Live data from Hacker News

DARPA, U.S. Air Force fly AI-controlled F-16

darpa.mil

201–210 of 357 posts

Re: DARPA, U.S. Air Force fly AI-controlled F-16

#201
post #80

All Stealth bombers are upgraded with Cyberdyne computers, becoming fully unmanned. Afterwards they fly with a perfect operational record. The Skynet funding bill is passed. The system goes on-line on August 4, 1997. Human decisions are removed from strategic defense. Skynet begins to learn at a geometric rate. It becomes self-aware at 2:14 a.m. Eastern time, August 29. In a panic, they try to pull the plug.

This is how AI will solve the paperclip problem.

And now I’m going to have nightmares about a super-smart, but just as annoying, Clippy.

Re: DARPA, U.S. Air Force fly AI-controlled F-16

#202

Earlier quoted context omitted.

That's straight up a lie. In all situations where autopilot was disengaged shortly before a crash, the autopilot is implicated in the crash. Both by Tesla's own standards and by NHTSA standards. And yet, people who really should know better repeat this as gospel.

Tesla is such a politically-charged brand at this point that none of these discussions are in good faith. It's all fanboys vs haters. The lidar vs cameras arguments are good example of this. This is a technical engineering question that could go either way - maybe cameras are good enough, maybe you need lidar, nobody on the internet is really in a position to judge. But Tesla did it with cameras, and everybody else i…

$TSLAQ was a thing long, long before Musk was openly donating to Republicans or running DOGE. It's always been a shit company making shitty cars that fall apart and kill their drivers and make their shitty CEO rich.

Re: DARPA, U.S. Air Force fly AI-controlled F-16

#203

> The kit utilizes a novel interface with the aircraft’s flight controls and mission systems, allowing a pilot to toggle between traditional human control and AI control with the flip of a switch. This ensures a safe, reliable environment for human-on-the-loop experimentation. How, exactly, does that ensure a safe reliable environment? Humans are pretty bad at suddenly taking over when an automatic system reaches its…

It doesn't, flying these is potentially very dangerous, there are many non-recoverable situations. But AFRL/NASA has been working on this for decades. The manoeuvres the higher level AI requests are actioned by highly assured dynamical systems solvers using model predictive control, like the autopilot, that maintain the aircraft in the safe envelope.

The unknowns in this system are (a) how agressive is the flight controller? (b) how clever is the autonomy?

Re: DARPA, U.S. Air Force fly AI-controlled F-16

#204

All Stealth bombers are upgraded with Cyberdyne computers, becoming fully unmanned. Afterwards they fly with a perfect operational record. The Skynet funding bill is passed. The system goes on-line on August 4, 1997. Human decisions are removed from strategic defense. Skynet begins to learn at a geometric rate. It becomes self-aware at 2:14 a.m. Eastern time, August 29. In a panic, they try to pull the plug.

1st place my head went.

Realistically though, whats worse - this thing going rogue or OpenClaud breaking out of its container and shredding a nation's infrastructure like a platoon of gremlins? Genuine question...

Re: DARPA, U.S. Air Force fly AI-controlled F-16

#205

> The kit utilizes a novel interface with the aircraft’s flight controls and mission systems, allowing a pilot to toggle between traditional human control and AI control with the flip of a switch. This ensures a safe, reliable environment for human-on-the-loop experimentation. How, exactly, does that ensure a safe reliable environment? Humans are pretty bad at suddenly taking over when an automatic system reaches its…

Probably a duplicate autopilot-like system that an AI controls. Flip a switch and controls revert back to the pilot. It may not even be a real switch, many planes disengage AP when you touch the controls. I doubt AI will be doing dogfighting anytime soon, this is likely just an AI following predetermined routes and maneuvers and making the choice on how to accomplish them. The great thing about computers is that they…

That is underestimating it. DARPA already showed dogfighting that can beat humans in simulation, and they may allow this do execute that.

However, dogfighting isn't very interesting, all combat occurs BVR and in the RF/EW domain.

Pilots are already spending most of their time thinking about epistemic problems: can I see my opponent, can they see me, how do I deceive, am I being deceived, how do I avoid being detected/tracked/targeted/engaged, while possibly trying to do that to everything else. AI is not yet very good at that.

Re: DARPA, U.S. Air Force fly AI-controlled F-16

#206
post #204

All Stealth bombers are upgraded with Cyberdyne computers, becoming fully unmanned. Afterwards they fly with a perfect operational record. The Skynet funding bill is passed. The system goes on-line on August 4, 1997. Human decisions are removed from strategic defense. Skynet begins to learn at a geometric rate. It becomes self-aware at 2:14 a.m. Eastern time, August 29. In a panic, they try to pull the plug.

1st place my head went. Realistically though, whats worse - this thing going rogue or OpenClaud breaking out of its container and shredding a nation's infrastructure like a platoon of gremlins? Genuine question...

One kills people, one kills 1s and 0s.

Sure life will suck for a while without the 1s and 0s that make our life better, but at least it will be life.

Re: DARPA, U.S. Air Force fly AI-controlled F-16

#207

The article smells of AI psychosis. They are already "orchestrating". In the first Iraq war, they went in and won. In the second one, it took years to sort of occupy the country. Now they struggle against Iran and "orchestrate" F-16s that can be shot down instantly by AA weapons. Since COVID the world has collectively lost its mind.

Unless those very F-16s shred their way through the AA weapons first, that is. Or follow up once the stealthier, more survivable platforms pave the way. Iran had a solid AA system - and the level of impunity with which US operated with over Iran should give every country that wants to rely on its AA a pause. US SEAD/DEAD capabilities are fucking scary. Of all the things that went wrong with Iran: USAF's capability to…

I would argue that their dominance in SEAD made them overconfident about suppression of OWD and IRBM launch. They failed to address the key risk to the strategic mission because of an obsession with air dominance.

Iran and Russia have likewise been overconfident about GBAD, which is effective against crewed fighters but useless against massed drone attacks.

If the US LUCAS (Shaheed clone) had better ISR and semi-autonomous engagement it could have achieved much better suppression. But that is not a pilot in the loop option and USAF is deeply uncomfortable with that.

Re: DARPA, U.S. Air Force fly AI-controlled F-16

#208
post #182

Earlier quoted context omitted.

What they really got wrong, is that there aren't even enough missiles, ammunition or fuel production to defeat a country that has been sanctioned for decades and these bottlenecks would make it trivial for humanity to prevail over AI, it'd be like Futurama "for unknown reasons, a kill limit of 999,999 is hardcoded into killbots...".

Except Skynet has nukes. But what isn't addressed is that radiation is terrible for semiconductors, both in operation and during production. Basically a nuclear war would end GPU production, GPUs have a 5% annual failure rate, higher with more background radiation, so it would get dumber instead of smarter, until it stopped working or partnered with biological compute (Daleks or The Matrix) or a human faction.

The idea that Skynet would use nukes seems very quaint to me.

Subversion and sabotage of communication, food production and transport will kill human civilization far more effectively.

Re: DARPA, U.S. Air Force fly AI-controlled F-16

#209
post #142

Earlier quoted context omitted.

Drones really changed the equation more than Covid. Autonomous F-16's already can win against all human pilots. This would have been big 20 years ago. Now the cheap drones make it not the biggest deal anymore.

Which cheap drones? Specifically which systems are currently out-competing F-16s in air to air combat in either trials or combat?

You don't need F-16s at all if you have cheap drones. Why send an expensive plane with a pilot to bomb some country if you can send a drone that costs as little as a basic SUV instead?

Re: DARPA, U.S. Air Force fly AI-controlled F-16

#210

Earlier quoted context omitted.

Still, the F16 is 16 hours of maintenance per hour of flight or something close to that. Which probably isn’t terrible compared to other aircraft but doesn’t seem sensible for something without aircrew

To make it safe for the pilot plus anyone in the flight path. Could the hours of maintenance be dropped significantly in wartime: if you don't care about catastrophic failure modes over a foreign country?

I assume there are shortcuts like extending the service life of particular consumables and whatnot.

Even something like a Gripen, designed to be run from makeshift airfields needs 10 hours of maintenance per hour of flight. Maybe it’s just hard to build supersonic, combat capable airframes as was suggested above.

Post reply on HN