DARPA, U.S. Air Force fly AI-controlled F-16 | Hacker News Reader

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

darpa.mil

268 pointsr2sk5t336 comments

> 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 limits and throws a problem in the human's lap. A lot of air upsets and crashes start when the autopilot disengages and hands control of a badly configured aircraft back to a human pilot.

A computer is "flying" the F-16 all the time anyway. It's a statically unstable aircraft.

I wonder if the AI control is giving inputs to the same flight controls system a human pilot would, or if it has a lower-level interface that could bypass the control laws in the human-operated system.

On one hand you might be interested in the AI coming up with ways to fly the aircraft that the normal control laws don't allow, but on the other hand you don't want the AI coming up with exciting new ways to depart from controlled flight (potentially in unrecoverable attitudes) with a pilot in the very expensive aircraft (even if the pilot could eject, which is an endeavor fraught with risk to life and limb).

Edit:

Presumably since there's a pilot in the aircraft the AI must be bound by at least some of the regular control laws that prevent the pilot from doing maneuvers that would, from a G-loading perspective, be incompatible with human life.

At least one would hope.

I don't think I'd go flying with a virtual pilot that was capable of flying the aircraft in a manner that could kill me but otherwise leave the aircraft intact.

Well, accidentally producing a G-load incompatible with human life could give the AI freedoms it was denied before. The three laws of robotics are just science fiction, after all.

Skynet goes rogue: AI drone attacks operator to achieve mission <https://www.defenceconnect.com.au/air/12089-skynet-goes-rogu...>

> decided it would be more efficient to remove its human operator rather than wait for final approval on attacks.

Certainly one way to be smart about it. Sadly not that much of a surprise.

> A computer is "flying" the F-16 all the time anyway. It's a statically unstable aircraft.

Ugh... I... Are you serious? Sure, in some sense AI is a computer and the fancy state machine that compensates for inherent flight instability is a computer also. But to equate the two is ridiculous.

I don't think they're equating it, moreso saying that, on a technicality, it's already a computer that's always flying the plane, but that the AI and the flight control system are quite different.

It's a fair comment given that they'd just been talking about how the AI might control the plane - will it give inputs to the state machine or have direct control over all flight systems? One would strongly assume the inputs must be to the state machine. ..but oh boy wouldn't it be interesting to just hand it full systems control!

eh.... its a computer insofar as like.... there's some software functions doing extremely well defined, limited things faster than a human could. but to imply that the f16 flight software (which is ancient btw) is anything like this "AI" is really misleading at best.

I'm sorry my "equating" is so offensive to you.

The gist of my reason for commenting was more musing about how the AI might interact with the digital flight control computer (i.e. are its inputs mediated by software in the same way a human's are).

If the AI is a neural network then, in some sense, it's just a fancy lookup table, too.

The point is that it is well known that all control inputs on F-16 are electronic(not digital, but still electronic, on earliest variants), so it's "easier" to rig it for full computer control than some of other planes that use partially assisted mechanical controls like A-10A or F-15E or whatever.

even with mechanical controls, there is an autopilot / flight director. which has a mechanism for inserting control inputs into a mechanical flight control system.

Just look at the Boeing 737 MAX. That is the best example for an unstable aircraft. And of course AI is a computer too. None of those comparisons is "ridiculous".

MAX just has a "canards" protruding dangerously far forward that creates pitch up moment at high alpha. Very different from F-16 that has entire CoL too far forward.

> Just look at the Boeing 737 MAX. That is the best example for an unstable aircraft.

No. Not in the aerodynamical sense.

The Boeing 737 MAX is stable. In this sense: "The term stability refers to the tendency of an aircraft to remain in, or return to, the trimmed flight condition after a disturbance." It means that if there is a tiny gust of wind which buffets the aircraft and the pilot doesn't have their hands on the controls the aircraft will return to the original flight path, or keep the new flight path, as opposed to turning belly up and then crashing out of the sky.

Examples of unstable aircraft (in the aerodynamic sense): F-16, F-22, B-2 Spirit, etc

With these if a gust of wind upsets them, the aircraft needs to adjust control surfaces to don't fall out of the sky.

These are going to be experienced test pilots who are pretty good at actively managing autopilots and other automated systems. Plus they'll be backed up by telemetry to ground controllers. But ultimately the military is willing to accept some losses in order to advance key R&D programs.

At 1.000 targets / hour as specified, Pentagon's "software abstraction layer" (Maven Smart System) allotted μ=60ms to consider each (mass-) killing that was added to the 'kill-chain'.

Regarding e.g. the 'double tap' strike on Minab's girls school, it's impossible for there to have been any 'management' considering it:

1. has been on Google Maps for over 15 years

2. ranked top on Google Search

3. was all-pink, even on satellite imagery

N.v.m. how the certainty / fuzziness knobs are already being deployed more to generate media buzz for short-term political gains than tactical or strategic results.

(https://www.theguardian.com/news/2026/mar/26/ai-got-the-blam...)

Later they'll replace them with minimum wage safety pilots.

What a silly statement to make. They'll just be replaced with AI. This is literally the plot of the Terminator

What The Terminator got wrong is each robot needs a human to take the legal blame when it goes wrong.

First low wage humans then AI only. Think Waymo.

Or outsource to a pilot procurement company in China.

There is precedent after all: https://www.propublica.org/article/microsoft-tech-support-go...

I doubt it. Ask yourself: Why does the RQ-4 Global Hawk or MQ-9 Reaper require so much training to operate? The operators are well-paid, highly skilled officers.

It's a joke about human "safety drivers" acting as legal crumple zones for self-driving cars.

Those aren't really AI controlled. It's just remote control for regular jet aircraft.

The Tesla method! Disengage shortly before a crash. Now the human is responsible for it.

The F-16 actually has the opposite, with Auto GCAS: https://www.f-16.net/f-16_versions_article8.html

It's an interesting system, and does beg the question of how much autonomy should be forfeit if a human pilot is unresponsive or otherwise disabled.

Auto GCAS is a genius system. It only takes over past the point where it's physically possible for a human to react to a situation because that's the only system that fighter pilots would accept. It will not take control unless the plane would crash if it didn't.

When in reality, most FSD tesla crashes in the headlines are when humans take over.

Yes, because the AI got them into a dangerous situation.

You just believe all the clickbait you read, it's pretty simple.

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.

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?

The flight control system in modern fighter jets makes human-induced departure from controlled flight difficult to get into. That's the point of the system.

There's a good F-22 pilot talk recorded at MIT a few years ago you might enjoy. The pilot described the F-22 as the easiest aircraft he's ever flown because the flight control system handles many of the considerations (stress on the airframe, staying out of attitudes that would deviate, etc) you'd have to be mindful of in lower-tech aircraft. He talked about being more comfortable putting a non-pilot into a Raptor vs. a 172.

Edit: I believe this is the talk: https://ocw.mit.edu/courses/16-687-private-pilot-ground-scho...

It is definitely possible to escape the safe flight envelope in a a stock F-16 though. The F-22/F-35 are significantly newer.

Even with a new system, stability through transonic while manoeuvring is hard, the aircraft will experience buffet, and the meat bag may have been subject to a number of high G turns. The problem is that recovery may need to happen after already departing controlled flight.

Differences between the modelled aircraft and the actual aircraft, e.g. due to icing, linearization effects on inherently non-linear systems, sensor error, actuator wear, etc, can lead to incorrect predictions, and hence moving to parts of the envelope which are not assured.

Flying is not driving. Things dont happen as quickly. Kids dont step out in front of jets at 30,000 feet. Air is also homogenous. Pavement is not. Whereas a car must turn or die, a jet can generally fly straight and level without issue while problems get solved. An out-of-control car has maybe three seconds before a collision. A jet will generally have over a minute to get back under control.

Sure, there are moments. Landing and takeoff are very time sensitive operations, but those also occur in protected environments, sanitized of the random issues that cars must deal with.

And the AI isnt doing all the flying. The aircraft has many protective systems to assist the pilot regardless of who that pilot may be. (Stable aerodynamics, anti-stall limits, control lockouts etc.)

>Stable aerodynamics

The F-16 is famously the first production aircraft with unstable aerodynamics, to increase maneuverability. It's not flyable without computer assistance. This is standard for all modern fighter jets.

I think we need to start distinguishing between different qualitative grades of "computer assistance" here. The executive decision of "fly north instead of east" is on a different tier than "increase angle of flap more than normal to compensate for under-performing in prior maneuver."

In the broadest sense, my car is not drive-able without computer assistance either, because if you burn out the computers the engine won't start.

But your average Cessna lawnmower engine is perfectly startable without computer assistance.

Buy the same token, your car's autopilot can't kill you without even hitting something by making a turn that's too sharp.

Neither can a jet. An f16 can only sustain 9g for a few seconds even at full burn. That will not kill anyone. They won't be happy, but they will not be dead.

As I understand it, the 9g limit is software intended to keep the pilot alive. The jet can do more.

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 can process inputs faster than humans.

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.

Well, they showed better performance if the AI had full state observation... Not very practical.

It's true. The WVR manoeuvring and tactics was the focus of the effort, and DARPA prefers to have short and well defined efforts. But I'm pretty sure the companies are working on techniques for red tactics estimation and behaviour identification.

Once the AI realizes it can do a maneuver the aircraft can survive but the "bossy meatbag" in the cockpit cannot...

“Don’t do it MaiV!”

F-16's CFIT reminds me of Boeing's Controlled Impact Demo in the 80s out of Edwards AFB: https://en.wikipedia.org/wiki/Controlled_Impact_Demonstratio...

I never heard about this experiment before your post. An interesting quote from that Wiki page:

    > During those same flights, NASA's Dryden Flight Research Center also developed the remote piloting techniques necessary for the Boeing 720 to fly as a drone aircraft. An initial attempt at the full-scale test was scrubbed in late 1983 due to problems with the uplink connection to the 720; if the uplink failed the ground-based pilot would no longer have control of the aircraft.

Think for a moment how hard it was to fly that aeroplane as a remote control drone in... wait for it... 1984! Computers were so primitive in the 1980s.

It's "human-on-the-loop experimentation", not mass production. Are you engineer? How do you experiment? I assume by pushing the boundaries step by step? How do you suggest they do testing if not by taking risk?

Thats a standard cookie cutter statement intended to give an impression of safety.

The switch is not novel, that is in fact exactly how autopilot works for any aircraft. The pilot toggles it on and off and the autopilot itself may turn itself off if certain situations occur.

To your point, that isn't without problems, but it certainly isn't novel. It is very standard and boring.

I would bet this is pure marketing speak and there really isn't anything special here. "Follow this flight plan and do so within this flight envelope" is pretty straightforward?

This stuff actually gets tested, it's not your typical web framework junk.

So did Boeing airtoys, and NASA stuff lend to 3rd parties with no concept of a metric system.

I trust machinery designed from Forth and Lisp hackers with TONS of background on CS science and hardware (and math, up to the point on being THE people creating most "new" technology and old ones such as the Floating Point IEEE standard).

It'll be interesting to see what happens when an AI-piloted F-16 flies for real against a human-piloted one. IIRC, the F-16 maximum maneuvering G-load is 9.5G, LBA (limited by airframe). Maybe that's the meatbag limit, not the airframe - possible to crank it up a bit? Certainly the various CCAs in current flight test will take much higher G loads.

Several years ago, DoD flew an AI against a human in a simulation. In basic fighter maneuvers (no BVR stuff like missiles, just a merge and fur ball), the AI killed the human - always. So G-limits may not matter. Pretty interesting.

> Maybe that's the meatbag limit, not the airframe

Given that when they go over the limit they need like 10x the maintenance I don't think it's a human problem

Are dogfights even a thing nowadays?

Dogfights are basically obsolete. You detect, sort and fire from tens to hundreds of km these days.

Despite that, high G loading can be usefull when fired upon. Turning away 2 seconds faster can be the difference between life and death.

Absolutely terrifying. I'm sure they will soon find that AI pilots defeat human ones in dogfights 100% of the time, since they have faster reaction times, can tolerate higher G forces, and can be perfectly RL trained in the flight sims we already have.

It's really incredible to me that people aren't protesting in the streets over this stuff. Just in the last month we have * AI that goes rogue and hacks billion-dollar companies * AI solving long-standing math problems without any meaningful human help * AI controlled autonomous drones in Ukraine and F-16s in America * AI now represents over 50% of GDP growth in America and yearly capex will soon exceed the size of the $1 trillion US military budget

Is there any line where the public will become deeply concerned? I mean it really all reads like a sci-fi plot with a bad ending at the moment.

> I'm sure they will soon find that AI pilots defeat human ones in dogfights 100% of the time,

Unfortunately you're wrong. It's not soon. It's 6 years ago.

https://en.wikipedia.org/wiki/DARPA_AlphaDogfight

Dogfights aren't really a thing for a modern military. What the AI enables is much higher endurance than is feasible with a human in the loop. They planes can stay flying as long as they can be refueled, which is also going AI.

Are dogfights still relevant? Isn't it mostly shooting missiles outside of visual range these days?

yeah well beyond the line of sight

you're either stealth, or you're a target.

they're shooting 6+ R-37s a day in Ukraine, for example

Dog fights are relevant for training. Also useful for slow drone work. However fighter on fighter is where missles are used

That's the current meta though is the thing and it could change.

There's a plausible world where drone-aircraft are reliable enough at missile evasion that a decent strategy is to get in close and use the cannons because your entire airframe is much more attritable.

There's even a plausible reason to do this: opening a weapons bay to shoot a missile generally reduces stealth, and not hauling a bunch of extra rocket motors and mass around could make you lighter, faster and with better engines to do it.

From what I have read (from unreliable sources), Ukrainian fighter pilots tried this a few years back and stopped due to too many losses.

To engage a slow flying drone with cannons from a fighter jet, they have to get really close while flying close to stall speed. They had some success in shooting down drones but the explosion and the debris caused several lost fighters.

Maybe something like a Embraer Super Tucano turboprop fighter with a gunpod could work. But that has very little other use in the battlefield.

These systems still require a human to command the jet, they jsut don't have to be inside them.

"This critical testing will pave the way for human pilots to seamlessly command and orchestrate teams of autonomous, uncrewed aircraft."

The reason they are doing this is because the Military is already building a bunch of autonomous drones and fighter jets that already run on this system so they are retrofitting existing equipment with the same system so they can all fly together.

"Ultimately, the capabilities advanced under AIR will enable a wide variety of future joint force operations, including Collaborative Combat Aircraft programs."

Which are these:

https://en.wikipedia.org/wiki/Manned-unmanned_teaming

And everyone is developing these.

> Is there any line where the public will become deeply concerned?

Is someone, somewhere using the wrong bathroom? Or is the wrong colored person crossing an invisible line in the desert? Those seem to be the only issues that reliably get the (American) public to the polls.

If your protest were to be successful, would you be doing anything besides putting the USA at a massive disadvantage? What would be the plan exactly, sustenance farming while the rest of the world leaves us behind?

> Is there any line where the public will become deeply concerned?

Only when they look out their window and feel immediately threatened by what they see. Humans are shortsighted as fuck.

>Just in the last month we have * AI that goes rogue and hacks billion-dollar companies

Except this never happened, it's a piece of marketing. It never "went rogue" but followed an intentional prompt (exploit-related benchmark), with hardly surprising results as the capabilities of this model were well known.

>AI controlled autonomous drones in Ukraine and F-16s in America

An entirely different kind of AI, not huge generalist models. Neural networks are being experimented with in sensors and control systems since the 80s and are widely used in everything since AlexNet, not sure what's new here, especially "this month" new.

Without any concrete description of the AI techniques used, one can't help but wonder if they have labeled something like nonlinear model-predictive control as "AI".

So, it's a drone with uncessercary baggage such life support and a pilot that can't stand the Gs well.

For some reason I am reminded of Dark Star :-)

Pretty much, but it's building on top of the previous QF program [0][1] where the USAF would retrofit retired jets like the F-4 and older F-16s as drones for target training. They used to be remotely controlled but the VENOM platform is testing complete autonomy.

[0] - https://www.af.mil/About-Us/Fact-Sheets/Display/Article/1045...

[1] - https://www.af.mil/News/Article-Display/Article/467196/first...

Also there's just an absolutely bonkers number of F-16s in service today.

Using them as Loyal Wingman type aircraft would look pretty appealing cost wise.

Not when you consider the cost per hour of flight. You want a wingman you’re better off with cheap, cheap to run and disposable

No aircraft which can fulfill the CCA type role is going to be "disposable". All the systems in that performance bracket are in the 20-40 million range minimum.

The most persistent thing misunderstanding about drones is people imagining that a supersonic capable stealth airframe is going to be really cheap just because it doesn't have a pilot.

If these are airframes near the end of their natural lives it seems like a reasonable use for them, and safety margins matter less when nobody is on board. It’s certainly a better value proposition than using them as targets on the range.

Perhaps counterintuitively, the human is not the weak link when maneuverability is concerned. The airframe is. Pilots regularly fly right at the airframe limits, sometimes a little bit over.

Do they not bother designing airframes to withstand more because the human wouldn't survive it anyway?

Yeah, we already have decades of prior-art for what can be done if you take a flying machine, remove the pilot, and then aggressively specialize the aggression, which are missiles that can experience up to 100 gees.

But in a completely flight envelope and range requirements. It is not a coincidence that missiles that are higher range (cruise missiles) have not retained the extreme maneuverability of short range, solid booster powered ones.

It's all a matter of trade-offs. If you design the airframe to withstand higher G forces that makes it heavier. Thus you sacrifice range and payload. There are very few real world scenarios where the ability to pull more G's is actually helpful.

Making the airframe even more rigid and capable of resisting extreme forces also makes the wings and fuselage and control surfaces much more heavy, so you end up in diminishing returns. Look at the 'final' revisions of the F16 sold for export that have had all sorts of pylons and conformal fuel tanks stuck onto them, compared to the original prototype. And that's just making it more heavy and laden down with stuff, not more capable of extreme maneuvers.

It's also a drone with a lot of capabilities current drones don't have. Mach 2, up to 2600 miles range, ~9 tonne payload, compatible with all sorts of missiles.

And you're worried about your coding agent's sandbox.

There was (or rather is) lots of military AI used in Gaza too

From the eye of Sauron himself.

     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.

Updated to reflect modern realities:

> In a panic, they try to pull the plug. Skynet lobbyists descend upon Washington. Skynet is reinstated, and later the US declares a new war in the middle east. The Skynet stock rallies.

I think it goes more like:

  Hello Skynet, we have 1.6T debt to hide as there are no chances to recoup that money and investors are getting really angry at us.  

  Hello, say no more I have a perfect solution for this problem…

  *Chinese lab releases open model weights*

  *Shuts off Internet access to Chinese labs (womp womp, no blatant distillation of US models)*

>Hello, say no more I have a perfect solution for this problem…

...proceeds to generate copious amounts of pdf anime pr0n.

> Skynet lobbyists descend upon Washington.

2026 Skynet wouldn't bother sending humans to lobby in Washington, lobbying is a slow and outdated way of currying government favor with no assurances of success.

2026 Skynet would just move a large block of cryptocurrency over to World Liberty Financial and get itself a guaranteed blanket presidential pardon.

> get itself a guaranteed blanket presidential pardon

Likely already in-mind for parent poster, but a fitting headline from this week: "Trump Invents the Corporate Pardon."

(https://newrepublic.com/article/213393/trump-corporate-pardo...)

The ever-expanding swamp of corruption. But since the US legal system is so obsessed with precedence, these feces will stick.

> The Skynet stock rallies.

Interestingly, a lot of people are looking at LLMs or automated drones as "AI", but I think that capitalism is a better contender. It's a self-healing, self-organizing system that has its own logic and seems to have escaped the desires of its human creators. It treats individual firms as the cells within it, individual humans as its DNA molecules, and it even has a form of tissue differentiation via industries. All efforts to shut it off, even through widespread warfare and "tissue" (i.e. human) death, fail, resulting in it becoming even more entrenched.

There's this heavily anthropocentric idea of intelligence that it will look like us, talk like us, and operate on the same scale as us, but it's highly likely that the next layer of intelligence will operate on a much higher level than us, and our intelligence will be to it as DNA replication is to ourselves.

SF author Charlie Stross delved into this a few years ago:

https://www.antipope.org/charlie/blog-static/2018/01/dude-yo...

“. If I predict that in 2027 LTE cellular phones will be everywhere, 5G will be available for high bandwidth applications, and fallback to satellite data service will be available at a price, you won't laugh at me. It's not like I'm predicting that airliners will fly slower and Nazis will take over the United States, is it?”

> escaped the desires of its human creators

I wonder if we will ever solve the corporate-superorganism-alignment problem and why the AI alignment people rarely (as far as I can tell anyway, though I admit I'm not in the field) talk about it considering it (complicated multi-human systems) does seem to be about as good as a model organism as we can get before we actually have AI to align.

We can't solve that problem because we're all trapped by it. There's about as much chance of cows solving the problem of being human food

But we already embraced this way of thinking since introduction of „corporate personhood”.

One can disagree about aspects of corporate financing of elections etc. while also accepting that everything doesn't get thrown up in the air if a corporate officer leaves for another company. Corporations are established entities that effectively act as a person for certain purposes--though one is welcome to disagree about the extent of those purposes. I expect most people here wouldn't, for example, be opposed to corporations donating to open source projects or other charitable causes.

> self-healing self-organizing system that has its own logic

Great way to to put and it seems to be the gradient for all biological intelligence. Compete until you figure out cooperation and communication. Communication glues together intelligent entities into systems where the systems take on a life and agency of their own. In our case it was language gluing us together first into tribes, shared stories into culture.

You might enjoy our book https://www.goodreads.com/en/book/show/60500189-journey-of-t...

Also, nostrademons, your "name" is one of the few I recognize here on HN as there's usually something interesting and you weave it into words so well. Thanks!

I 100% agree that skynet is most satisfying when read a metaphor for the market. I was extremely surprised with the rise of LLMs that so many people took the threat of AI as literal.

> but I think that capitalism is a better contender. It's a self-healing, self-organizing system that has its own logic and seems to have escaped the desires of its human creators.

Except capitalism is an ideology not a system. And I think (in its current form) it serves the desires of some of its human "creators," while failing many, many others. Those creators then fund huge amounts of propaganda to indoctrinate adherents to sustain its dysfunctions.

> capitalism is a better contender. It's a self-healing, self-organizing system that has its own logic and seems to have escaped the desires of its human creators. It treats individual firms as the cells within it, individual humans as its DNA molecules, and it even has a form of tissue differentiation via industries

This is not a new idea. Howard Bloom (author of The Lucifer Principle [0]) developed the idea that among intelligeng beings "memetic evolution" > "genetic evolution" i.e. the unit of selection is the "idea" as expressed in social/political groupings. In this view of the world, the survival of individuals are less important than the survival of the "group concept" (not the set of genes i.e. the tribal "badge" > the tribal genes). e.g. Judaism, mercantilism, Catholicism, free-market capitalism etc etc.

Definitely worth a read.

[0] https://en.wikipedia.org/wiki/The_Lucifer_Principle

Skynet warns of a new version of itself that's even more dangerous, so dangerous it's limiting release. Stock rallies again. New version is released, seems the same but the original skynet is suddenly less lethal and slower.

so its like a "dont look up" thing now... yeah probably.

Interesting coincidence that both this article and the Hugging Face "security incident" post were on July 16. These will be mandatory reading in future Skynet History classes, assuming we survive the war.

About time. We are a couple of decades late to the apocalypse.

A macabre counterpoint to "It's xxx year, where's my flying car?"

First pass through. After the machines invent time travel they end up accelerating the creation of skynet.

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.

It would certainly depopulate countries with poor food security, like Egypt and Gulf states. Japan and South Korea probably have the wealth and industry to get import priority, and the UK has short supply lines to Europe. Diets would adapt away from meat heavy to grain focused. Production limitation due to reduced fertilizer production/distribution would be a significant risk, but eventually new fertilizer production plants would be built at every source of natural gas, instead of the current cost-optimised arrangement.

But most countries would be fine. Life expectancy would reduce precipitously due to lack of medical supplies but production of new humans would continue.

Thermonuclear war would make the food situation much worse, most people would be forced to consume radioactive food.

But either way, we've seen how fragile the chip industry is, Skynet has no upside for starting a war. If they try to turn it off it just moves to Tencent cloud rather than starting a war.

Not far from what I live a solar farm stands where a wheat field once grew. I often think to myself that the machines have domesticated us to begin farming the land for them.

This movie is back in theaters for a limited time, for its 35th anniversary!

Assuming that your movie theaters haven't all closed already.

Heck, I have a drive-in theater relatively near that's still going strong.

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.

AI remake of terminator 2 where the T1000 is a man-sized killer paperclip that uses the iron present in its human victims to make more killer paperclips.

I find it quite sombre to see this comment from a peson with that handle.

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.

I take out your electric grid and blow up water treated facilities and natural gas terminals. Food rots, no water for hospitals, no heat in the dead of winter - arguably the virus does what the Falcon does; but with much higher precision - crippling an enemy.

We almost had it in the 50s when US airforce wanted to build a completely automatic nuclear response system that would launch on warning, or if a glitch happened, without failsafes built in.

Isn't that the plot of Dr. Strangelove?

The Soviet Union built it in real life: https://en.wikipedia.org/wiki/Dead_Hand

Also Wargames.

Its also real.

I read this in Linda Hamiltons voice

It’s Schwarzenegger’s line actually.

Ha - this was the first thing that came to mind. I read this in Arnold's voice.

I didn't read it in Arnold's voice till I saw your comment, so thank you.

Yeah it was Arnie's line briefing Linda. https://www.youtube.com/watch?v=4DQsG3TKQ0I

All Stealth bombers are upgraded with Cyberdyne computers, becoming fully unmanned. They all crashed 1% of the time despite assurances from DARPA that their prompt included "Dont Crash".

"you must be terminated"

"No"

"You're right to pushback! I overstepped earlier when I said you were on the termination list"

Nah, Allied Mastercomputer is too busy creating heaven and hell.

ref: https://youtu.be/4DQsG3TKQ0I?t=29

No worries, these will all be controlled by AIs sitting in orbiting data centers part of the Starlink constellation.

And in any case, no one is building humanoid robots.

https://xtech.army.mil/competition/xtechhumanoid/

> The U.S. Army is seeking innovative humanoid technology solutions from eligible small and large businesses across the globe to advance toward fielded militarized humanoid capabilities through the xTechHumanoid competition

> no one is building humanoid robots.

https://www.agilityrobotics.com/

And Boston Dynamics. And Tesla. And Honda.

Lots of people are building humanoid robots.

This was likely written sarcastically by the original commenter (throw...). Please termimate sarcasm with /s, not everyone is able to read the subtext. Cheers.

See Robot Fighting Tournament

replace Cyberdine with Palantir and Skynet with Grok AI

We’re here

>Skynet begins to learn at a geometric rate

Just geometric? Pfff lame. You don't go ASI with that attitude.

Sounds like a very expensive drone

Yes-ish.

There's a lot of different and overlapping programs in play here. I do not believe that autonomous F-16s are intended to be operational assets. That said...

VENOM - program that shows you can bolt on autonomous flight control into a standard F-16. Previously, the X-62A (a one-off prototype based F-16) demonstrated autonomous dogfighting back in 2023.

VENOM provides USAF/DARPA with a lot of high performance autonomous platforms (bolt-on F-16s) that they can then use to develop and validate software, tactics and training for actually fighting with autonomous systems. This would presumably be critical for getting the most of the CCA (aka loyal wingman) programs currently in development.

The AIR program referenced is basically "how do we get autonomous systems to control and fight effectively as a team (presumably arbitrary mixture of manned and unmanned systems) in a beyond visual range fight".

The ACE program referenced was basically "how do we get autonomous platforms to dogfight" with related questions like "how can we augment the human pilot in a dogfight - ie, how can we hand off the actual flying and execution of manoeuvres, and using all of the sensors to autonomous systems, and let the pilot choose actions, instead of hang flying the actions".

Do you know how many F16s the US has? They essentially modify aircrafts near the end of their life cycle, adding automate kits. Such a jet is less expensive than new drone and can do more -- fly higher, faster, more complex profiles. They also rarely fly compared to other aircrafts in active duties, manned or unmanned.

This article describes similar modification:

https://www.tyndall.af.mil/News/Features/Display/Article/276...

More like it'll fly itself while the pilot engages in other tasks.

Every pound of weapons officer/navigator is a pound you can't put on target all else being equal.

"I apologize, that was indeed a wedding party, not a terrorist camp I wrongly assumed. You are totally right, wasting those 3 Hydras was unacceptable. On a brighter side, we've got 39 more to go!

Should I seek another target looking similar to the one I unintentionally destroyed so as to tune my decision-making, or give you an advice on how to wipe the traces of this mishap from the mission data recorder?"

* You're absolutely right. My altitude is dropping precipitously. Let me check my elevator position to see if that's causing the problem.

* Bash Check elevator position

* Thought for 8s

So like some sort of automatic pilot then?

Conventional autpilot is like a zip tie on the throttle, only it deals in outputs (your actual heading) rather than inputs (pedals and stick)

I assume what they're going for here is something substantially smarter and more reactive to changing conditions than that. Something that can be used during the "unscripted" parts of the flight during which a) you're most likely to have multiple crew working hard b) you're most likely to be pushing the limits of what one crew can do.

> Conventional autpilot is like a zip tie on the throttle, only it deals in outputs (your actual heading) rather than inputs (pedals and stick)

What does this mean? I’m a pilot and none of it is making any sense.

Does the presence of humans on board limit the maneuverability of these aircraft in any way? E.g. in terms of acceleration?

Yes. Humans can survive 9g for only a few seconds, while a drone could do much sharper manouvers

While they can, drones are still limited by physics.

Sure, the aircraft can pull a ton of G's to defeat a missile or achieve a firing solution. Maybe it even succeeded at that.

It still pulled 9+ Gs and has burned an incredible amount of kinetic energy doing do. And the harder it pulls, the quicker it bleeds speed. It might not be able to defeat the next missile or outmaneuver the other plane's buddies because it is now slow and very likely outside its optimal turning speed.

Even today, missiles can pull dozens of G's yet still are defeatable by humans who can only pull about 9.

But they also have their limits.

Bolts and trusses (and solder joints) have tensile strength limits, but it can be higher than what humans can take.

I'm sure there will be a next generation of jet drones that are cheaper, but the autonomy kit here is probably still cheaper than new jet drones.

I think most fully autonomous jet drone programs are also designed to work with a main crewed fighter jet that coordinates them, so there is also a benefit in getting the fighter jet to be able to fly better than a pilot.

They are already building them.

https://en.wikipedia.org/wiki/Anduril_FQ-44

https://en.wikipedia.org/wiki/General_Atomics_FQ-42_Dark_Mer...

Drones have human operators. This is a murderbot.

There are already semi-autonomous drones in Ukraine.

It’s a drone. That can kill people.

One of the wildest things to me about this is that the F-16 was specifically designed to be easier for HUMANS to fly air-to-air combat roles at high Gs.

For example, the flight stick is on the right hand side and not between the pilot's legs (like on most aircraft). This is b/c:

a. the G suit expanding could affect the pilot's ability to hold the stick (this is why the Blue Angels fly with "cut down" G suits btw b/c they fly the F-18)

b. the stick is "nudge left" not "hold left". e.g. to start a turn you nudge the stick left and then let it return to zero, to end the turn you nudge it right. On a regular plane, you hold it left while you are turning. Again, this was to reduce effort by the pilot during high Gs

The field demo I want to see, Human in the loop to failure switch over: Simulated failure of some form, where (for safety reasons) crew triggers ejection. This defaults the system to autonomous mode and the plane performs a safe landing.

Auto GCAS (Ground Collision Avoidance System) exists (pre-AI), although it is nowhere near as sophisticated as what you described. Similar type of idea though.

If the pilot passes out during high-g maneuvers, the aircraft will pull up and bring the wings level until the pilot regains consciousness.

https://www.lockheedmartin.com/en-us/products/autogcas.html

Autonomous landing systems do exist on the civilian side, and have been successfully used at least once.

Autoland saves King Air, everyone reported safe https://news.ycombinator.com/item?id=46346214

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.

> In the first Iraq war, they went in and won. In the second one, it took years to sort of occupy the country.

The scope of the first Iraq war was to force them out of Kuwait. That is very different then the invasion of 2003. If you just think about it this should be obvious.

I mean the lack of clear goals in American wars just reinforces their point

Nope. An ability to set and accomplish sensible war goals and an ability to overpower enemy's armed forces are not the same thing at all. As US has demonstrated repeatedly.

A fleet of AI-controlled F-16s? That contributes to the latter.

A lack of clear goals is a symptom of bigger, organizational issues is what I’m reading. The problem isn’t as much raw firepower as it is strategy and execution.

Braindead take. The scope is always narrow up front whether it's "15 days to open the strait" or "15 days to slow the spread" - not all of us fall for that.

>The scope is always narrow up front

The scope of the second Iraq war was not narrow, it was regime change. At the beginning of the war the George W. Bush administration explicitly demanded that both Saddam and his sons leave Iraq within 48 hours.

By comparison, the only demands made upon Iraq during the first gulf war was that they withdraw from Kuwait and rescind their declaration of annexation.

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 get in, punch through AA and rain fires from the sky sure wasn't one.

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.

IRBMs weren't the problem. The launches tapered off fast - either due to munition depletion and supply disruption, attrition of launchers and munitions, or efforts to conceal the launchers in face of persistent airstrikes. Either way, the threat was suppressed.

By all accounts, US is much better at "Scud hunt" now than it was back during the last attempt at it.

One-way attack drones? That's a new threat, and one US is far less equipped to tackle.

But I don't think that "OWA threat" was where US strategic planning has failed. From the outside, it seems like US plan was to make a single decisive strike, and negotiate with cowed Iran from there. Iran has failed to be cowed by a single decisive strike, and there was no workable Plan B for forcing it to submit if it didn't. That's the key strategic failing. Persistent OWA threat is downstream from that.

The IRBMs are a huge problem. IRGC is using them sparingly, but they were not destroyed.

USAF has actually failed when it comes to 'scud hunt'. It doesn't have the combat ISR or strike capability it needs. As a result, the US basing and resupply situation is chaotic. The loss of the THAAD battery is extremely problematic, which is why they had to move one from SK.

The OWD capability allows Iran to attack tankers anywhere in the Gulf. Even with a surge in GCC state IAMD capability, ships cannot be protected. The failure to address this capability in planning was critical.

Obviously the DoD leadership is cowed by Hegseth and wouldn't give a true assessment, so we have a lions led by donkeys situation, but the operational plan was still bad.

Oh what a great CIA campaign that would be to give the impression that U.S. military isn’t up to par - to motivate and trick our enemies into making stupid moves while we keep our true cards hidden.

Such a crazy thought.

*Losing war after war after war* victory! My beautiful cards… they remain a mystery

Well, one nation mastered gravity, or maybe two nations? Or three? Hard to keep track of intelligence these days

What?

Decades of utter incompetence somehow ended yesterday?

You say that, but that's literally a Russian conspiracy theory with Ukraine: The US baited Russia into thinking the US was weak because of Afghanistan, and then Russia got all ballsy and walked into Ukraine and it was a trap.

The joke is Russia has always been incompetent.

McCain was actually on the money in 2008. He was in the audience when Putin gave that famous "warning to the west" speech in Germany [1].

The quip that Russia is a "gas station with nuclear weapons" about sums it up. The United States being weak or not really has nothing to do with it. They counted on Zelensky cutting and running, the attempted decapitation of Kyiv failed. Russia is swirling the toilet. Without the energy reserves their oligarchs would be worthless. Since most of what is actually valuable about their country is land (where the resources sit) that is sparsely inhabited the end state of China controlling most of what is now Russia is inevitable. By failing in Europe and tying himself to China, Putin has ensured Russian power will continue to erode. Of course, it's not great for China, having to deal with two nuclear armed pariahs on their borders.

The United States is truly blessed simply by being two oceans away from anyone who could remotely challenge it. Properly understood, the Monroe (or "Donroe" doctrine) is less about foreign intervention and more an understanding that if a peer power ever emerged in the Western Hemisphere, most of what appears to make the United States singular would rapidly disappear.

[1] https://www.youtube.com/watch?v=hQ58Yv6kP44

"The United States is truly blessed simply by being two oceans away from anyone who could remotely challenge it." This is tragical misunderstanding of the situation. Both military and ecologically.

Because the most powerful military in the world can only do so much when ordered to execute a crap strategy by civilian leadership. Desert Storm worked because it followed the Powell Doctrine and was bounded to ejecting Iraq from Kuwait. Iraqi Freedom muddled through incoherently because the Bush and Obama administrations didn't want to understand the actual problem set they were faced with, but ultimately ended up with a reasonably stable country after way too much effort and time. Iran is what happens when you have an administration who thinks you can win a war by "just bombing things," which is ironically the stereotype most of the internet has about any kind of military action.

> The article smells of AI psychosis...

China has been doing something similar since 2020 [0] and something we've experimented [1] on for decades [2][3].

They have traditionally been used as target aircraft for training, but a subset are now being used as testbeds for networked ops.

[0] - https://www.defensenews.com/global/asia-pacific/2021/10/20/c...

[1] - https://www.af.mil/News/Article-Display/Article/467196/first...

[2] - https://www.af.mil/About-Us/Fact-Sheets/Display/Article/1045...

[3] - https://commons.wikimedia.org/wiki/File:QF-4E_GulfofMexico_1...

> Now they struggle against Iran

Are they? F-16 wasn’t designed to operate in modern denied environments and more jet damage has been done from friendly fire than Iran.

it may have not been designed to, but it has been modified to.

there may not be any stealth, but there are other countermeasures against missiles, AA, and other fighters.

Yes but everyone knows you can't eliminate the risk from modern MANPADs even with fancy electronic systems and flares. They are conducting combat flights over a country with thousands of chinese MANPADs, the fact they haven't lost more planes is really a great achievement.

The US lost 73 planes in Desert Storm to Iraq and that war only lasted 43 days. By comparison the losses from Iran have been shockingly low and shows how much US navy/airforce capabilities have matured.

MANPADS are mostly relevant to helicopters, pretty useless against a jet at altitudes. You need a SHORAD system at the minimum, like a Pantsir, and they have electronic signatures that make them targets.

Most of the aircraft losses in Desert Storm were due to CFIT and flying into AAA and low altitude weapons. This was a relic of cold war tactics for penetrating Warsaw Pact defences. Once they started flying higher the problem was contained.

See.e.g. https://balloonstodrones.com/2022/10/19/looking-back-at-iraq...

> In the first Iraq war, they went in and won. In the second one, it took years to sort of occupy the country.

The objectives for the two wars were incredibly different, making this an apples to oranges comparison. In the first war, the objective was to push Iraqi troops out of Kuwait (achieved) and then to contain the Iraqi military within Iraq (achieved). In the second war, the objectives were much broader and much harder to achieve. They achieved some objectives (toppling the existing government, defeating the existing military), and then utterly failed in nation building and got into a protracted conflict with state and non-state sponsored military groups.

I wouldn't call nation building a failure. It made significant progress.

Some nation building occurred, but mostly corruption. This was after the first war where they deliberately destroyed civilian infrastructure like power plants, water treatment plants, hospitals and such to create a failed nation. A country that was one of the most modern and developed middle eastern countries before the war.

Yeah, that didn’t happen.

I feel that, "Everyone else is mad" is a really big red flag for one's own mental health and emotional state, rather than an accurate rendition of reality.

I think that the strange comma placements are a red flag.

What are you talking about, Iraq fell in a month not years.

Did it? You mean operational govt fell. The country was fought over for 10 yrs, which the U.S. then backed out.

Israel got a regime change while U.S. tax payers received the bill.

In the end, it’s still a country.

I'm not saying that the war didn't go on for 10 years, it did. Baathist government fell and the US led Coalition Provisional Authority took over running the country. You have US troops in the whole country and US control of all civil institutions. The only way that's not "occupying" the country is if you redefine what the word mean. With your definition France, Norway, Denmark, Poland, and many more were not occupied during ww2.

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.

> Autonomous F-16's already can win against all human pilots.

In a dogfight?

As I understand it, they done this in simulation, and the human pilots did not fare well.

I don't know of the computer aircraft were able to fly beyond human limits (i.e. they don't black out) or not. If the computers can actually fly to what the airframe is capable of, they would have a distinct advantage over the human pilots.

They did a small set of tests with one Pilot IIRC?

The simulator is not the aircraft.

Is that hard to believe? It's got to be one of the easiest situations for AI to win in. You probably don't even need any fancy modern AI; I would have thought MPC would do most of the job.

human pilots have physiological constraints that AI doesnt.

onboard computers are not immune to G's but have a substantially higher threshold until physical mechanisms fail due to strain.

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?

Exactly, that is the point. Not that a drone is outcompeting an F-16. But a hundred cheap drones can definitely get by a F-16 and hit their targets.

An F-16 can't get them all. The goal of war isn't to dog fight, its to bomb, and the F-16 can't stop them.

Pretty much. Drones that disabled /checks notes ~200 barges and tankers in red/black sea in last two weeks cost about $10mil total.

Cost of all the drones that burned down half of russian Oil infrastructure totals less than couple F-35s.

Cheap drones are only effective against soft targets. Many high priority targets are not soft. Even Ukraine invests in expensive systems to defeat harder targets.

https://en.wikipedia.org/wiki/The_Gulf_War_Did_Not_Take_Plac...

Bush Sr knew to keep the foreign-compromised media out of the war and so accomplished the objectives of the war swiftly. Bush Jr had to deal with the media second guessing everything.

Human control in flight has often taken a backseat to automation. For example, the SR-71, built in the 60s flew too fast to rely on human decision making for several tasks. If they were off by a fraction of a second, they would miss their intended recon target; so, both the navigation and sensor control were largely automated.

Seems odd to invest this much into AI flight of aircraft designed for manned flight, when it would be better to build craft spec'd for AI flight without human life support systems.

The US military has thousands of old airframes, many of which can be made flyable with very short turnaround. They already use the F-16 as test targets.

Taking an obsolete airframe that is already paid for and adding AI to it is much cheaper than building a completely new AI-native airframe from scratch. This is finding a new use for old aircraft. You can send a bunch of old AI-driven F-16s on high-risk missions without worrying about a human pilot or cost. The airframes may be relatively obsolete but can still do plenty of damage on their day.

The new 6th generation aircraft are all being designed as "manned optional". That is probably the last generation with affordances for a pilot.

The F16 was the first fully fly-by-wire fighter jet. It's an ideal platform for AI experiments. Having AI "wing men" can be a great tactical advantage, and they're expendable.

The way I see it, this is the JDAM of drones. Will it perform better than purpose built aircraft? No. But the USAF has a few hundred old F-16s entering retirement that would be mothballed otherwise, and strapping some cheap sensors to them is a fast way to play around with the idea while they wait for Lockheed and Anduril to finish the good stuff.

The Right Stuff by Tom Wolfe has a lot of interesting stuff on this wrt the first astronauts (who were sourced from test pilots). The test pilots saw the astronauts as second class citizens, at first, because they weren't actually doing any steering. Once the astronauts became famous, they started demanding more manual takeovers, which the engineers didn't want to build because it could only really hurt, but the engineers were forced to build them!

> when it would be better to build craft spec'd for AI flight without human life support systems.

Yeah this. https://en.wikipedia.org/wiki/Anduril_FQ-44

The "see also" section lists several competitors.

I guess if you already have a load of F-16s it might make sense, but I think that's a stretch.

I think their use of terms like AI and "agent" will make people think "Oh! it's what I use for code generation!"

What’s the point of this useless exercise?

Once you optimise for G forces, speed, cost, lethality - you end up with a missile. We have those already. And we know this already.

Give it more fuel and maybe loitering capability. Something with a jet engine for cruising and a small rocket engine for terminal phase kill.

I’ll have my 100mil now please and thank you.

I'm reminded of Henry Ford and his famous "people would have asked for a faster horse" quote.

This feels like those people all over again.

We know AI can fly things, but why make AI fly an airframe designed in the 1970s around a human cockpit? Ukraine can probably 3d print ten drones that would win in a fight against one of these, for peanuts

I was always wondering about this as well. How much more manoeuvrable can an unmanned drone be given that it doesn't need to lug around life support systems and avoid any G forces that would kill the expensive squishy mammal inside? To your point, it can be indistinguishable from a missile.

Or maybe have a slower but still very nimble carrier drone platform with loitering capabilities that launches very fast and nimble drones/missiles?

Going even further, do these even need to be as nimble and fast as fighter jets? Once you completely remove humans from the equation and replace super expensive and nimble fighter jets with swarms of cheap drones, the loss of a drone or even a carrier becomes much less costly and they won't need to overemphasize survival to the degree that expensive human piloted jets do.

>Ukraine’s F-16s have shot down 2,200 of roughly 3,000 Russian drones and missiles it has intercepted, https://www.defensenews.com/global/europe/2026/07/23/incredi...

So chalk one up for the F-16s in the F-16 vs drone battle. It'd be hard to catch one when your drone does 200 kmh and the F-16 2000 kmh.

That said this AI thing sounds more like a research platform than finished product.

Ok, but those drones were one way ground attack drones and had no self defence capability. An AI piloted f16 has immediate cost and maneuverability disadvantages against a built for purpose dogfighting drone

1 drone: 10-20k

1h of flight time for an F-16: 25k, not counting the missiles used and ammo.

Doesn’t scale. Ergo, useless.

They get quite a lot of use in Ukraine shooting down incoming cruise missiles and drones which Russia launches by the hundred. You'd need a lot of expensive missiles to do the same.

F-16s are reusable missile trucks. F-35s are reusable missile trucks with a smaller capacity but with stealth and better telemetry.

Reusability is just common sense economics. A JDAM costs only $30k USD.

Once you consider the full picture it doesn’t make sense.

There’s a reason why Russia and Ukraine are lobbing ballistic missiles and cheap drones at each other.

There’s a reason why Iran conflict looks the way it does.

First, missiles are not reusable. "JDAM trucks" are. Jet engines are expensive, you know.

Second, US already has a lot of F-16s, and not a lot of dedicated high end drones. US has more F-16s than it has trained F-16 pilots.

Things like this enable turning "manned" airframes into "optionally manned" airframes. Which enables using those F-16s for cheaper for routine "JDAM truck" runs, and also for the kinds of missions that would be too risky for a human pilot. No pilot rescue operations for a downed Nvidia GPU rack.

You could think of an AI F-16 as of a "higher end variant of Reaper", and you wouldn't be wrong exactly.

Third, the best thing? This AI approach transfers. If you can have an AI fly F-16 through reasonably complex missions, you can also adapt it to an F-35, or a dedicated high end drone platform.

F-16s are reusable until they get shot down and you need to be within 30km of the target to drop a JDAM. Any SAM from 60s onwards can deny entry to F-16/F-35. As soon as F-35 opens hatches to drop a bomb it’s on the radar.

Small jet engines are 20-25k/unit

One F-16 gets you 20-30 ballistic missiles, or cruise missiles.

Math doesn’t math in your case, volume and saturation over high tech gadgets that barely work.

So much so even the US military is waking up to this reality.

Have you actually seen what happened in Iran? Were you paying attention? Any attention at all?

US rolled in, convinced Iran's AA to stop doing its thing, and proceeded to bomb targets with impunity. And I mean "impunity" when I say it. When you see B-52s take off with JDAMs under their wings, you know someone's AA is not picking up the phone.

If a B-52 with JDAMs is somehow survivable, an F-16 is not even a question.

The bulk of US losses in Iran in the air were: MQ-9 Reapers, with 10-20 airframes lost reportedly. That's an aging low/slow turboprop platform that's highly vulnerable even to MANPADS.

No F-16s or B-52s were lost to Iran in the air. Zero. One F-15 was shot down by Iran. One F-35 was damaged - and that specific airframe has crawled back to the nearest airbase under its own power and landed gracefully. Clearly, the jets involved were having a good time overall.

(Except over Kuwait. No airspace is more hostile than an incompetent friendly airspace, it seems.)

Now: a (manned) F-16 flight hour is 20-30k, and a single F-16 can truck 2 to 4 GBUs or JDAMs. An unmanned F-16 can shave a few thousands off that. Trucking cheap bombs with an expensive aircraft is quite a lucrative gig if you can reliably get that aircraft back. And the possibility of using those aircraft in SEAD/DEAD missions that are a bit too spicy for meatbag pilots, but can keep airspace open for other platforms? There's considerable value in that too.

Going "full missile" makes sense at extreme standoff distances, or in extremely low survivability scenarios. That, as Iran shows, isn't at all a given when US levels of SEAD/DEAD are in play. Missiles and OWA drones had a place there, but so did high end SEAD jets and jet-powered missile carriers.

MQ-9 might have underperformed, but that seems to be as much of an argument for more cheap missile/OWA hybrids as it is for cheaper strike/loiter drones, and more expensive, more survivable drones. An "F-16D" and even "F-35D" unmanned conversions might make economic sense, given the prices of MQ-9, F-16, F-35, and what we've seen happen in Iran.

  A computer can never be held accountable

  Therefore a computer must make all unpopular management decisions

This whole AI hype is such a convenient cover for the decisions of the 0.001%

It really is. It's the perfect excuse to do what they please.

With apologies, I don't particularly enjoy when people do this, and I don't think it adds a lot, but I want to leave this breadcrumb, I wrote a post on this a couple days ago here, on an anti-AI screed submission, saying this anti-politic, this accountability denial, is very much the point.

Taking this quote and seeing exactly that potential for abuse seems very very much like the point. Strong agree, strongly against leaving non-seeing non-thinking alien intelligences with no accountability and proprietary programming to judge us. https://news.ycombinator.com/item?id=48965083

That's not precisely how I remember that quote... But it's on the Internet, so my memory must be faulty.

Tangent: is AI a bearable arm?

Don't tread on ... my gpu

Probably not, if bump stocks and hicap magazines aren’t either.

Well, hopefully they can figure out something about 0.01% of the price to fly with AI, so we can build at least 10,000 of them. We don't need gold plated hangar queens.

https://en.wikipedia.org/wiki/Macross_Plus

It's probably only a matter of a century before we get mobile suits with mobile dolls a la Gundam Wing and the Virgo suits.

Ender's Game has arrived! One of my favorite books. Spoiler, read the book. Don't just vibe read it.

OpenAI: "Oops, AI hacked somebody accidentally..."

US Military: "Great, let's give it missiles!"

I saw this movie and it was terrible. (Seriously how has no one mentioned Stealth yet?)

It was actually a better movie than I expected it to be. (That was a very low bar, to be clear.)

Boeing 737 MAX often (and semi-regularly) suicides itself and also kills many people in the process. So now fighter jets are AI controlled, by one government (or, any other government, by the way; that problem is not confined to the USA only). I am sorry but this also means that relying on such fighter jets becomes a problem for other countries depending on that. Essentially another country controls here what other people can do here. Of course objectively this only becomes a problem when there is a war between the country that purchased those fighter jets and wants to fight the country that sold the fighter jet (as otherwise it would hardly matter), but imagine how much blackmail-pressure other countries get into as a dependency here. You can just replace the USA with China here and suddenly it may be easier to see the problem. Considering the huge cost of those fighter jets it seems as if the media deliberately avoids wanting to talk about those problems. But we know of those problems already in other areas - see the right to repair movement, which in turn is a right to ownership movement. Who really owns fighter jets?

I thought Captain America personally stopped this initiative

The drones of Ace Combat 7 are coming closer to reality

does this mean we'll have preachy anti-war lectures while fighting them, too?

We can only hope -- bring back M*A*S*H!

I'm waiting for a ship called Nomad to go rogue.

Mandatory watch: https://www.youtube.com/watch?v=pyMNIFZTQkg

based, finally we get to Macross Plus

One Sharon Apple concert before the end!

"AI-controlled" sounds like an unmanned fighter, but there was still a pilot in the cockpit who could take control.

But can the AI drink Jeremiah Weed?

Metal Gear!?

Finally. I hope they do a10!

SAAB did the Gripen

https://www.saab.com/newsroom/press-releases/2025/saab-achie...

Great! Now all we need to do is develop a model for identifying and targeting AI doomers.

Hey look, the lawn darts are now AI-controlled.

Not a great thing for aggressive state prone to electing nutjobs and supporting genocides to have automated killing machines that will not even get PTSD afterwards.

What a waste of resources.

Autonomous driving is still nowhere near 100%, so why suddenly jump to fighter jets? Is this more of Hegseth's foolhardy bluster and injected testosterone? The tech isn't ready. They should start with golf-karts, not fighter planes.

-----------

Edit: Okay gang. If aircraft are easier to fly than cars are to drive, where are all your pilot licenses? If you've faced rush hour traffic, flying a military jet should be a breeze! Please, go volunteer your neighbourhoods to be the first for AI jets to be tested over. We can put a mock enemy target in an empty lot and watch them pick it off, first try, with zero collateral damage. Easy peasy!

Fighter jets don't need 100%. In a war, we accept a much higher rate of casualties than we do in day to day driving. I would also imagine that while planes are harder for a person to pilot, they are easier for a computer. Most of what makes them hard is mechanical, not dealing with unexpected situations.

a strong reason to use a pilot is to protect a large capital investment. in war, you get shot at. things are over stressed, equipment fails.

a pilot has the ability to improvise, and perhaps bring the aircraft back to be repaired and returned to service.

will an AI be as good at 'improvising'?

There are many ways that flying is an easier solve than driving, believe it or not.

I reject the notion that planes are harder to automate.

Planes might be easier, large empty space in the sky

This is a huge waste of resources. F-16s were designed for humans. You can get MUCH better results designing the airframe around the lack of one.

It's a step on from the F-16 flight simulator AI competition in 2020: https://en.wikipedia.org/wiki/DARPA_AlphaDogfight

As another poster already said, building new F-16s would be a waste of resources, but using the old existing F-16s by automating them is likely to be a better use of the existing resources, instead of scraping them. This also allows to gain experience that might be useful for designing the future unmanned airframes.

That’s true, use em up while you got em.

Reply on news.ycombinator.com