This is more complicated than it seems. You are comparing a job (HFT) with a pursuit (science). The actual comparison is HFT with academia (the job part). Academia is a terrible job and getting worse over time. Academics have been talking about how the job is bad and getting worse for 100 years. There is a long list of problems. Working in science is also an exercise of joining a community where your peers seem willing to put up with the terrible working conditions, almost to the point of masochism or martyrdom.

The other issue is that intelligent people tend to feel entitled to success, particularly if they succeed on traditional measures which rank them according to others. This makes them less likely to enter very difficult and uncertain fields when there are other options. Science is often inhumane in a sense, you can waste your life on the wrong idea, certainly many people have. You can also be right, but only proven so after you die.

Science also has developed a large swarm capacity, which reduces the ability of an individual to succeed and make an impact. The best example is deep learning... the explosion in deep learning research output is frankly astonishing. In real terms, a lot of researchers outside the field managed to drop whatever they were doing at the time and switch to deep learning research. Along with the excellent tooling, accessible hardware and data and low barrier to publication, there is literally a flood of papers. It is now very difficult to get noticed, and much harder to make any progress. This has happened in the space of one PhD duration. It is reasonable to assume this will keep happening as new fields open up. I presume it is because there are large numbers of people waiting for the next big thing, underemployed or underutilised in their own current work. In this scheme, research success comes to either those that found a field (even more difficult and unlikely) or have the best resources (often corporate research these days). Your talent, intelligence etc all become secondary. You can say that this is just 'rapid progress' and good for humanity. Maybe. There is also a lot of low quality work. But in the end one is an individual, not humanity, and you need to act on that basis.

All these factors conspire to send smart people away from science, and it is hard to blame them.

>> Science is often inhumane in a sense, you can waste your life on the wrong idea, certainly many people have. You can also be right, but only proven so after you die.

I’m no scientist, but the above really resonated with me.

Yep. Some of the brightest minds in the world, including many of my friends, are wasted figuring out how to make more money from online clicks.

Conversely: science pays junk money compared to IT. I have a Masters in chemistry, my wife has a Ph D. I make more then twice what she does off of skills I taught myself, and when I'm between work recruiters call me.

If we want more people in the sciences, we've got to create the jobs.

That's what my chemistry teacher told us in high school. If you want to really 'make it' in chemistry you better be insanely good, which is has a small probability. Otherwise you sorta end up as a lab technician.

Made me choose the other thing I loved in high school which was math and computer science.

Haven't looked back.

Once you have enough liquidity that the markets function (i.e. the stuff of actual value to people doesn't get jammed up) it seem the marginal utility of adding liquidity is near zero.

Have I missed something?

capital deteriorates and becomes obsolete so executives are constantly maintaining and optimizing their businesses as they respond to market conditions. These trades are how the market chooses who is better and worse at this constant maintenance/optimization. unfortunately capital isn't something you can set up and allow to run indefinitely.

I’m pretty sure it can be allowed to run unattended for several milliseconds at least.

it can but this is "late" capitalism, we are at a point where capital that updated every year was outcompeted by capital that updated every quarter, which was outcompeted by capital that updated every month, and so on. Each time this happened there were efficiency gains that required everyone to keep up or go out of business, and the efficiency gains resulted in surpluses that were allocated somewhere in the economy. If we turned back the clock then these businesses would become less efficient and we would have to sink more resources into them to continue the same output.

Iterative optimization resulted in smaller and smaller (and more frequent) corrections.

I see where you are going here, but it seems to be many orders off magnitude off in time relative to HFT.

its highly counterintuitive to see the modern system in the same terms as the capital markets of last millennium. and I'm convinced the system isn't perfect, I'm sure there are things that should be changed. I'm merely suggesting that the market sped up as people learned to do market stuff faster, and that market stuff is ultimately a process of allocating capital resources.

Ok, I get that - I don't think it really addresses my concern/question about whether or not this is far in the range of diminishing utility.

Put it another way, if the number of HFT trades were suddenly cut in half, what 1st order impacts would be felt by people outside those markets? The heart of the "but we're providing liquidity" argument is that this is a service provided that has real value outside the market. If that's not true the particular argument falls apart, doesn't it?

> far in the range of diminishing utility.

it probably is because of distortions in the flow of money

> Put it another way, if the number of HFT trades were suddenly cut in half, what 1st order impacts would be felt by people outside those markets?

it would probably result in capital markets taking longer to account for changes in the market/world such as shortages. Then the capital markets that weren't subject to the restriction would be more competitive as a whole. Depending on the specifics, it might open the door to people with relevant information (who weren't hf traders) to begin to make some of those trades.

> The heart of the "but we're providing liquidity" argument is that this is a service provided that has real value outside the market. If that's not true the particular argument falls apart, doesn't it?

I don't use the liquidity form of the argument because I think it focuses on the wrong things. Capital markets affect the rest of the markets by optimizing the creation and maintenance of capital. All consumer goods rely on capital. So policies that make the capital markets less efficient would have broad downstream effect in commodities markets, essentially there would be in aggregate fewer goods.

I haven’t seen any claims that HFT is allocating capital in better directions. They’re mostly reacting to the same signal and trying to trade the same asset at the same price, right? That’s the only case where the latency matters, because the market uses order time to break ties.

the market allocates better, the hot allocates faster. the information (prices) is better because its more current.

>As a result, it attracts people who should be out there curing cancer, developing new technologies, researching fundamental sciences, and so forth.

I'm in HFT, and I'm confident in speaking not only for myself when I say that if HFT didn't exist I'd be doing something else fun and profitable (e.g. investment banking, non-high-frequency trading), not something boring and/or poorly compensated like medical research. People on this site have no right to tell other people how to live their lives, especially when they themselves probably work on something like ad-tech which is arguably actively very harmful to democracy and society.

I like that term "“bug” in capitalism". I have always felt the "bug" is basically incorrectly pricing externalities. We don't properly charge you for the carbon you put into the atmosphere or the mental health strain you put on workers and so on. If societal costs were also real costs the allocation of resources could be more optimized.

It's an interesting of reasoning but comes with a caveat - Reducing externalities to a price figure allows you to trade them away on a balance sheet as nothing more than a price figure. This is fine if pricing accurately reflects value, but that's not the same as allowing a market to set the price. You lose a huge amount of detail in the process that can't simply be captured in a price signal moving up or down.

You may end up with things that have infinite price, at which point I think we're begging the question.

There's also a question of how commodified the things we're talking about here can actually be - If you can sell it one day and it's impossible to 'buy back' tomorrow at any price, then it's a fundamentally different kind of trade that doesn't really fit the framework as neatly as first appears.

Without pricing the externalities, we're basically taking on debt, a debt against the future number of years the planet is habitable. I imagine it probably started at 1M years of supporting human life. I wonder if that number is now down to 50000 left. And I feel like it went from 1M to 50K in the last 50 years.

Indeed, it seems the longer we run under the status quo the greater the cost in externalities, to the point we’re beginning to understand represents an existential threat globally.

However we also risk a self-fulfilling prophecy if we commodify those externalities and reduce them to lines on a balance sheet. I know that this is not quite what’s being argued for, but I am equally certain that is how it would land in practice.

I’m not convinced re-framing/codifying our social, environmental, and human capitals as financial instruments is any kind of progress towards salvation, rather it’s a doomed attempt to take everything that never fit inside the tidy box of capitalism and cram it inside of the box anyway. A try at making the world comprehensible to a mind accustomed to that manner of framing.

This problem needs novel thinking, more novel than finding a direct way to account for it within the existing framework. I regrettably don’t have a better idea.

Which social costs? Is it a social cost having too few kids? Or too many kids? Should we price that in? And what’s the cost we should price in? Who decides? Can we all agree on one price?