Robots Are Coming for These Wall Street Jobs | Hacker News Reader

Robots Are Coming for These Wall Street Jobs

bloomberg.com

73 pointschampagnepapi73 comments

Kevin Slavin gave an excellent talk [1] on the need for trading systems to be built with humans in mind rather than just for beating the other bots. I highly recommend it, it's certainly one of my favorite talks.

[1] https://www.youtube.com/watch?v=9Goa1Y9OBHU

Good talk.

And I, for one, am confident that Wall Street will accept the limits of their intelligence and introduce software constraints that require human oversight for the benefit of market stability. /s

Seriously though, any system with many actors is going to see unanticipated failures resulting from runaway interactions. There is no way around it. And I think that if the threat of losing money isn't a powerful enough incentive to stop it, then nothing is.

I still think there is too much hype around AI with all the doomsday prediction. Yeah, data processing will get automated away, but most jobs that require thinking within an ill defined framework will be fine for at least the next 20 years. The human brain is vastly superior to any machine learning algorithm we currently have.

I think a big factor that people are failing to take into account is time to market and deployment. Imagine we had amazing AI that could do a lot of office jobs now. We’d still need to make useful software out of it, distribute and sell that software, customize it for different needs. That, by itself, could take years.

Forecasts are being pushed in the mainstream media that tout that 50% of all existing US jobs will be wiped out in the next 10 to 15 years. It has reached manic levels.

35-50 years, probably still wildly over the top. 10-15? Zero chance. If you had the perfect solution in a lab that could, for example, replace truck drivers today, you couldn't even get that fully implemented & distributed such that it cleared out 50% of all trucking jobs in 10-15 years. The entire economy? It'd be laughable if there weren't so much fear mongering involved in pushing these bogus forecasts.

It's another example of people over-estimating change in the near-term. In 1999, ecommerce was going to immediately threaten all physical retail with extinction. If you didn't get on that hype train, you were toast. 23 years post founding, with year after year of incredible growth, Amazon is still just 1/4 the size of Walmart. As it turns out, it more often takes a long time to change the world.

Forget the human brain, we are a long way away from doing what an ant does with it's spec of dust size mind. And that still requires acres of racks.

But the hype does help with funding. Which is a good thing.

Firstly, the human brain is not vastly superior. There are already tasks such that a reasonably competent developer can build and train a deep learning model that beats human perception on commodity hardware with open source software.

Secondly, the number of jobs requiring thinking in an ill-defined framework these days is shrinking rapidly.

This is not specifically caused by machine learning, but rather by the desire for predictable timelines / cost savings / metrics gathering. But it has set the stage for AI automation.

Take law and health care, for example:

* Law firms are consolidating, and lawyers can now spend their entire career working on one small, specialized aspect of a field.

* Primary care physicians barely are allotted just enough time to refer you to a specialist. The specialist has just enough time to order tests.

Yeah, on the ImageNet data set, where all the images are pretty unclear, low res and cropped weirdly leaving out context, deep learning "beats" human vision system.

Human brain beats everything for now. It is a much more generic system.

I don't know enough about the specific fields you talked about to comment.

Deep learning is an incredibly powerful tool, but we should differentiate between its actual merit and the huge hype around it, mostly done for PR and misleading investors/financial analyst.

You are moving the goalposts.

If we're talking about doing a specific task within a well-defined framework, computers are already proving to be superhuman. That includes image recognition beyond ImageNet, as well as games like Go.

But yes, if we're talking about a generic system, then humans are better. And I think they will be for a long time.

And yes, I agree, there is a lot of noise and sizzle around AI (but there is also quite a bit of steak there too.)

Were they ever making any meaningful difference? How many of them have beaten the returns given by an index fund over last 20 years? Also, I hate the sensationalism of such articles, more sensationalism usually means less sense.

Quite a few. Famously Medallion:

https://www.bloomberg.com/news/articles/2016-11-21/how-renai...

I know about medallion but its also run by quants and afaik largely algorithmic. My argument is against the more traditional fund houses and stock pickers.

Most of the jobs listed are on the sell-side and predominately in market making roles.

What is market making?

http://www.investopedia.com/terms/m/marketmaker.asp

tl;dr They provide liquidity to markets.

tldr their wealthy benefactors throw money at them and they in turn throw money at an unexplored corner of an unintuitive industry

If you want an inadvertent AI doomsday scenario, how about black-box trading models that figure they can make money by betting on a market crash caused by a war, then manipulate the market to induce an economic conflict between different states, without every really having any understanding of the meaning of the outcomes they're optimising for.

AI doesn't work that way.

Not yet.

You might find the history of Doug Lennat's Eurisko informative.

"Lenat fed the massive Traveller rulebook into the system and asked it to find the best way of winning. Each night, after several hours of trial and error, Eurisko would spit out a few strategies—some interesting, some impractical, some ridiculous. At one point, Lenat remembers, it suggested he could win the game by changing the rules."

https://www.wired.com/2016/03/doug-lenat-artificial-intellig...

ah this was Traveller book 5 Trillion Credit Squadron Publish in 1980 or so.

The authors commented that having a computer (when Pc's where ultra rare and expensive) would be useful :-)

I commented at that time that a PC was essential and working at the cutting edge of AI would be "useful"

Traveller (book3) was the basis for Elite as well

Bear in mind that Lenat won't release the Eurisko source code, and some people have speculated that it didn't do all that he said it did.

I used to work for Doug and let’s say there is some “dramatization” in that characterization.

Also Eurisko and other symbolic/semantic systems are fundamentally different from the NNs popular today.

AI will eventually be better at doing what the best people do.

Although I agree with this sentiment normally, this doomsday scenario would require a great degree of cunning and deception. An AI can succeed without practicing this kind of manipulation. I'm certain other humans/AIs would be able to identify this and build around it (or maybe even pull the plug)

The scenario may work with a human actor/corporation working with the AI, but then it's no longer that cyberpunk (and could probably be done with a large enough grid today).

AI is generally strictly bounded. It only looks at certain data and can only take certain actions based on that data.

Which is how I feel about most of the supposed Ai paradoxes like "The stamp collector problem" or even "Grey goo". They're like a 10 ten ton block of ice in the middle of a sun. Sure, it's physically plausible, there's nothing in the laws of physics that says it can't happen. It's only that there's no path leading up to it happening.

I think grey goo is fairly realistic. Having small, self-replicating machines has many useful applications. Flaws during reproduction could trigger a runaway replication scenario.

The latter already happens with the existing small replicating machines that we know. We usually call that cancer.

Life is already tiny self replicating machines, it still can't replicate inside of a wood wall very well. Further, life and nano-machines are stuck with the same building blocks. Outside of organic chemistry there are not a lot of options for self replication. At which point your nano machines are also just food for existing nano-machines with billions of years of evolution behind them.

TLDR; The problem is chemistry not design.

Wood wasn't digested by anything for millions of years. It seems not completely impossible that whatever we use to create the nanomachines wouldn't be eaten by anything. But of course you're right in stating that it's unlikely that we'll produce machines that are capable of self-replicating using virtually any available substrate. So the runaway scenario where the whole Earth is turned into grey goo is unrealistic.

Small bits of Wood get eaten by any number of organisms and much faster when wet. https://en.wikipedia.org/wiki/Dry_rot So, it's probably a poor choice for building blocks if you don't want the robots eaten especially if you want the to move around..

I was referring to the carboniferous period where literally nothing could digest wood, wet or not.

> where literally nothing could digest wood

Yea, that's not actually well supported. There is no where near enough coal for all wood over that time span.

Larger creatures like termites where probably missing. But, micro organisms don't eat dry wood on their own very quickly.

AI doesn't actually know what it's doing. It just optimizes for hidden patterns.

I wrote a whole blog post on this but tl;dr AI/ML can do unethical things without anybody knowing.

Of course, the trading bot AI assumedly would only have outputs as either buy or sell shares. But what if the AI after a long time of analysing news and correlating it with stock prices determines that the best returns are somehow by buying and selling in a way that crashes the market temporarily.

I guess the central point is that even if they aren't advanced enough to enact OP's scenario, they do not understand negative repercussions outside of their specific domain and so can unscrupulously cause havoc by proxy once they get smart enough.

AI works exactly that way. Reinforcement learning involves training a model using some input features, some outputs, and a value function (usually, how closely the output(s) match a known value in the training set).

Consider a model whose inputs are market prices, outputs are market transactions, and which has a naive value function that simply optimises for the amount of profit is trades result in in a relative short space of time. It's plausible that it could detect trading patterns that correspond to market crashes that can be exploited for profit.

These trading patterns might actually be causing or supporting political / economic shocks, such as bubbles, corrupt or unethical behaviour or inter-state conflict that then cause these crashes, but, to the model, this would be invisible or irrelevant. It would learn to perform trades that encourage these same shocks (e.g. undermining a large employer, causing job losses that cause political anger), and profit from the consequences. All without ever "understanding" what is happening outside of its model or having malevolent intent.

Unfortunately, humans do.

Winning on other people’s disgrace? Already happened, The Big Short, etc

Accelerando?

Better yet, how about black box recommendation models that figure that people are more likely to click on things that outrage them, who don't realize that the end result of viral fake outrage can be war?

But that one would be current reality. People's tendency to click on things that outrage them (and therefore generate advertising) is one of the things that lead to social networks becoming echo chambers of various conspiracy theories. The result is sites like InfoWars, which helped make Donald Trump president. The war bit hasn't happened, but it well could.

My favorite book about this topic is _Trust Me, I'm Lying_. It lays out the human incentives that guarantee that the recommendation engines will produce terrible for us results.

> The war bit hasn't happened, but it well could.

Give it time.

tl;dr: data processing is getting automated. This is just a dramatisation (and a bit of a summary) of normal process automation. Nothing to see here.

Dramatization? How so? Is the continued loss of middle and upper-middle class jobs to outsourcing and automation not noteworthy?

Also, if you are a regular BBG reader, and anything in that article is news to you, than you should be worried and get updated.

So can we expect an influx of developers who were former traders?

That's me. I switched 3 years ago because I saw exactly this. I originally did it to gain more tech skills with the plan on going back into it, but I left and didn't go back. I took a bootcamp, and about 1/3 of my class were ex-wall street.

In your opinion how hard is for average ex-wall street workers to pick up programming ?

Really depends on what role they were in but the Bloomberg Excel API is very widely used along with VBA which I think can assist with some transition.

Finance is a huge industry. I can't really speak to the average, but I was a buy side equities trader.

Technical: All of us are well versed in excel and have working knowledge of SQL. Going from VBA to ruby/javascript was pretty trivial. While I don't really consider college all that meaningful, most people had been STEM majors, so the non job-specific knowledge tends to have some overlap. I was a math/econ double major with a stats minor. Had I dropped 2 econ courses and taken data structures + computer architecture, I would have had a minor in CS. The rest of the minor overlapped with my math/stats.

Work-life: Finance is a very hard-working industry and had a ton of continuing education, so that transition was extremely easy to the tech life. I previously worked 60-80 hours a week (dependent on earnings season) and went from that to 50 hours. All that free time went towards a few hours of extra sleep and MOOC work to catch up on some of the more fundamental CS concepts.

the current generation being hired into wall street (people around my age) have some kind of engineering, mathematics, physics, CS background in huge amounts. finance hires people like these by the boatload into all kinds of positions: risk, research, trading, sales. software development permeates the entire value chain in the industry.

Those infographics: unless they render a lot differently on non-iOS devices, Tufte must be having a coronary.

This doesn't even touch on any of the back/middle office jobs that are going to completely vanish in the next few years. Most of these people have a very narrow skill set that is not transferable to other fields.

The title should read "Robots Are Here..."

Great read for anyone interested in a history of the Wall Street bots: "Automate This" by Christopher Steiner

And, as per typical, the regulation won't start happening until the white collar workers making 120k+ start losing their jobs.

It's been happening for quite awhile: https://www.wsj.com/articles/end-of-an-era-as-cme-to-close-a....

Unrelated but if anyone's interested, KloudTrader is hiring for fintech developers. Looking for senior software engineers (fullstack web) and a CTO. We are an early stage startup that is willing to offer significant equity. Stack is Rails and python scientific software. Email's in profile!

Stephen Hawking says we should be more frightened of capitalism than robots

Who is behind robot? capitalism

I guess we just need some good ol' communism to fix this problem.

Maybe? Kind of?

I'm extremely pro-capitalism, but free market economics does rely on a number of assumptions being met in order to function, and only maximizes utility functions for those participating.

If you build a market that violates those assumptions (e.g., massive consolidation of supply) and effectively excludes people from participating (by reducing total volume of work amenable to human intervention), capitalism ... just doesn't work. Wrong tool for the job.

Traditionally we either regulate operations to make them better fit free market assumptions, or regulate outcomes. Whether that looks like communism is a secondary discussion, though if you consolidate supply heavily enough, then yes, you end up with some central supplier choosing what to produce and how to allocate it for all those people that have no resources to create meaningful demand.

I think we can make supply distributed (given micro manufacturing/recycling etc) so even if it is less efficient for a community to do it than something more centralised they still have some autonomy. See my profile for my website if you are interested where I am coming from.

I meant centralized in terms of ownership/control, not geographic placement. The current trend towards winner-take-all automation and resultant oligarchy.

So do I. Consider the technological extreme of open source manufacturing. A set of tools like repraps, that that can really make themselves (and computers etc)

It's pretty sad that whenever someone criticises capitalism that people immediately shout "communism" to kill the debate.

It's ultimately humans behind the robots. So what we should really be wary of is the values and assumptions of those who develop the robots.

>"When machines and computers, profit motives and property rights are considered more important than people, the giant triplets of racism, extreme materialism and militarism are incapable of being conquered."

-Dr. King

You cannot take away property rights and still claim that people are considered important. (Didn't King notice the word rights in property rights?)

> (Didn't King notice the word rights in property rights?)

What's that supposed to prove?

Capitalism, Communism, Socialiasm, and Collectivism all have property rights. They are structured somewhat differently, but they all recognize that some things can be owned, and some cannot.

You can't build a consistent legal, or ethical system from the first principle of "Property rights are the most important right."

He didn't say "take away property rights"

Yes; but rather that they aren't more important than people. Well, that is actually silly because under the right circumstances, it could be that the property (or any other) rights of person A are more important than all of person B as such. It's just trivially true when A = B (A and B are the same person). No right attributed to A is more important than A him or herself.

>it could be that the property (or any other) rights of person A are more important than all of person B as such.

How would you determine this? Is this not the same argument made by slave owners?

Stephen Hawking should stick to saying things in his area of expertise. I wouldn't expect Bernie Sanders, Slavoj Zizek or Elon Musk to start making pronouncements about black holes.

Trading, Sales, Portfolio Manager, Analyst

Reply on news.ycombinator.com