Really? Jobs differ, people differ. Existing AI (esp. the art generation stuff) can already do a lot of the things that people do. The fact that it can't do *everything* is irrelevant. Just to use FdB's post as an example, lots of artwork doesn't require an accurate representation of particular people (or any people).
For some people in some jobs, AI can already do what they do. If and when AI improves, it will do be able to take over more of what people do.
The problem is that we want systems to be MUCH SAFER than human drivers and to achieve that level driving in SNOW probably requires extra sensors, not just extra intelligence. The best option IтАЩve heard of for this is ground-penetrating radar. An MIT project showed you can use radar to recognize and map features UNDER the roadway which remain stable and recognizable under a blanket of snow. This makes it possible for auto-driving cars to drive in the exact same part of the road - down to the inch - at times when humans canтАЩt see the edges of the road at all. But cars need the extra sensors installed and you need to build up the database; each service needs somebody to drive all those roads in snowy places and record the new data.
IтАЩd expect cars with that specific capability - driving safely in snow - to show up for use by the general public SOMEWHERE in the next 5-10 years but if you want it EVERYWHERE youтАЩll be waiting a while.
As with the cars weтАЩve got now in Phoenix and SF itтАЩs really more down to what capabilities regulators ALLOW than what the technologists can provide.
The question of what you want the technology to accomplish is completely separate from what the technology can accomplish. Driving in snow or being able to handle East Coast drivers is a problem where, as far as I can tell, not a lot of progress has been made in the last decade.
Hope all you want but I am extremely skeptical that any manufacturer is going to try to sell a car that can only be driven in some parts of the country some of the time.
I don't even know what you mean by "being able to handle East Coast drivers". Why would you expect the same progress that finally allowed full autonomous driving in SF NOT to help with that?
(SF traffic can be just as congested as anywhere else. California drivers do TEND to be more polite but that's a difference in degree, not in kind - and you can't drive in SF without being able to handle SOME tough merges, narrow lanes and obnoxious drivers!)
People sell vehicles *all the time* that you can only use in some parts of the country. Segways are illegal in some towns, electric scooters illegal in others. Motorcycles or jet-skis are only practical in certain weather, road, or water conditions. Some mopeds and electric bikes can't go on the highway due to being underpowered to reach higher speeds. Let's not let the best be the enemy of the good!
A potential solution isn't really a solution until it's put into practice.
Merging onto the BQE is a well known issue for self driving cars because the East Coast is notorious for aggressive driving practices, to the point where said self driving cars give up even trying.
The problem with self driving cars that can't be used everywhere is the human machine interface. Who wants to be woken up from a deep slumber when your car starts screaming "I'm doing 75 on the highway and I can't handle it! You take over!!!"
If AI can't do what human beings can then how useful is it?
More practically how likely is it that it's going to take anybody's job?
Really? Jobs differ, people differ. Existing AI (esp. the art generation stuff) can already do a lot of the things that people do. The fact that it can't do *everything* is irrelevant. Just to use FdB's post as an example, lots of artwork doesn't require an accurate representation of particular people (or any people).
For some people in some jobs, AI can already do what they do. If and when AI improves, it will do be able to take over more of what people do.
That is not necessarily a bad thing - I've written about this: http://mugwumpery.com/?p=703
So when will the new AI advances result in completely autonomous self driving cars?
"Prediction is difficult, especially about the future" --N. Bohr
"If AI can't do what human beings can then how useful is it?
"More practically how likely is it that it's going to take anybody's job?" -- me, just now
1) Clearly the people who are at risk of losing their jobs aren't driving for Uber or Lyft.
2) My prediction is that LLM's will be completely useless (orthogonal to the problem space) for autonomous vehicles.
The problem is that we want systems to be MUCH SAFER than human drivers and to achieve that level driving in SNOW probably requires extra sensors, not just extra intelligence. The best option IтАЩve heard of for this is ground-penetrating radar. An MIT project showed you can use radar to recognize and map features UNDER the roadway which remain stable and recognizable under a blanket of snow. This makes it possible for auto-driving cars to drive in the exact same part of the road - down to the inch - at times when humans canтАЩt see the edges of the road at all. But cars need the extra sensors installed and you need to build up the database; each service needs somebody to drive all those roads in snowy places and record the new data.
IтАЩd expect cars with that specific capability - driving safely in snow - to show up for use by the general public SOMEWHERE in the next 5-10 years but if you want it EVERYWHERE youтАЩll be waiting a while.
As with the cars weтАЩve got now in Phoenix and SF itтАЩs really more down to what capabilities regulators ALLOW than what the technologists can provide.
The question of what you want the technology to accomplish is completely separate from what the technology can accomplish. Driving in snow or being able to handle East Coast drivers is a problem where, as far as I can tell, not a lot of progress has been made in the last decade.
Hope all you want but I am extremely skeptical that any manufacturer is going to try to sell a car that can only be driven in some parts of the country some of the time.
The tech I referred to just above would seem to constitute significant progress for driving in snow. Um, here:
https://www.caranddriver.com/news/a31098296/autonomous-cars-snow-mit-researchers-solution/
I don't even know what you mean by "being able to handle East Coast drivers". Why would you expect the same progress that finally allowed full autonomous driving in SF NOT to help with that?
(SF traffic can be just as congested as anywhere else. California drivers do TEND to be more polite but that's a difference in degree, not in kind - and you can't drive in SF without being able to handle SOME tough merges, narrow lanes and obnoxious drivers!)
People sell vehicles *all the time* that you can only use in some parts of the country. Segways are illegal in some towns, electric scooters illegal in others. Motorcycles or jet-skis are only practical in certain weather, road, or water conditions. Some mopeds and electric bikes can't go on the highway due to being underpowered to reach higher speeds. Let's not let the best be the enemy of the good!
A potential solution isn't really a solution until it's put into practice.
Merging onto the BQE is a well known issue for self driving cars because the East Coast is notorious for aggressive driving practices, to the point where said self driving cars give up even trying.
The problem with self driving cars that can't be used everywhere is the human machine interface. Who wants to be woken up from a deep slumber when your car starts screaming "I'm doing 75 on the highway and I can't handle it! You take over!!!"