First TL-9 items

[Farmer]

Originally Posted by Opellulo

"with deuterium and tritium" it's like in the title, and that's the problem because the ending result it's not the intended "Clean" Fusion Power.

Which is moving the goal posts. Your original point was that the temperatures were only 1,000,000 degrees. They're 100 times that. They are progessing and withing the scope (90 years) of TL9 medium, not unreasonable to think further progress will be made.

Originally Posted by Opellulo

My daily work is with Machine Learning, and this article is not unlike the tens i receive daily on my LinkedIn feed. Development is going fast sure, but in a specific direction that's not even "dumb AI" but simply "refined algorithm to better prey on your credit card". There are probably out there different approaches and researches for something different but, well "still not there"

Indeed, we're not. But machine learning is doing things like learning how to recognise faces, it's reviewing trials and papers and figuring out where something probably was discovered but not realised. It's learning how to tranlsate languages better, and so on. The underlying process of learning, is improving.

Originally Posted by Opellulo

In open road, without supervision, paid by customer and sold as "here" because the entire business model of some idiotic twitter troll is that of constantly feed its companies with fresh venture capital.

Huh? Sounds like a political interpretation. There are many self-driving prospects that work, that need some refinement and really some way of talking to each other to help make them flow in smart ways rather than just reactively.

Originally Posted by Opellulo

Facebook and Epic continuously spew out "metaverse" nonsense because their wet dream is streaming advertisement on your eyeballs 24/7. Still a technology searching for a purpose.

Again, actively being used for remote or microscopic medical procedures, engineering tasks, general technical operations, and many other things. Just because the "game" versions aren't popular, doesn't mean they're not being used in less "sexy" ways.

Originally Posted by Opellulo

How may of these tracks are more than a test or a "political show-off projects"? (like the Shangai maglev).

The ones that have been going for nearly 30 years and, by definition, are cost effective (your original complaint)?

Originally Posted by Opellulo

It's matter of MONTHS that one of those creepy robotic dogs will be unleashed against immigrant and celebrated as an effective way of "border control".

OK, definitely politiical. Let's come back in June next year and check this.

[maximara]

Originally Posted by Phantasm

As far as I'm concerned, we're already TL9 in computer technology. My own hope is that bionic prostheses come along soon.

IMHO computer technology is more a refinement of TL 7-8 with a few TL 9 advances be made. Apple's M1 for example is a refinement of a chip design developed back in the 1985 that Apple has been improving for a about a decade.

One issue is unlike the TLs before it TL 9 doesn't really have "signature" tech

Never mind, in the real world the boundary between TL x and TL x+1 is fuzzy as all get out. For example does Watt's very small improvement to the Newcomen atmospheric engine (TL 5) make it the the first TL 6 steam engine or was it just a refinement of TL 5?

[Þorkell]

Originally Posted by Farmer

Huh? Sounds like a political interpretation. There are many self-driving prospects that work, that need some refinement and really some way of talking to each other to help make them flow in smart ways rather than just reactively.

The "some refinement" appears to be the hard part. There are many components of what is needed for a self-driving car available right now as I mentioned in my post above. However the trick appears to be to get that whole thing to work. I mean there have been stories of "self-driving" cars flummoxed by rain which makes them oh so useful in climates that get lots of it.

[the-red-scare]

Let’s go back to basics. And by basics I mean the Basic Set’s list of signature technologies.

Artificial Intelligence: Yes. It doesn’t say sapient, volitional artificial intelligence.

Real-Time Virtuality: Yes. This is totally possible but not particularly popular.

Robot Cars: Kinda. Autonomous robot cars have certainly driven, but they’re not at all ready for mass production.

Space Elevators: No way. I don’t even think they’d be TL9, but that’s what it says.

Manned Interplanetary Spaceflight: Not yet.

Electrolasers: No. I’m not sure these are ever going to be a thing.

Heavy Laser Weapons: No. we have heavy-as-in-massive laser weapons, but damage-wise they’re pathetic.

Battlesuits: No.

Combat Robots: Kinda. We have some (particularly the flying kind) but they’re remotely piloted.

Designer Viruses: Kinda. As far as we know not as deployable weapons but it sure seems inevitable.

Micro Fuel Cells: Not sure. I think advancing battery technology is kind of taking this role.

Deuterium-Hydrogen Fusion: No. Assuming this means deuterium-tritium and commercial production.

High-Temperature Superconductors: Kinda. Only on an experimental level at this time.

Human Genetic Engineering: Kinda. Very early CRISPR treatments are underway.

Tissue Engineering: Kinda. Some simple organs are grown, and cell-based synthetic meat is on the cusp of entering the market.

Artificial Wombs: No. Working on it.

Cybernetic Implants: No. there are implants and prosthetics, but none I’d describe as traditional cybernetics.

The fact that so many of these are at the prototype stage means depending on how you want to draw the fuzzy line either very few TL9 technologies exist or half of them do.

[Rupert]

Originally Posted by Opellulo

- Drones: this is tricky... Apparently someone thought that it was a good idea to arm a drone and let it go killing without human supervision. So yes probably the only operational TL9 item we got is also the worst one.

It's also something you can do at TL7 if you want. Recon drones date back to TL7, though they were something that was considered a very secret technology so the public didn't hear about them (especially as nobody got captured or killed when they got shot down, so it was easy for all concerned to simply not mention them). Putting missiles, bombs, or guns on them was well within the technology of the time - they weren't armed because the political climate of the time meant there was no real use for such a device.

[Rupert]

Originally Posted by Fred Brackin

......and some items wth stats that UT would make TL9 were probably available then. Other things that UT has as TL9 may never show up. That truck bed fission power generator might be one of them.

I think that's more a matter of regulation and demand than technical limitations. Whether we could make one that would retail for $100,000 is another matter.

[Opellulo]

Originally Posted by Þorkell

The "some refinement" appears to be the hard part. There are many components of what is needed for a self-driving car available right now as I mentioned in my post above. However the trick appears to be to get that whole thing to work. I mean there have been stories of "self-driving" cars flummoxed by rain which makes them oh so useful in climates that get lots of it.

Let's clear a thing about autonomous driving vehicles: the road to their implementation is long (if ever will be one) not because of technological limits (and there are many: from being confused to rain, to a slightly redder moon to a discolored crossing) but for a specific philosophical/ethical/legal problem which is LITERALLY the trolley problem.

In the event of an inevitable collision, whose safety the autonomous vehicle will prioritize?

The resulting collision will not be an "accident" but the expected result of a decision embedded within the system, a decision that someone will need to be accounted for.

This is the dark spot of any discussion about self driving cars: the answer that nobody will directly give (to my memory only the head or research for Mercedes gave an answer: the driver because it's the customer, and they quickly removed any trace of that interview). THIS is a major slippery slope that create all kind of legal and ethical mess: is the customer responsible of eventual damage caused by its car? Or the manufacturer? Neither? It's Glicol's Will and call it a day?

Heck considering that the vast majority of the algorithms used in those systems are simply black-box model trained ones and not even the developer will know how the car will respond to a given situation.

So yes, we may have the "technology itself" but its implementation is an heated political problem... Just to remind that there are few fields that have more political connotation than technology and its effects on humanity.

[Varyon]

Originally Posted by Opellulo

Let's clear a thing about autonomous driving vehicles: the road to their implementation is long (if ever will be one) not because of technological limits (and there are many: from being confused to rain, to a slightly redder moon to a discolored crossing) but for a specific philosophical/ethical/legal problem which is LITERALLY the trolley problem.

In the event of an inevitable collision, whose safety the autonomous vehicle will prioritize?

The resulting collision will not be an "accident" but the expected result of a decision embedded within the system, a decision that someone will need to be accounted for.

I hadn't considered this, honestly. So, a situation where braking is insufficient to avoid a collision, and swerving left or right will still result in a collision. This means an oncoming vehicle (one in the wrong lane, or one going in reverse), something unexpectedly coming in from the side (a vehicle running a red light, a kid running into the street, etc) or the car in front of you coming to a sudden stop due to it experiencing a collision (it suddenly braking should be insufficient - the self-driving car should keep enough distance between itself and the vehicle in front to be able to brake in such an instance; for humans, this is the "two-second rule"). There's also cases where your car slows/stops as appropriate, and the vehicle behind you fails to do so. In addition to the above, you must have nowhere to safely go - boxed in by cars on either side on the highway, an occupied sidewalk to one side and cars coming from the opposite direction on the other, etc.

Having all vehicles be self-driving would get rid of the bulk of the above issues, leaving only the problem of a person (or animal, etc) suddenly running out in front of you. Humans can avoid most such collisions via situational awareness, so at least in theory a properly-designed AI should be able to do the same (possibly with greater effect, given better reaction time). If a collision is well and truly unavoidable... make the human decide. Unless all vehicles are self-driving (and possibly even then), actions by the person in the driver seat (turning the wheel, pumping the brakes, etc) should serve as a manual override. The vehicle should prep any crash mitigation systems (lock seatbelts, ready airbags, send a burst communication to emergency services, arm fire suppression systems, etc), go to max deceleration, and flash some sort of "collision unavoidable" warning. If the human reacts in time, he/she can override the computer's decision, swerving to one side or the other, hitting the gas to accelerate, etc. If the human fails to react in time, that means he/she likely would have failed to react in time if he/she were the one driving, and at least the vehicle slows down some prior to the collision. At worst, you're looking at the same result as if the human were the one driving in the first place (at least in theory - in practice, a person driving is probably going to be paying better attention than one riding along, all else being equal). The car's computer isn't causing a collision... it's just failing to avoid an unavoidable one (or, rather, failing to avoid one collision by not causing a different one).

[Ulzgoroth]

Originally Posted by Varyon

I hadn't considered this, honestly. So, a situation where braking is insufficient to avoid a collision, and swerving left or right will still result in a collision. This means an oncoming vehicle (one in the wrong lane, or one going in reverse), something unexpectedly coming in from the side (a vehicle running a red light, a kid running into the street, etc) or the car in front of you coming to a sudden stop due to it experiencing a collision (it suddenly braking should be insufficient - the self-driving car should keep enough distance between itself and the vehicle in front to be able to brake in such an instance; for humans, this is the "two-second rule"). There's also cases where your car slows/stops as appropriate, and the vehicle behind you fails to do so. In addition to the above, you must have nowhere to safely go - boxed in by cars on either side on the highway, an occupied sidewalk to one side and cars coming from the opposite direction on the other, etc.

Having all vehicles be self-driving would get rid of the bulk of the above issues, leaving only the problem of a person (or animal, etc) suddenly running out in front of you. Humans can avoid most such collisions via situational awareness, so at least in theory a properly-designed AI should be able to do the same (possibly with greater effect, given better reaction time). If a collision is well and truly unavoidable... make the human decide. Unless all vehicles are self-driving (and possibly even then), actions by the person in the driver seat (turning the wheel, pumping the brakes, etc) should serve as a manual override. The vehicle should prep any crash mitigation systems (lock seatbelts, ready airbags, send a burst communication to emergency services, arm fire suppression systems, etc), go to max deceleration, and flash some sort of "collision unavoidable" warning. If the human reacts in time, he/she can override the computer's decision, swerving to one side or the other, hitting the gas to accelerate, etc. If the human fails to react in time, that means he/she likely would have failed to react in time if he/she were the one driving, and at least the vehicle slows down some prior to the collision. At worst, you're looking at the same result as if the human were the one driving in the first place (at least in theory - in practice, a person driving is probably going to be paying better attention than one riding along, all else being equal). The car's computer isn't causing a collision... it's just failing to avoid an unavoidable one (or, rather, failing to avoid one collision by not causing a different one).

It might be possible to fool the law into accepting that, but it would be a pretty bad argument. They've already tried the 'automatic drive but a human sits there to take over in emergencies' and what I've heard is that there's no sign that it's got any redeeming qualities. The 'driver' who has not been driving doesn't have appropriate reactions for being yanked from full disengagement to full emergency with no warning, (even compared to normal human emergency reaction standards) because how could they possibly be expected to?

[Rupert]

Originally Posted by Ulzgoroth

It might be possible to fool the law into accepting that, but it would be a pretty bad argument. They've already tried the 'automatic drive but a human sits there to take over in emergencies' and what I've heard is that there's no sign that it's got any redeeming qualities. The 'driver' who has not been driving doesn't have appropriate reactions for being yanked from full disengagement to full emergency with no warning, (even compared to normal human emergency reaction standards) because how could they possibly be expected to?

This also defeats the purpose of a self-driving vehicle. If the human 'overseer' has to maintain alertness and situational awareness at the same levels as they would if actually driving they'll find it at least as fatiguing as driving, and will lose concentration more often (because there's nothing really holding their attention on the task), and can't do anything else whilst overseeing the computer, so there's no productivity gain or leisure time gain.

Also consider that aeroplanes can automatically take-off, fly a route, and land, and this is, compared to driving on crowded roads, a very simple task and when emergencies happen the correct response is usually fairly straightforward and obvious, yet the supervising humans find it hard to rapidly take over and manage such emergencies, and they are highly trained individuals and the emergencies are almost always during times when the pilots are at their most vigilant (take-off and landing).

I have little confidence in people's ability to do better in cars so until the computer controls can manage everything, and we have rules in place to reasonably assign liability (so the manufacturers aren't constantly sued when their vehicle did as well as possible in a 'no-win' emergency, but can be sued or charged if their system messes up), I'll not be expecting self-driving vehicles on our roads. OR, if they are, it'll be a very expensive circus of litigation, etc.