Stuck In The Middle With You (Common Ground Thread)

[steve dave]

Marc Andreesen is not someone I want in charge of about anything. I say this mostly as a CNBC watcher.


Sent from my iPhone using Tapatalk

[ben ji]

How can anyone be excited about AI drones?

Think of all the police we can defund. Just have drones hovering all over the place, someone robs a store? The drone tasers them. Some college kids rush the field after a big win? 2,000 taser drones descend. Bet that wont happen again.

[nicname]

He might build the most efficient rockets! But I don't trust him!

I also don’t trust him.

[Pete]

Of course he’s untrustworthy.  The point is that his business is a model for how others can disrupt the most corrupt industry on earth.

I want AI making the kill decisions in war. Much better decision making ability than sleep deprived, morally deprived, angry, scared, hurt, intoxicated, vengeful humans.

[michigancat]

Of course he’s untrustworthy.  The point is that his business is a model for how others can disrupt the most corrupt industry on earth.

What model is that?

I want AI making the kill decisions in war. Much better decision making ability than sleep deprived, morally deprived, angry, scared, hurt, intoxicated, vengeful humans.

There's a ton of stuff to unpack here, but Pete, who do you think tells the AI how and when to kill? I mean the programmers might not be sleep deprived but they'd still be everything else on your list.

Should AI be able to launch nuclear weapons?

[Spracne]

Of course he’s untrustworthy.  The point is that his business is a model for how others can disrupt the most corrupt industry on earth.

What model is that?

I want AI making the kill decisions in war. Much better decision making ability than sleep deprived, morally deprived, angry, scared, hurt, intoxicated, vengeful humans.

There's a ton of stuff to unpack here, but Pete, who do you think tells the AI how and when to kill? I mean the programmers might not be sleep deprived but they'd still be everything else on your list.

Should AI be able to launch nuclear weapons?

Never say never and what not, but having a human making the ultimate decision when and why to take life is a moral and ethical imperative that I (thankfully) don't see us moving away from any time in the conceivably near future. When we use drones is missions today, the spotter will confer with his superior, and if a decision to engage is considered, then one or more military lawyers are getting looped in, too, before the decision is made.

[Pete]

Humans, particularly in the heat of battle, are terrible decision makers. I want a dispassionate decision when it comes to taking life.

I am OK with humans assisting as long as they have clear parameters on when to override and the human is in a safe decision making position. For example, humans should decide nukes, as we currently understand that type of warfare. Dog fights, obscured tank battles, nighttime door-crashing infiltrations…these sorts of fast paced high stress situations have proven over and over again that humans are horrible at fast, stressful decision making.  For example, I once sat next to an army infantryman in a bar at an airport who told me he was re-upping so he could “kill more Muslims.” I bet an AI would do a better job of honoring human dignity and the value of life than him.

[Pete]

Of course he’s untrustworthy.  The point is that his business is a model for how others can disrupt the most corrupt industry on earth.

What model is that?

The same models at work in every industry, except for defense it seems.

- greater spend on high-risk R&D, relative to size, than the average defense contractor

- top-tier software and a new generation of STEM talent with fluency in digital technologies such as AI, quantum computing, and advanced microelectronics

- product-oriented business models that tend to be faster, cheaper, and more innovative

- a focus on commercially priced, scalable products and services

https://www.mckinsey.com/industries/aerospace-and-defense/our-insights/a-rising-wave-of-tech-disruptors-the-future-of-defense-innovation

[Pete]

How can anyone be excited about AI drones?

I’m not limiting this to small flying drones. I think AI should eventually be running all weaponized flying aircraft. Probably ground based patrol and attack vehicles as well.

[michigancat]

Humans, particularly in the heat of battle, are terrible decision makers. I want a dispassionate decision when it comes to taking life.

I am OK with humans assisting as long as they have clear parameters on when to override and the human is in a safe decision making position. For example, humans should decide nukes, as we currently understand that type of warfare. Dog fights, obscured tank battles, nighttime door-crashing infiltrations…these sorts of fast paced high stress situations have proven over again have horrible humans are at decision making.  For example, I once sat next to an army infantryman in a bar at an airport who told me he was re-upping so he could “kill more Muslims.” I bet an AI would do a better job of honoring human dignity and the value of life than him.

It could very likely be much more ruthless than that guy. You have to remember folks like Dick Cheney will be leading this type of AI development, not people concerned with human dignity and the value of Muslim lives.

[Pete]

Humans, particularly in the heat of battle, are terrible decision makers. I want a dispassionate decision when it comes to taking life.

I am OK with humans assisting as long as they have clear parameters on when to override and the human is in a safe decision making position. For example, humans should decide nukes, as we currently understand that type of warfare. Dog fights, obscured tank battles, nighttime door-crashing infiltrations…these sorts of fast paced high stress situations have proven over again have horrible humans are at decision making.  For example, I once sat next to an army infantryman in a bar at an airport who told me he was re-upping so he could “kill more Muslims.” I bet an AI would do a better job of honoring human dignity and the value of life than him.

It could very likely be much more ruthless than that guy. You have to remember folks like Dick Cheney will be leading this type of AI development, not people concerned with human dignity and the value of Muslim lives.

Then that will be fully auditable, both before and after deployment. That’s an engineering problem.

[Pete]

ALL military activity is ruthless. It’s the business of killing. I want it to be mathematical and dispassionate.

[Pete]

Also, here are several other benefits of AI:

- no sense of vengeance

- no bigotry

- no pleasure from dispensing pain and suffering

- no rape

- fully explainable actions, and predictable outcomes when “re-trained” to adjust behavior…no clinging to old ideas and customs.

[Pete]

In war, AI can be completely free from the motives that lead to war crimes.

I will gladly accept the risk of the agenda set in a boardroom in DC as the “motivator” for the AI warriors’ actions over the agenda set by disillusioned soldiers living in unimaginable hardship as the “motivator” for their actions.


What must a human mind have to go through in order to become a highly efficient killer? Is that mind to be trusted?

[michigancat]

In war, AI can be completely free from the motives that lead to war crimes.

I will gladly accept the risk of the agenda set in a boardroom in DC as the “motivator” for the AI warriors’ actions over the agenda set by disillusioned soldiers living in unimaginable hardship as the “motivator” for their actions.


What must a human mind have to go through in order to become a highly efficient killer? Is that mind to be trusted?

I honestly can't trust the minds in a DC boardroom more than a trained soldier. Who do you think has put disillusioned soldiers in the position they're in? I think it's more likely that removing humans from doing the killing would lead to even more ruthless killing, not less. Mainly because military leaders wouldn't need to put anyone on their side at risk.

It's just a bad/scary road to go down. War sucks without AI doing the killing, but at least we have a pretty good idea of what it's like, you know? The unknown repercussions of AI killing machines scare me way away more than the Muslim-killing infantryman.

[star seed 7]

What if ai decides that war is immoral and refuses to fight?

[steve dave]

I’ve seen Terminator Pete


Sent from my iPhone using Tapatalk

[chum1]

AI simulates human intelligence. If humans think killing is a good idea, AI will too.

[john "teach me how to" dougie]

The winners of all future wars will be the country that can shut off data flow, power grids, and disable satellites. War over.