an account at a bank that started deploying tellerless branches.  Yet, a vast majority of banks still use tellers. 

[Jacob]

The result will be higher unemployment, higher lottery sales and more people speaking out against lazy welfare bums.

That sounds about right.

[Baron von Schtinkenbutt]

Yeah, I have an account at a bank that started deploying tellerless branches.  Yet, a vast majority of banks still use tellers.

Yes, but the vast majority of customers don't use them.  I can count on one hand the number of times in the past year I have actually spoken to someone at my bank.

[The Brain]

Banks with tellers? Jesus, you guys still have post offices as well?

[The Minsky Moment]

While Ideologue's concern for the future of investment bankers is touching, perhaps Wharton should hold off a few more years before closing its finance programs:

Algorithms are already commonplace on City trading floors, and are used in many industries, from online retail to internet dating. High-frequency trading, governed by algorithms, is already one of the most profitable trading classes.

Oh yeah, right.  Algorithmic trading is huge.  And it hasn't eliminated bankers, it's added a whole new layer of them to develop and program the algorithms, and integrate them into the broader set of things that banks do.

But, according to Mr. Coplin, in 10 years people will no longer be required to manage these algorithms. Decisions will be taken directly by the artificial intelligence

No doubt Microsoft's "chief envisioning officer" knows a lot more about the future capabilities of computers than I do, although apparently not enough to prevent MSFT from getting its clocked cleaned in every new tech evolution and market since the Browser Wars of the 90s.  But this needs a little more explanation.  Like what does he mean about "managing algorithms" - I don't think anyone does that now - algorithms do run themselves, no one needs to turn a crank.  What are the "decisions" that are going to made directly by AI?  Is this for prop trading, retail wealth management, brokerage operations?  The fact that Mr. Collin appears to have no professional experience or exposure to the financial services industry gives me pause.

[The Minsky Moment]

History has shown that all economic shifts in the modern era ultimately result in not just greater wealth, but wealth more broadly shared, resulting in a higher standard of living. 

Not a true statement.
However, if the focus on the period since 1978 it would be true if one looked at the world economy as a whole.
For the prior 30 year period it would be true for the US, Western Europe and Japan. 

[mongers]

I just read a review of one of the better robo-vaccum cleaners, it encountered a fresh cat turd and preceded to very efficiently cover one whole level of his house with a fine coating of shit.

[alfred russel]

I just read a review of one of the better robo-vaccum cleaners, it encountered a fresh cat turd and preceded to very efficiently cover one whole level of his house with a fine coating of shit.

Robotics are proceeding faster than I expected; it seems the subprime mortgage executive robot is ready to take over.

[Ideologue]

While Ideologue's concern for the future of investment bankers is touching

Look, there are many problems, one solution.

[grumbler]

I just read a review of one of the better robo-vaccum cleaners, it encountered a fresh cat turd and preceded to very efficiently cover one whole level of his house with a fine coating of shit.

Yeah, if you leave a big puddle of paint on your floor, you will get the same thing.  In fact, if you vacuum using a regular, non-robotic vacuum cleaners, and run over a fresh cat turd, you will manually cover one whole level of your house with a fine coating of shit.

[Tamas]

I have an easy time saying good bye to these luddite fears when I think of my previous job, where a huge building was full of around 1000 IT guys, and these were the ones actually supporting and overseeing the computers others worked with.

So all of this will just create new kind of jobs. Only possible concern is that there will be less and less jobs you are capable of doing if you haven't learned anything past 3rd grade in school, but boo fucking hoo. I, for one, am not ready to stop human progress just so lazy people can remain comfortable.

[DontSayBanana]

While Ideologue's concern for the future of investment bankers is touching, perhaps Wharton should hold off a few more years before closing its finance programs:

Algorithms are already commonplace on City trading floors, and are used in many industries, from online retail to internet dating. High-frequency trading, governed by algorithms, is already one of the most profitable trading classes.

Oh yeah, right.  Algorithmic trading is huge.  And it hasn't eliminated bankers, it's added a whole new layer of them to develop and program the algorithms, and integrate them into the broader set of things that banks do.

But, according to Mr. Coplin, in 10 years people will no longer be required to manage these algorithms. Decisions will be taken directly by the artificial intelligence

No doubt Microsoft's "chief envisioning officer" knows a lot more about the future capabilities of computers than I do, although apparently not enough to prevent MSFT from getting its clocked cleaned in every new tech evolution and market since the Browser Wars of the 90s.  But this needs a little more explanation.  Like what does he mean about "managing algorithms" - I don't think anyone does that now - algorithms do run themselves, no one needs to turn a crank.  What are the "decisions" that are going to made directly by AI?  Is this for prop trading, retail wealth management, brokerage operations?  The fact that Mr. Collin appears to have no professional experience or exposure to the financial services industry gives me pause.

Yes, algorithms do need to be managed.  Many different algorithms could get to the same result with differing operational complexity and performance drain.  It's not just a case of picking the shortest route, either- especially in large database situations, the optimal algorithm from a complexity standpoint could put an untenable strain on system resources.

It was a poor choice of words, but what he's getting at is that computers have reached the operational capability and flexibility that once we've wound them up, they need almost no human interaction apart from tuning and repairs, and computers are starting to get to where we could an algorithm-based program to tune itself.

I'm going to go one further on this guy, in fact.  These units are hitting so much computational power with such robust connections that I'd be willing to wager in the next few decades, we'll start to see more devices like Chromebooks, where all the heavy lifting is done by just a few servers, and the houshold devices are decreasingly complex but better-connected to the networks.  Eventually, the IT market is going to collapse to a few highly-fought-over positions maintaining a few high-powered devices.  Something like the Internet's current dozen or so root servers, applied to a broader category of devices.

[Ideologue]

That seems scarily plausible.

It will all be fine, though.

[DontSayBanana]

That seems scarily plausible.

It will all be fine, though.

Software guys should be fine for a while longer.  One of the few bright lines that hasn't really been crossed in systems development is that we still prefer humans for "design" tasks.  While we design systems to be "intelligent," we don't design them to have "aesthetics."  Once we've given a mechanical intelligence freedom of choice, a mechanism to choose, and Skynet doesn't happen, that will probably change, but until then, machines are programmed to perform a task and not be picky about how they do it.

TL;DR- hardware guys will die out because complex hardware won't need to be as ubiquitous as it is now.  Software guys are (for now) okay because we don't design programs to design programs.

[The Minsky Moment]

Yes, algorithms do need to be managed.  Many different algorithms could get to the same result with differing operational complexity and performance drain.  It's not just a case of picking the shortest route, either- especially in large database situations, the optimal algorithm from a complexity standpoint could put an untenable strain on system resources.

It was a poor choice of words, but what he's getting at is that computers have reached the operational capability and flexibility that once we've wound them up, they need almost no human interaction apart from tuning and repairs, and computers are starting to get to where we could an algorithm-based program to tune itself.

I don't see anything in this description that would be a big hit to human investment bankers, just on some of the back office IT folks. Even there I am dubious based on my own admittedly anecdotal interactions with corporate IT systems.