-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Autonomous Car Progress
- Thread starter stopcrazypp
- Start date
-
- Tags
- Autonomous Vehicles
scottf200
Well-Known Member
That is pretty cool although it is unclear if that is just when they would use it. Just to verify some specific scene / scenario that caused a problem. It is like getting a view inside the ML/AI black box in a sense.
Unveiling LINGO-1's Show and Tell Capability with Referential Segmentation - Wayve
We have developed a new "show and tell" feature in LINGO-1, which uses referential segmentation to visually highlight the focus of LINGO-1's attention and enhances the connection between language and vision tasks, ultimately allowing for more accurate and trustworthy responses.
wayve.ai
diplomat33
Average guy who loves autonomous vehicles
We know the black box nature of end-to-end is one of the main drawbacks of that approach. I think LINGO-1 seeks to address this issue. And the ability for the AI to actually show what it is thinking will help even further. So now, Wayve can use this to see what their AI Driver is thinking and why and that will help them troubleshoot and validate their autonomous driving. I do think making sure LINGO is accurate will be an issue. If LINGO makes a mistake, they will need to fix that. So LINGO does not automatically mean their autonomous driving is reliable, it is just a tool that will help them validate their system. But it is definitely an impressive tool IMO.
That's possbile. Or it's just another clever marketing strategy. "Hey look our driver is very smart, please more capital".
Doggydogworld
Active Member
Yeah, I saw that. But it includes 10k+ square miles of almost completely uninhabited desert and mountains. I excluded all that as irrelevant to a Robotaxi service.It's even more than that. A quick Google search says it is 14,600 sq miles.
LNGO-1 looks kinda cool, thanks for posting it.
diplomat33
Average guy who loves autonomous vehicles
Wayve shows some nice driving examples. It is nice to see their car be able to handle some city driving cases with end-to-end. But I am curious what their overall intervention rate is. To be able to actually deploy driverless at scale, you have to be more reliable than just a few nice "demos".
Those edge cases can be challenging. Got to watch out for those Goats in the Bay Area
Good for the board. It's getting scary so many AI/AV leaders are turning out be be so untrustworthy (Altman, Vogt, Musk, Milton, ..).
If only Tesla had a real BoD...
scottf200
Well-Known Member
Yes, lots of moving parts. Also a few big reasons to stay with OpenAI like stock options and an eventual IPO. Even offers to match equity with MSFT or CRM stock wouldn't come close.
Huh? Those texts end with the summary that Microsoft has obtained OpenAI's CEO and most employees for essentially no cost. If Microsoft has them, why would they continue to fund the company? Almost 90% of OpenAI's funding comes from Microsoft. That dries up, OpenAI dries up, and the stock is worthless.
What am I missing?
EVNow
Well-Known Member
I read the same thing but apparently it wasn't all true. Per MSFT's CEO, over the weekend he encouraged the OpenAI board to rehire them. But that would only solve part of the problem as OpenAI employees want the board of directors replaced. So MSFT is now planning to talk to the board in the hopes of a final solution.
EVNow
Well-Known Member
Is it delayed or has it come out ?
I wonder how the Cruise fiasco will affect Ultra Cruise ....
My understanding is they have very little to do with each other, with minimal if any collaboration, which may be a good thing given the recent turn of events.
Similar threads
- Article
