Crash at Reagan National Airport, DC. Small aircraft down in the Potomac.

How long between hiring and actually acting as a controller? I honestly don’t know.
Acting as a controller can vary a lot depending on previous experience and complexity of facility. I have never seen it, and I’ve seen a lot, be more than a year. Now becoming an FPL, full performance level, can be long time, like more than 3 years at Centers.
 
Factored landing distance is a 121 reg (121.195) and it's for planning purposes only.
We discussed 121.195 at length. The post you quoted was not about 121.195.
 
I had an up close encounter Friday at about 1445 ~E/NE of Outlaw Field with a flight of two Blackhawks. They weren't squawking ADSB as far as I can tell and I didn't see them with ADSB/IN or on ADS-B Exchange when reviewing my flight

I'm confident they would have been talking with Ft Campbell Approach and I have ADSB out. Why they were allowed to come so close to me is a mystery. I made what I call an evasive maneuver to avoid a very close encounter or worse.

Of course I thought of this incident in DC where the chopper apparently wasn't squawking ADSB either.
A lot of times they don't use Campbell Approach and utilize the Campbell route structure for recovery into the KHOP. I wouldn't expect to see a lot of Army helo traffic on ADS-B. You're better off monitoring or talking to Tower and/or Approach.

Be safe out there, and see ya in the AO!
 
….. I wouldn't expect to see a lot of Army helo traffic on ADS-B. You're better off monitoring or talking to Tower and/or Approach.
Be safe out there, and see ya in the AO!
Why is that ? Is it because of some kind of policy ( obviously we are not talking here warzone deployments) or just never got around installing ADS-B on their helos ?
 
This profession will be a prime candidate for AI taking over … practically tailor made for this sort of repetitive but “ never have a bad day and never make a mistake” workflow.
LMAO! This shows a fundamental lack of understanding of two topics. What we call AI today is more A than I. They're built on LLM and would be terrible for any type of creative role requiring dynamic, outside-the-box thinking like ATC has to do often.

Obligatory XKCD: https://imgs.xkcd.com/comics/machine_learning_2x.png
 
LMAO! This shows a fundamental lack of understanding of two topics. What we call AI today is more A than I. They're built on LLM and would be terrible for any type of creative role requiring dynamic, outside-the-box thinking like ATC has to do often.

Obligatory XKCD: https://imgs.xkcd.com/comics/machine_learning_2x.png
Really ?
We are build on LLM like information acquisition and retrieval too and 99% of what we do is recalling and applying existing (or parallel ) patterns that we have seen before ( it is generally referred to as being experienced and “ having seen it all”)
You are overestimating human creativity as very few people engage in truly “outside of the box” thinking and come up with completely new stuff and these that actually do we tend to refer to as geniuses - everything’s else is just recalling and applying patterns.
 
Really ?
We are build on LLM like information acquisition and retrieval too and 99% of what we do is recalling and applying existing (or parallel ) patterns that we have seen before ( it is generally referred to as being experienced and “ having seen it all”)
You are overestimating human creativity as very few people engage in truly “outside of the box” thinking and come up with completely new stuff and these that actually do we tend to refer to as geniuses - everything’s else is just recalling and applying patterns.
I agree with pretty much everything you said. Except for applying it to ATC. I'll let them speak for themselves, but I don't imagine much of even the "mundane" things they do other than maybe issuing IFR clearances are always the same. Between weather, stupid pilot tricks, dynamic traffic levels, etc. I can't imagine any AI tech we currently have handling that very well unless we move to a point where there's no part 91 traffic and everything is HEAVILY metered.
 
Start with the JO7110.65AA and associated notices. The pdf is 794 pages.

Then you have to memorize the airspace over which you will have control. All navaids, airway radials, procedures, etc.

Then you have the applicable letters of agreement, coordination protocols specific to your airspace, etc.

Then you need experience in evaluating the traffic situation, formulating a plan, and implementing it. That takes a lot of practice.

It takes several years from hire to reaching full qualification at your facility. Washout rates are high.
I live near KLEX in Lexington Ky. I'm acquainted with one of the controllers there. After the take-off accident several years ago leading to a fatal crash of an airliner, I saw firsthand the devastating effects such an accident has on the ATC community. Our church actually implemented a 2+ years free counseling program for any who wanted it, and many availed themselves of it. But obviously for some people, that kind of heartbreak is the end of their career in ATC. these folks not only need to be smart, resourceful, and cool under fire, they also need the emotional reserves and support in dealing with potential tragedy that might meet them on any given day at work.
 
Why is that ? Is it because of some kind of policy ( obviously we are not talking here warzone deployments) or just never got around installing ADS-B on their helos ?
I will say that Operations Security (OPSEC) plays large part of it. Some might argue the legality or necessity of it, but those decisions are made far above our level. Some of these assets are highly specialized as are the precious cargo they carry, and it has been determined that they should not be able to be tracked.
 
Give me enough data I can know Unit readiness…OR rates…Unit capabilities, which plenty of hostile actors would love to have. We already have plenty of hostiles logging take offs and landings all over the world. Don’t plan on any changes anytime soon…but I could make a case for VIP units…
 
Give me enough data I can know Unit readiness…OR rates…Unit capabilities, which plenty of hostile actors would love to have. We already have plenty of hostiles logging take offs and landings all over the world. Don’t plan on any changes anytime soon…but I could make a case for VIP units…
Yeah, certainly there is no acute shortage of 70+ year olds VIPs waiting in the wings ready to assume the incredible burden of being a public servant in case something happens to the current one …
 
The authority for a blanket masking of movements by removing the safety benefits of ADS-B should be questioned.
Agree. There is no reason to go dark. The Senators/etc just don't want us to know that the PATs are a taxi ride for them. Training......lol.

If the real deal was happening, it will be like the fall of Saigon, so not enough seats.
 
Yeah, certainly there is no acute shortage of 70+ year olds VIPs waiting in the wings ready to assume the incredible burden of being a public servant in case something happens to the current one …
Do you want our enemies choosing who serves us?
 
We should be starting to work our way out of that hole at this point. A lot of the people who started careers after the PATCO strike are retirement age now, so that is really hurting the system too. Losing a pretty big group of the most experienced people.
PATCO replacements are all long gone. ATC is forced retirement at age 56. It is a cognitively heavy job.

PATCO-replacement retirements were what started the understaffing dang near 20 years ago. Not only have we not recovered, the second post-PATCO generation is starting to become eligible to retire... And the way ATC seems to be these days (woefully understaffed, 6 day weeks, hard to get vacation, etc) is not going to lead those folks to stick around until they're 56.
What’s the latest with the NTSB on this? Have they done a press conference since the one they did a couple days after where they said they couldn’t comment on altitude until they learned more about altitude readouts on the Tower’s Radar?
I think they have, since additional facts have come out, but they didn't put video up.
This profession will be a prime candidate for AI taking over … practically tailor made for this sort of repetitive but “ never have a bad day and never make a mistake” workflow.
"AI" as we know it today is pretty much just for written words and images, and it doesn't "know" anything. It merely strings together words that have the highest probability of occurring after one another.

It also requires an incredible amount of data. Unless someone's been saving the "tapes" for the past few decades, there's probably not much in the way of training data available. Feed it everything that exists now, and there's probably still enough nuance and little exceptions to everything that it would not work well enough for a LONG time.

If they try to make AI take over ATC I'm gonna go flying and ask for contact approaches, cruise clearances, etc until it crashes.
I will say that Operations Security (OPSEC) plays large part of it. Some might argue the legality or necessity of it, but those decisions are made far above our level. Some of these assets are highly specialized as are the precious cargo they carry, and it has been determined that they should not be able to be tracked.
I'm sorry, but if Air Force One with POTUS aboard can squawk ADS-B, so can a random helicopter on a training flight.
Do you want our enemies choosing who serves us?
Do you think they're not already heavily influencing our elections?
 
I have no direct experience with this subject but it does sound like there are some deep inefficiencies built into the system.
I would think this is the kind of job that does not require an extensive body of knowledge ( how many relevant rules and regulations are there after all ) and is more the type of job that relies on natural ability to think and act calmly and quickly - in other words you will either make a good ATC or you wont and which is which should be apparent rather than quickly.
Of course, as I mentioned I have no direct experience with the subject so I may be completely off here …
You are not off here. There are certain abilities that are required. Can you evaluate things in motion. Can you see where things are ‘now’ and mentally picture where the will be a minute from now. Can it be taught? Only if the ability already exists.
 
I think that's where the DEI factor comes in. In the long past, if one didn't qualify on position after a fixed period of time, they were let go. Now it's train until the trainee can meet minimum standards, no time limit. Need an active ATC'er to weigh in on that.
Not active but I can weigh in. There was the ‘train to succeed program’ that came about in the era after the strike. Congress was upset about the ‘failure rate’ of new hires reaching certification. The ‘wash out’ rate was to high. Ergo, train them until they can. So numerous days, weeks, months and years, yes years, were spent on persons who didn’t have the inborn skills to separate two flys with a screen door, much less the cheeks of their azz with a crow bar.
 
The authority for a blanket masking of movements by removing the safety benefits of ADS-B should be questioned.
I believe they have run risk analysis and find it acceptable risk to run in the blind.

Maybe the bean counters will reconsider under further review of current events.

In my case I know Ft Campbell had me on ADSB. They watch that area like a hawk as CKV sits within the KHOP veil. And CKV is on the doorstep of R-3701 A and 3702 A-B. Campbell will pop up on CTAF and call you out if they are the least bit concerned about you.
 
I believe they have run risk analysis and find it acceptable risk to run in the blind….
Uh, no. One of the biggest challenges .AF had with ADS-B was cost effectively retrofitting it to specialized platforms with custom FMS code. Simply plugging a box into the 1553 databus doesn’t always work.

C-21s amortized out to $1.7M/per aircraft for ADS-B out and that’s on an off the shelf Lear 35. OTOH, the T-6’s ran about $32k/plane.
 
Uh, no. One of the biggest challenges .AF had with ADS-B was cost effectively retrofitting it to specialized platforms with custom FMS code. Simply plugging a box into the 1553 databus doesn’t always work.

C-21s amortized out to $1.7M/per aircraft for ADS-B out and that’s on an off the shelf Lear 35. OTOH, the T-6’s ran about $32k/plane.
This. It took a long time to get the box to play nice with the integrated Avionics System. It's still kinda finicky.

That said, it's STILL not a case of flying VIP's, at least around here. The pipe hitters being flown around here are particularly edgy about being seen. Ya know, the whole quiet professional thing and all. And that's all I'm gonna say about that.
 
"AI" as we know it today is pretty much just for written words and images, and it doesn't "know" anything. It merely strings together words that have the highest probability of occurring after one another.
This is true of the major LLM packages, but not the MoE AI software. There is a big difference between the AI you see on Google and the AI being used in industry.

There are AI packages doing low-level engineering work and creating drawing packages that can go straight into automated manufacturing systems.

Ready for ATC? Probably not.

Ready to screen bags at TSA checkpoints? Probably already better than humans.
 
In order to avoid opsec issues, there was a requirement to have the cability to modify tail-specific identifiers (e.g., Mode S transponders and communication equipment). I don't know if the Army took advantage of that capability.

The cost of that capability is on top of the usual aircraft integration costs. wasn't cheap.
 
PATCO replacements are all long gone. ATC is forced retirement at age 56. It is a cognitively heavy job.

PATCO-replacement retirements were what started the understaffing dang near 20 years ago. Not only have we not recovered, the second post-PATCO generation is starting to become eligible to retire... And the way ATC seems to be these days (woefully understaffed, 6 day weeks, hard to get vacation, etc) is not going to lead those folks to stick around until they're 56.

I think they have, since additional facts have come out, but they didn't put video up.

"AI" as we know it today is pretty much just for written words and images, and it doesn't "know" anything. It merely strings together words that have the highest probability of occurring after one another.

It also requires an incredible amount of data. Unless someone's been saving the "tapes" for the past few decades, there's probably not much in the way of training data available. Feed it everything that exists now, and there's probably still enough nuance and little exceptions to everything that it would not work well enough for a LONG time.

If they try to make AI take over ATC I'm gonna go flying and ask for contact approaches, cruise clearances, etc until it crashes.

How do you know that you know anything and what does that even mean knowing something ?
Our brains work using the same basic binary primitive switches numbering in billions as do computers so how do you know that whatever you consider knowledge is not just some sort of multidimensional vector using basic vector maths and dot product like comparisons producing an illusion of what we refer to as intelligence … it has to be something along these lines given that there is nothing rally magically more complicated inside of our heads beyond a network of binary switches.

Btw , given huge advancements in AI perception and scoring abilities , I fully expect AI to start using almost exclusively self reinforcing learning where it simply produces effectively unlimited number of randomly generated scenarios and relies on what is basically an insanely accelerated , brute force , natural selection like process to accumulate knowledge …
We shall see …
 
Last edited:
How do you know that you know anything and what does that even mean knowing something ?
Our brains work using the same basic binary primitive switches numbering in billions as do computers...

Hmmm
 
The main obstacle to AI for ATC is cost, not technical. I'm sure the same brains that developed self driving cars and LLMs could automate substantial portions of ATC. But nobody is going to throw the billions at the problem that would take, because there's no profit in it.
 
The main obstacle to AI for ATC is cost, not technical. I'm sure the same brains that developed self driving cars and LLMs could automate substantial portions of ATC. But nobody is going to throw the billions at the problem that would take, because there's no profit in it.
We are in early stages of AI … 30 years ago it took months and computing render farms worth hundreds of thousands of dollars to generate the original Jurassic Park dinos - these days a kid with a $300 graphic card and an open source program can work the magic in an afternoon …

 
How do you know that you know anything and what does that even mean knowing something ?
Our brains work using the same basic binary primitive switches numbering in billions as do computers so how do you know that whatever you consider knowledge is not just some sort of multidimensional vector using basic vector maths and dot product like comparisons producing an illusion of what we refer to as intelligence … it has to be something along these lines given that there is nothing rally magically more complicated inside of our heads beyond a network of binary switches.

Btw , given huge advancements in AI perception and scoring abilities , I fully expect AI to start using almost exclusively self reinforcing learning where it simply produces effectively unlimited number of randomly generated scenarios and relies on what is basically an insanely accelerated , brute force , natural selection like process to accumulate knowledge …
We shall see

"'Oh, that was easy,’ says Man, and for an encore goes on to prove that black is white and gets himself killed on the next zebra crossing.

Douglas Adams, Hitchiker's Guide to the Galaxy
 
FWIW: AI works fine as do most computers till the power goes out. Carbon units may not be perfect but most run fine without it.
Last I checked radios also don’t run so well without power. What’s the human ATC without power gonna do, yell really loud?
 
Light signals :cool:
That works if you're in a tower I guess. Of course, if the tower goes dark then you could just revert to pilot control. Light gun signals probably don't work well in the en route or terminal environment.
 
How do you know that you know anything and what does that even mean knowing something ?
Our brains work using the same basic binary primitive switches numbering in billions as do computers so how do you know that whatever you consider knowledge is not just some sort of multidimensional vector using basic vector maths and dot product like comparisons producing an illusion of what we refer to as intelligence … it has to be something along these lines given that there is nothing rally magically more complicated inside of our heads beyond a network of binary switches.
You can get into all kinds of philosophical discussions about this kind of stuff and I rather enjoy that... But there is a 10,000 word per post limit on PoA and that's a lot of typing. It'd also only scratch the surface, and it's much more fun in person over a beverage.

But stuff like ChatGPT gets things wrong frequently because it doesn't associate any meaning with the words - It merely looks at what words are most likely to occur next based on all of the training text it's been fed. I think they are starting to work on this, but AI at this point should not be trusted whatsoever. Always verify the answers you get.
Btw , given huge advancements in AI perception and scoring abilities , I fully expect AI to start using almost exclusively self reinforcing learning where it simply produces effectively unlimited number of randomly generated scenarios and relies on what is basically an insanely accelerated , brute force , natural selection like process to accumulate knowledge …
We shall see …
I had an interesting discussion with my brother about pilots vs technology (he works for a company that is BIG in the AI space) and I said something about how pilots can use their knowledge to come up with creative solutions to any given problem. He came back with "Well, we can just run 10,000 different simulations per second and choose the one with the best probably of a successful outcome." When you have that level of brute force, creativity isn't as big a deal any more.
 
I had an interesting discussion with my brother about pilots vs technology (he works for a company that is BIG in the AI space) and I said something about how pilots can use their knowledge to come up with creative solutions to any given problem. He came back with "Well, we can just run 10,000 different simulations per second and choose the one with the best probably of a successful outcome." When you have that level of brute force, creativity isn't as big a deal any more.

Computers will never overcome human intelligence. As proof I offer you this from Captain James T Kirk:


:devil:
 
Computers will never overcome human intelligence. As proof I offer you this from Captain James T Kirk:


:devil:

ah, but sometimes the computer can learn (see the end of "WarGames")
 
What’s the human ATC without power gonna do, yell really loud?
Not at all. They and the aircraft will simply follow the existing guidance for just such an occurance as in past events.
 
Not at all. They and the aircraft will simply follow the existing guidance for just such an occurance as in past events.
And that’s different for when an AI ATC goes out how?

Look I’m not advocating for AI ATC. Despite being somewhat in the tech space, I mostly feel that AI is massively overhyped as a solution to most problems people try to throw it at, but “AI can’t function without electricity” is not a very compelling argument against it. Many many things we rely on require electricity and in safety critical applications we provide redundant sources of power. Safety critical AI would be no different than a safety critical non-AI computer system or electro-mechanical device in this respect.
 
Back
Top