No fully automated slaughterbots. There has to be a man in the loop that at least clicks a mouse every time the robot kills an enemy combatant. No one should ever say: "Oh yeah, we had a bug in the auto-kill algo and our killer robot shot a whole bunch of innocent civilians. Whoops! That's nobody's fault. Nobody's responsible. The algo did it all by itself!"

In the context of a military that accepts collateral damage as par for the course, I don't really see the distinction. Did anybody but Manning go to prison over that Collateral Murder video? We have humans approve airstrikes that we know will kill unidentified civilians when they're close to an identified target. We have humans approve airstrikes that target unidentified individuals which we probabilistically identify as combatants based on movement and cellphone metadata. What difference does it make if a human pulls the trigger --err, presses the button?

I'm completely with you, but eventually "we" won't have a choice when adversaries go full autonomous.

Now if someone will make some sort of union I can join so I can strike and support others who share the same fundamental ethics. :)

Out of curiosity... why won't you work with HFT?

Yeah, that one surprised me too. I've read Flash Boys, but I've read a lot of pushback on it, including the idea that it's essentially a very one-sided pitch for IEX.

There are some good arguments that the entire point of HFT is to manipulate the market by front running people who are trying to make actual transactions. Here is a cool video of a few hundredths of a second: https://www.youtube.com/watch?v=K4ENWtfeYp8

There are many purposes for HFT, most of which have nothing to do with front running. Market making is an entirely legal and moral activity which benefits "people who are trying to make actual transactions" and happens to gain a lot from better tech.

Market making is indeed a useful activity, in isolation. And market making is indeed technically what HFT companies do.

However, pouring unspeakable amounts of money, watts and person-hours into completely unnecessary and wasteful infrastructure just because emergent properties of a system mean that billion dollar companies and their owners can make a lot of money by being a picosecond faster than each other in an eternal battle is of extremely questionable utility.

What exactly do we as a society get in exchange?

We get markets that are incredibly fair and nearly frictionless on a global scale.

How much fairer and frictionlessier do they become for every picosecond? Is there a limit to fairness and frictionlessyness?

Arguing about HFT feels like you're questioning people about why they're spending billions on excavating single, individual sand grains and the defense you get is "well, sand is an incredibly important building material. It is used in concrete, and we wouldn't have modern society without it". Sure that's true, but why is it that nobody can tell me how what you're doing is actually necessary for that.

Also, from a control engineering point of view, a lot more unstable.

With market "liquidity" that evaporates in milliseconds exactly when it's needed most...

Better HFT than bitcoin...

> Market making is an entirely legal and moral activity which benefits "people who are trying to make actual transactions" and happens to gain a lot from better tech.

Does that require HFT, or just T?

The HF bit lets prices be adjusted faster, which means bid-ask spreads can be narrower.

That's the theory, anyways.

Then add the caveat "outside regulatory frameworks" to the orginal statement.

It’s amazing when people are vehemently against something and take a strong moral stand, but you dig a little bit and realize that their basic facts are wrong about the issue.

It is illegal to offer a security for sale that you have no intention of selling. I'm pretty sure the way they make money is to flash out orders and see if there is a response for the now unavailable security, allowing them pricing signals they wouldn't otherwise have. The basic solutions people offer are often of the theme "require the order to live long enough for others to actually trade on it".

> I'm pretty sure the way they make money is to flash out orders and see if there is a response for the now unavailable security

No.

Market event A is seen to correlate strongly with market event B. The correlation is so obvious that multiple participants compete to make the B trade first in reaction to A. It's part of having a healthy market, electronic or otherwise.

No software-only interlocks on anything that requires a safety interlock

While I think these are incredibly important and extremely neglected in the industry, these ethical concerns are really something different from the spirit of the "Things I won't work with" series, which is mainly about safety and sanity.

No voting-machine software. Paper ballots guarded at all times by well-armed disabled vets. Ballot-scanners produce a paper checksum slip for each voter that they can later compare online to verify ballot has been counted correctly.

Why would your first rule be so limited? Why help anyone spy on anyone else?

Are you against spying on, say, white supremacists who you know are plotting terror attacks on minorities?

How would you know they are plotting terrorist attacks before you started spying?

Tips to law enforcement, violent history, threats.

Anything that can be used to "spy on white supremacists" will almost certainly be (ab)used to spy on other people. There are very few real, actual white supremacists, and even fewer of those who are "plotting terror attacks". In recent memory (~15 years) I can only recall three major terror attacks against minorities perpetrated by real, actual white supremacists: Breivik, the Christchurch shooting, and Dylann Roof.

Besides the definition of what constitutes a "white supremacist" has been getting more ambiguous by the minute over the past few years. Nowadays it pretty much just means "anyone who doesn't like illegal immigration" or "anyone I don't like, but I'm too lazy to articulate why".

All of this is completely off topic from my post. "White supremacists" were just an example. It could have just as easily been neo nazis, or Islamic extremists, or whatever.

Your post is similar to if I proposed a hypothetical about addition using apples, and you posted a screed about how apples are overused in American cuisine and their nutritional value is suspect, especially relative to oranges and pears.

All of those groups represent a tiny subset of the population ill suited to mass surveillance dragnets and well suited to on-the-ground policing and community engagement (as in, getting sane members of those groups to snitch on the insane).

Just the other week I saw the news that until recently we had 300 (!) Saudi nationals training to fly planes right here in the US. Software won't help us with stupidity/sabotage of this magnitude.

https://abcnews.go.com/Politics/navy-suspends-flight-trainin...

You can add the Pittsburg synagogue shooting to that list.

I'm not in favor of surveillance panopticons either, for what it's worth.

Seems more of an antisemite rather than white supremacist per se. Didn't publish any manifestos either. But I'll allow it.

Still, that's very few events, which supports my point. It's basically a more potent "think of the children" trojan horse that routinely gives the government even more power (and potential for abuse) through harebrained and ineffectual laws. See e.g. the current FISA abuse scandal where the FBI _falsified the supporting documentation, and knowingly supplied fake documentation to the court_ to spy on a presidential campaign. Had it been you or me they spied on, nobody would even know.

I think you'll find that isn't the case: https://en.wikipedia.org/wiki/Pittsburgh_synagogue_shooting#...

Again, I agree with you that domestic extremism (be it white supremacist, Muslim, or Black Hebrew Israelites) should be countered through ordinary policing, rather than building the Stasi's wet dream.

Charlottesville car attack?

Spying is not necessarily immoral, I presume.