Connected/Automated vehicles and infrastructure in Boston

Interesting video of Tesla FSD in San Diego.


2:40 runs red light
3:22 runs over bollards
6:25 tries to drive on rail line
7:45 bollards again

he does this same loop every month, so another example of the AI failing to learn. The tracks arent moving, but it still doesnt understand.
The video above is among those confirmed and reported by the Washington Post. In the WaPo article, other drivers do report that their car is learning from their driving:
Some drivers said they have run their own experiments to test and improve the software. Kevin Smith, who uses FSD on his Tesla Model Y in Murfreesboro, Tenn., said he identified 13 locations near his hometown that stumped his car and created a route that hit all of them. “Each time, it gets a little bit better,” he said.

Full story in WaPo may require a subscription, but has the advantage that they sampled many youtube videos of crazy FSD, talked to the video's makers, and also had experts react frame by frame:

‘Full Self-Driving’ clips show Teslas on train tracks and fighting for control. Experts see deep flaws.
The Washington Post verified Full Self-Driving footage posted by beta testers and had it reviewed by a panel of experts
 
Computers that can do the heavy inference work that benefits from very large training data sets are server farm sized, not car sized. Moore's law might change this someday, but you can't cheat physics with code, no matter how much machine learning you can leverage.
Breaking out from Moore’s Law (or more specifically Dennard Scaling) will change this some day:
shameless plug for my company
 
Tesla gives driverless cars a bad name. There are many ways to really do what Tesla is saying it can do, with tech that Lord Musk has refused. LiDAR is one. Our problem is he has become something of a cult leader when he is in reality just another bro who made some good business decisions demanding fealty. Wall Street cannot question its golden calf. His blind disciples are no better than Appalachian snake dancers, but Muskovites dance with 3000 lb snakes on our roads. Someuns gonna git bit.
Good batteries in a cool car? Sure. Safe driving system? Sometimes… and that’s not good enough.

Yes, rockets and solar yaaaay. He’s still just a lucky Aspergian dreamer with a work ethic.
Don’t genuflect.
And don’t use his toddler-level self-drive.

Related side note: how many downtown Boston self driving videos are there? Why are the company test areas in optimal California neighborhoods with post-war street builds? Why is snow never factored in demo drives?
When the robo-car testers are getting perfect results in Lowell back alleys after a January Nor’easter, I’ll buy in. Until then, it’s all pump and dump wall street hype.
 
Our community rocks - - thank you, @fattony!


"...... In the past, compute performance improvements guaranteed by technology scaling, enabled incremental improvements of the computational capacity of transistor-based computers. However, a slowing of the rate of improvement based on technology scaling is creating a widening gap between compute performance and system need. This means that deploying advanced algorithms, such as deep learning in AVs, requires a new paradigm in computer design that exploits not only innovative architectures but also radically novel physics.

The team’s hybrid electro-photonic approach has been motivated by the development of photonic chips that compute using photons, not electrons, at speeds on the order of tera operations per second, while consuming much less energy. A critical competitive feature of the team’s approach is the ability to fabricate photonic chip based compute systems using standard semiconductor fabrication and OSAT processes within existing manufacturing facilities....... "

Open that door, let's go! (y)

1644593700548.png
 
Last edited:
Our community rocks - - thank you, @fattony!


"...... In the past, compute performance improvements guaranteed by technology scaling, enabled incremental improvements of the computational capacity of transistor-based computers. However, a slowing of the rate of improvement based on technology scaling is creating a widening gap between compute performance and system need. This means that deploying advanced algorithms, such as deep learning in AVs, requires a new paradigm in computer design that exploits not only innovative architectures but also radically novel physics.

The team’s hybrid electro-photonic approach has been motivated by the development of photonic chips that compute using photons, not electrons, at speeds on the order of tera operations per second, while consuming much less energy. A critical competitive feature of the team’s approach is the ability to fabricate photonic chip based compute systems using standard semiconductor fabrication and OSAT processes within existing manufacturing facilities....... "

Open that door, let's go! (y)

View attachment 21306
Hell yes! And thanks for posting. I also Googled Lightmatter after your comment. Insanely cool. Represent, Boston!
Now let's build enough housing to keep the genius here instead of letting Palo Alto CONSTANTLY pick our pocket!
 
Imagine the economic growth and human activity that will be unleashed around 2035 when the city might start banning individually owned and stored (on sides of roads) cars. The on-demand mobility should be available then.

That very street will look so different and alive with HUMANS (along with more restaurants, theatres, apartments, stores, etc.).

Moving this here, but the thing I can't follow with this is: everybody that is currently out driving their car right now, filling streets and creating traffic, if they aren't going to switch to public transit, are going to still need to occupy a significant amount of street space, whether its in their personal vehicle they own, or a rideshare vehicle. The demand will be the same, so the amount of traffic will be the same.

While they have the potential to eliminate parking in areas, there will still be, as anti-car puts it, "chaos," perhaps a lot more of it, as self-driving vehicles will pull in and out of spots quite frequently if they are making these kinds of trips for riders, no?

If we can solve AV's to a point where nobody has to worry about them in the same space as pedestrians and other car/service drivers, I think where they can help best in cities is as a 4-6-8-person pod as an extension of public transit. The T stations directly outside of Downtown Boston, and each Regional Rail Stop should be individual network hubs for on-demand rideshare AV's for the last-mile inefficiencies public transit can't solve. Incentivize trips to/from transit stations, and price point-to-point travel across cities according to demand, much like a demand-based tolling system on a highway. You can also reallocate your local busses to serve longer/further reaching communities.
 
Moving this here, but the thing I can't follow with this is: everybody that is currently out driving their car right now, filling streets and creating traffic, if they aren't going to switch to public transit, are going to still need to occupy a significant amount of street space, whether its in their personal vehicle they own, or a rideshare vehicle. The demand will be the same, so the amount of traffic will be the same.

While they have the potential to eliminate parking in areas, there will still be, as anti-car puts it, "chaos," perhaps a lot more of it, as self-driving vehicles will pull in and out of spots quite frequently if they are making these kinds of trips for riders, no?

If we can solve AV's to a point where nobody has to worry about them in the same space as pedestrians and other car/service drivers, I think where they can help best in cities is as a 4-6-8-person pod as an extension of public transit. The T stations directly outside of Downtown Boston, and each Regional Rail Stop should be individual network hubs for on-demand rideshare AV's for the last-mile inefficiencies public transit can't solve. Incentivize trips to/from transit stations, and price point-to-point travel across cities according to demand, much like a demand-based tolling system on a highway. You can also reallocate your local busses to serve longer/further reaching communities.

1) They won't be able to bring a 'vehicle they own' in to the city.

2) The AI pods operating within urban centers would be made out of heavy duty styrofoam (or another lightwight composite like it) and would be limited to no more than 20 mph..

3) Rideshare or better- mobility on demand (by 2035 kinks worked out on AI pods) means no need for storing these things on sides of streets. The "chaos" isn't caused so much by droppign off/picking ups - - it's caused by the ubiquitous space taken up by the sleeping cars on the sides......this would be more evident to the readers in this thread if you also moved the video along with my post. In that video, no more than 10% of the space on the sides (or middle) of those streets are being used for active pickup/dropoff.

4) There WILL still be "chaos" - (God bless chaos!!!) - but it will be much less harmful chaos - - more human dynamism and density and life - - not rusting hulls of vehicles at rest taking up space and choking the urban street like plaque in an artery.
 
Last edited:

Back
Top