Will latest F-35 problems push Norway towards a European solution?

[Dalregementet]

When it comes to recognising patterns computers out perform humans because they can process so many calculations in such a short time. With the right algorithims they can determine all sorts of things, which is why computers are used to determine where buried IEDs are located despite the planters covering their tracks.

Yes, the "right algoritms is the key... But when the computers with the algoritms pop into something they were not programmed to handle then they are dead meat/metal. There are no such thing as the "right" algoritms and there will never be - that's the flaw in your argumentation. That´s why teenagers are able to penetrate high security networks wit state if the art firewalls even though the security algoritms were the right ones. There´s always, always a loop hole. A human can "on the fly" handle that a machine can´t. An example would be two machines flying together - suddenly one explodes. The other machine haven't got any indications from it´s sensors that something is wrong - what does it do? Gut feeling is a powerful sensor

 

[Abraham Gubler]

Well what's going to happen? A UFO? China's FLANKERS tranform into Deceptacons? At least if such a huge phase shift happens no human lives are at stake. Also modern AIs are a lot more adaptive than people give them credit.

The thing is autonomous UAVs are proving themselves every day. People can find semantic holes in simplistic descriptions of the capability but that doesn't mean it is so. This idea that the human can always fluff it on the day is attractive but just crap. And in the case of UAVs if such a situation comes to head it can refer back to large numbers of human operators at a GCS. Those pilots of Apollo 13 would have been dead if they had to solve their problems on their own.

 

[Dalregementet]

Well what's going to happen? A UFO? China's FLANKERS tranform into Deceptacons? At least if such a huge phase shift happens no human lives are at stake. Also modern AIs are a lot more adaptive than people give them credit.

The thing is autonomous UAVs are proving themselves every day. People can find semantic holes in simplistic descriptions of the capability but that doesn't mean it is so. This idea that the human can always fluff it on the day is attractive but just crap. And in the case of UAVs if such a situation comes to head it can refer back to large numbers of human operators at a GCS. Those pilots of Apollo 13 would have been dead if they had to solve their problems on their own.

Well what's going to happen? - I don´t know, you don´t know and so on - that´s the issue! Mans inability to foresee "everything" that can happen is the bottle neck - that also goes for the machines. Their weak spot is their imperfect algoritms which limits them. This is an universal problem. Why don´t we have 100% automatic nuclear power plants? Well, no one is so stupid that they would run a nuclear power plant with imperfect algoritms. Would you?

 

[gf0012-aust]

Well what's going to happen? - I don´t know, you don´t know and so on - that´s the issue! Mans inability to foresee "everything" that can happen is the bottle neck - that also goes for the machines. Their weak spot is their imperfect algoritms which limits them. This is an universal problem. Why don´t we have 100% automatic nuclear power plants? Well, no one is so stupid that they would run a nuclear power plant with imperfect algoritms. Would you?

Thank goodness its not the middle ages - because the capability is already here and already in use (and in some good old swedish kit as well!)

fail safe and redundancy in 2008 is a whole lot different than it was in 2001 when Predators went AWOL due to failed programing.

If you think that its dangerous then you better start telling SAAB to stop some of their own developments - because they obviously agree with AI specialists in the US, UK, Australia, Germany, Singapore etc.....

In fact, if thats your attitude to tech then you'll be surprised to know that the Gripen without it would be unflyable (just like its distant cousin the F-16)

The technical luddites can indulge in some baskervillian wailing to hold their ground - but it's a done deal that you'll see mainstream AI in UAV's by 2015-2020

 

[Dalregementet]

Thank goodness its not the middle ages - because the capability is already here and already in use (and in some good old swedish kit as well!)

fail safe and redundancy in 2008 is a whole lot different than it was in 2001 when Predators went AWOL due to failed programing.

If you think that its dangerous then you better start telling SAAB to stop some of their own developments - because they obviously agree with AI specialists in the US, UK, Australia, Germany, Singapore etc.....

In fact, if thats your attitude to tech then you'll be surprised to know that the Gripen without it would be unflyable (just like its distant cousin the F-16)

The technical luddites can indulge in some baskervillian wailing to hold their ground - but it's a done deal that you'll see mainstream AI in UAV's by 2015-2020

I don´t think we disagree. But the issue was to replace humans with AI and I think I proved my case very well with what I wrote earlier. I admit that I lack specific detailed knowledge about fighter aircrafts but in the case of IT, that is not the case - that is my realm. The company I work for provides systems that have redundancy etc but "fail safe" that doens't exist except in an optimists dream. It´s all about probability - 1 in a million, 1 in a billion etc.

We don´t have nuclear powerplants that run on AI alone and we will never have one´s either. I would be interested to develop one though.... in Australia.

 

[gf0012-aust]

I don´t think we disagree. But the issue was to replace humans with AI and I think I proved my case very well with what I wrote earlier. I admit that I lack specific detailed knowledge about fighter aircrafts but in the case of IT, that is not the case - that is my realm. The company I work for provides systems that have redundancy etc but "fail safe" that doens't exist except in an optimists dream. It´s all about probability - 1 in a million, 1 in a billion etc.

We don´t have nuclear powerplants that run on AI alone and we will never have one´s either. I would be interested to develop one though.... in Australia.

One of the things I find extraordinary in these debates is the use of empirical logic to support a position.

For someone who professes to be in IT (and I spent some 6 years running an IT security team in our Parlt) - I find it interesting that you're attempting to extrapolate plant management and even network management as examples of an asimovian world where AI runs unfettered. To be frank - that 'aint so.


It's especially so wrt to weapons management. there are always going to be "men in the loop" for legal and redundancy reasons.

however whats really cute is the deliberate failure to acknowledge that even now, there are situations where combat is weapons free - ie in "declared space".

Guess what? a UAV in a weaps free environment will have permission to kill whatever is in that grid because the human controllers have already pre-assessd that there are no blue forces in place. It happens now with manned weapons release, and the controls for UAV release are tougher.

Besides, its here and now.

PS if you want to discuss nuke safety, then the swedes should speak to the danes - they seem to have a far more different view of manned swedish power stations WVR of their public beaches. AI control won't make much diff there.

 

[JohanGrön]

PS if you want to discuss nuke safety, then the swedes should speak to the danes - they seem to have a far more different view of manned swedish power stations WVR of their public beaches. AI control won't make much diff there.

Barsebäck, which is what you meant I guess, is closed nowadays.

 

[gf0012-aust]

Barsebäck, which is what you meant I guess, is closed nowadays.

Yes. I should have tightened up on my usage of "tense". (past/present)

 

[Dalregementet]

Guess what? a UAV in a weaps free environment will have permission to kill whatever is in that grid because the human controllers have already pre-assessd that there are no blue forces in place.

Wonder how many innocent women and children that are killed in that type of situations?

 

[gf0012-aust]

Wonder how many innocent women and children that are killed in that type of situations?

Instead of auto-replying all the time make the effort to read exactly what I said.

It's becoming tiresome.

... and please, no more IT and Industry stories - there are others in here who also have worked or work in both.

 

[Dalregementet]

Instead of auto-replying all the time make the effort to read exactly what I said.

It's becoming tiresome.

... and please, no more IT and Industry stories - there are others in here who also have worked or work in both.

This discussion was, as I interpreted it, a discussion about having fighter aircraft with AI, thus replacing human fighter pilots. I objected to that idea due to that there are no perfect algoritms in this world - they can be good but there are always flaws. I do see that AI will be more and more used, but, hopefully, with a human that takes the ultimate decison to open fire or at least to designate an area to operate in where there are no blue forces or civilians. IT and industry is key in this area and in many respects more advanced than what the has military.

Besides, the type of IT I work with is not of the standard type, it is true state of the art and it will now be used in Civil Security applications.

 

[Vivendi]

latest in Norwegian Newspapers:

http://www.dagsavisen.no/innenriks/article378751.ece

Yes, we will give a clear recommendation, says Pål Bjørseth leader of the project team which will recommend which fighter Norway should buy.
...
Roughly six months ago the two competitors Saab and LM their binding offers on 2,500 pages each. "Since then we have worked day and night going through the offers and evaluating" says Bjørseth.

Both [competitors] had to respond to roughly 1000 requirements. These responses the team has used six months to evaluate. 27 groups with up to 100 people have evaluated the technical capabilities for the planes. People from DoD, FLO, FFI and LUKS have done this.

Purchase, maintainance and upgrade costs have been evaluated by people from FFI and FLO, wheraeas offsets have been evaluated by FFI, DoD, Dept. of Commerce, and 'Innovasjon Norge'.

Only Mr. Bjørseth and a few others know the result of all these evaluations.
The people in the different evaluation groups have not had access to eachothers work.

Mr Bjørseth stresses that even if his team will give a clear recommendation on which plane to choose, the politicians will make the decision.

Defence politicians in the parlament makes it clear that the political room for manouvring will be less the more clear the recommendation from the professional evaluation team. If Bjørseth and his team recommends that the F-35 is clearly best for Norway, not in the least in a 30-40 years perspective, it hardly matters that the Swedes offer huge offsets.

Bjørseth also says that the government has made it clear that the capabilities and costs are more important than industrial collaboration.

"But still I don't believe that LM feel completely confident that they will be chosen. In my opinion the government has built the foundations for a credible competition, and our project has benefited from this, he says."

"A clear recommendation" can only mean the F-35 of course. Funny, they have tried to keep the conclusion secret, and then they almost openly advertise it like this. OTOH for people knowledgable about both planes their conclusion was already a given.

The next step now is (to my knowledge) December 19 when the government will make public their decision. Of course given the huge offsets offered by Saab one should not yet completely dismiss the possibility that they will pick Gripen however in my opinion, with the above news release the possibility of that happening has dropped considerably. To me it looks like LM will win this now.

There has been a lot about this in the Norwegian media for several weeks -- I wonder if it will calm down now. Or will Saab try to "fight back"?


Vivendi

 

[gf0012-aust]

This discussion was, as I interpreted it, a discussion about having fighter aircraft with AI, thus replacing human fighter pilots. I objected to that idea due to that there are no perfect algoritms in this world - they can be good but there are always flaws. I do see that AI will be more and more used, but, hopefully, with a human that takes the ultimate decison to open fire or at least to designate an area to operate in where there are no blue forces or civilians.

read my numerous responses (again)

IT and industry is key in this area and in many respects more advanced than what the has military.

Am curious as the depth of your experience in IT when you make comments like this. All civil IT systems (and esp western incl scandinavian countries) have to be accredited by national intelligence - and that includes defence intelligence. Accreditation means that the entire solution needs to be vetted, it means that the entire solution set is provided to the mil-intel and def-intel people to see whether it can breach and compromise anything in the mil-IT, def-IT complex. In addition, we (as in friendly countries) share that material out of common necessity.

Besides, the type of IT I work with is not of the standard type, it is true state of the art and it will now be used in Civil Security applications.

See above. Sweden is no different. we work with the swedes on these types of issues.

I could give you a recent horror story on how a def-crypto specialist broke a civil system that was regarded as the most secure in the southern hemisphere in under 30 seconds. Like a lot of people he seemed to think that commercial systems were more robust than military ones.

It's never a perfect world, but I'd bank on a mil crypto breaking a civil system any day of the week.

COTS/MOTS/GOTS all have their place - but MOTS is the only one that gets to crypto break and interrogate all others - your civil systems aren't even allowed on the national system unless cleared - and that means absolutely gutted by analysts to make sure that we don't get any nasty surprises and that they don't overlap other undeclared systems not in the public view.

 

[Feanor]

Instead of auto-replying all the time make the effort to read exactly what I said.

It's becoming tiresome.

... and please, no more IT and Industry stories - there are others in here who also have worked or work in both.

He has a point. The recent reports of a UAV bombing of Pakistan that left civilians dead?

 

[Abraham Gubler]

He has a point. The recent reports of a UAV bombing of Pakistan that left civilians dead?

He and this are no points. You are talking about a Predator attack that would have had humans sitting there pointing and clicking on the building to be enaged. Its not as if manned aircraft haven't killed civilians by mistake either... ever heard of Dresden? Targeting includes calculating acceptable collateral damage in this case something like taking X number of terrorists but killing or injuring Y number of civilians. The targeters would have quite literally worked out how much damage those terrorists would do if left untargeted... and if that's more than the collateral damage ordered the attack.

In fact an unmanned aircraft with high speed image processing is less likely to mistake a friendly vehicle for an unfriendly one as in all the recent blue on blue kills. Because the computer system can do things like precisely measure the spatial relationship between things on the image rather than rely on an impression of the image. Like "is that BMP-2 or a Warrior with add on armour... they both have turrets and pointed bows?" The computer would calculate "5.32m between real road wheel and gun barrel indicates a Warrior as same distance on a BMP-2 is 4.19m."

This kind of debating point is ill informed and not helpful. If you are just trying to score some points save your time and effort. All you will achieve is burning up what ever capital I have in respect for you, your knowledge, your motivations and your opinions.

 

[SlyDog]

One example there a human mind are superior to AI:

A human can recognise a dog "as a dog" even if it is the first time that human seeing the dog (let's say it is a dog of an unusual breed). I don't think a computer have reached that far get.

 

[Dalregementet]

read my numerous responses (again)

Am curious as the depth of your experience in IT when you make comments like this. All civil IT systems (and esp western incl scandinavian countries) have to be accredited by national intelligence - and that includes defence intelligence. Accreditation means that the entire solution needs to be vetted, it means that the entire solution set is provided to the mil-intel and def-intel people to see whether it can breach and compromise anything in the mil-IT, def-IT complex. In addition, we (as in friendly countries) share that material out of common necessity.

See above. Sweden is no different. we work with the swedes on these types of issues.

I could give you a recent horror story on how a def-crypto specialist broke a civil system that was regarded as the most secure in the southern hemisphere in under 30 seconds. Like a lot of people he seemed to think that commercial systems were more robust than military ones.

It's never a perfect world, but I'd bank on a mil crypto breaking a civil system any day of the week.

COTS/MOTS/GOTS all have their place - but MOTS is the only one that gets to crypto break and interrogate all others - your civil systems aren't even allowed on the national system unless cleared - and that means absolutely gutted by analysts to make sure that we don't get any nasty surprises and that they don't overlap other undeclared systems not in the public view.

In Sweden we have numerous sensitive systems, some are only used by the military, intelligence, police forces, government and so on. Some systems gives you an overview over the air space, the sea, the national grid (electricity) and so on. Most of these are supplied by defence related companies and in most cases, they rely on input from external civilian sources - it´s no rocket science. When we talk about IT-security, then there are many aspects to consider: network security, crypting, redundancy, data and information access to just mention a few. One big issue is that you want to bring in data/information from many sources, some classified, and when you then bring all that together, the sum, the synergy can be really top secret. The key thing is being able to integrate disparate systems, perform analysis and data fusion on top - keeping all secure and secret. Now, the system that my corporation has developed has so far costed more than 400 MUSD - no public funding in any way - Very few governments has the muscle to spend an equivalent sum of money one one single platform. On top of the platform we then create specific applications. I can´t estimate how much that is, but it´s far more than the previous figure. So yes, when the system are used in intelligence or military application, it has been accredited and the same goes on the civilian side though the accreditation is carried out by other entities.

 

[Dalregementet]

He and this are no points. You are talking about a Predator attack that would have had humans sitting there pointing and clicking on the building to be enaged. Its not as if manned aircraft haven't killed civilians by mistake either... ever heard of Dresden? Targeting includes calculating acceptable collateral damage in this case something like taking X number of terrorists but killing or injuring Y number of civilians. The targeters would have quite literally worked out how much damage those terrorists would do if left untargeted... and if that's more than the collateral damage ordered the attack.

In fact an unmanned aircraft with high speed image processing is less likely to mistake a friendly vehicle for an unfriendly one as in all the recent blue on blue kills. Because the computer system can do things like precisely measure the spatial relationship between things on the image rather than rely on an impression of the image. Like "is that BMP-2 or a Warrior with add on armour... they both have turrets and pointed bows?" The computer would calculate "5.32m between real road wheel and gun barrel indicates a Warrior as same distance on a BMP-2 is 4.19m."

This kind of debating point is ill informed and not helpful. If you are just trying to score some points save your time and effort. All you will achieve is burning up what ever capital I have in respect for you, your knowledge, your motivations and your opinions.

In an asymetric war, which vehicle is "unfriendly" if you spot one or more Totota Landcruisers? I guess it depends on the driver and/or the passengers?
If it´s a BMP-3, a warrior or a CV90 then it´s easier. Save your breath regarding "ill informed" etc, I´m just amazed that you doesn´t get the point.

 

[Abraham Gubler]

In an asymetric war, which vehicle is "unfriendly" if you spot one or more Totota Landcruisers? I guess it depends on the driver and/or the passengers?
If it´s a BMP-3, a warrior or a CV90 then it´s easier. Save your breath regarding "ill informed" etc, I´m just amazed that you doesn´t get the point.

So how does the human pilot tell apart two Landcruisers? I get the point you are trying to make but it just doesn't add up. You are not giving the autonomous sensor system its due and crediting human's with powers of intuition we just don't have.

 

[stigmata]

Its not as if manned aircraft haven't killed civilians by mistake either... ever heard of Dresden?

Mistake ?
The soul purpose of firebombing Dresden was to kill as many civilians as possible. Bomber Harris got so exited over the overwhelming numbers of civilians killed that he declared "Berlin next". In hope of murdering an even greater numbers of civilians, alltho that failed for a numbeer of reasons.

This kind of debating point is ill informed and not helpful

Yes indeed

http://www.spartacus.schoolnet.co.uk/2WWdresden.htm

Mod edit: It would be highly advisable to go back over this post and edit the content, as the site linked does not support some of the assertations made. Additionally, while the Bombing of Dresden in 1945 continues to be controversal, information such as that provided has a tendency to manipulated according to the agenda of the poster or site. As an example of this, three different 'normal' population totals for Dresden in 1945 were encountered in a few minutes of search, these were ~350,000 then ~650,000 and lastly 1.2 million and that as a result of POWs and refugees there would have been additional transients. Given that according to the City of Dresden on December 31st, 2006 it had a population 509,565 it would seem to suggest the population listed on the reference was incorrect, and by extension at least some of the rest of the information and links provided would also be inaccurate.
-Preceptor