GM bolts from Cruise

After plowing 10 billion into self-driving cars, GM calls it quits.

At least the program yielded some pretty good improvements to collision avoidance systems, lane keeping, and other driver assistance features.

True self driving won't happen until every car and truck on the road is linked up and communicating with all of the others. It will work great until one car's AI wigs out and commits suicide.

Kind like this flock of drones which appear to all be following one that broke and did a nose dive into the water.

MSN.com: Hundreds of drones malfunction and fall from sky at show in China

I drove several times in Champcar with a Cadillac engineer who was working on a system that would link GM cars.  He had a simple version that he would install on the non-GM car we were renting a seat in.

Tech companies can burn cash on things because they have super high profit margins on their stuff and their investors are used to higher risk projects.

Car companies can't burn cash on as many things because their profit margins are lower and their investors are more conservative and less willing to take those risks.

Will they be selling off the cars and the data to an investor company , 

Seems like most of the work has been done , just that  GM does not want the liability .

californiamilleghia said:

Will they be selling off the cars and the data to an investor company , 

Seems like most of the work has been done , just that  GM does not want the liability .

There's no reference point of what a finished product looks like, so we have no idea how much of the work has been done. It could still be an infinite amount of work required at this point. It's pretty clear that at this point in time, for better or for worse, that the system needs to be essentially flawless (ie basically zero failure/incident rate) for general public acceptance, so I'd guess there was a lot of work left. Plus the continuing maintenance/oversight/development that was going to make it a struggle to be profitable for probably quite a long time.

I feel like the tech for the self-driving car is like 85% there, but that last 15% is just a cast-iron bitch.

dps214 said:

It's pretty clear that at this point in time, for better or for worse, that the system needs to be essentially flawless (ie basically zero failure/incident rate) for general public acceptance

Oh, you mean like Tesla? /sarcasm

In reply to californiamilleghia :

Most of the mapping and tech was applicable to their SuperCruise safety suite in their normal consumer vehicles, so I'm guessing they'll keep a lot of the IP in house. Vehicles might be scrapped, or used for mapping new routes to keep Supercruise updated a la the "Google maps" cars out there that drive around.

Sinking large piles of cash into a project and then pulling the plug on it? 

I don't think GM has ever done anything like that!

I'm sure something could be created for interstate travel, involving electronically linked cars and such.

I want to see one operate in this.

AngryCorvair (Forum Supporter) said:

dps214 said:

It's pretty clear that at this point in time, for better or for worse, that the system needs to be essentially flawless (ie basically zero failure/incident rate) for general public acceptance

Oh, you mean like Tesla? /sarcasm

You mean their full self driving system which is still in development and its current form is heavily criticized with basically every incident? Yeah.

I thought I heard that tesla was getting into the robotaxi game. If there's anyone that can convince government agencies and the public (again, for better or for worse) that "average human" quality autonomous driving is acceptable, it's them. But I probably wouldn't hold my breath.