News for hackers. Stuff that matters.

A comment on Tesla’s Dojo Microarchitecture. Parent comment

Re:Tesla’s Dojo Microarchitecture by seunosewa on Monday September 05, 2022 @10:53AM (#32724971)
The self-driving AI they use this chip for appears to be less than impressive from my perspective. Its mistakes seem to come up in the news quite frequently.
[ Parent | on HN ]
Re:Tesla’s Dojo Microarchitecture by johnsimer on Monday September 05, 2022 @02:08PM (#32727265)
Dojo’s D1 chip for training AFAIK is not being used yet. Tesla still trains with their ~7000 cluster of A100s

And the D1 is for their training machine anyways. Not their inference chip in the cars

[ Parent | on HN ]
Re:Tesla’s Dojo Microarchitecture by jillesvangurp on Monday September 05, 2022 @10:58PM (#32732233)
Are you aware of a better implementation though? It might not be perfect; it's still miles ahead of competing systems. Yes, it makes mistakes. But that's a rather poor argument to dismiss the whole strategy and technology.

From my point of view, Tesla is converging on good enough more rapidly than anyone else in the industry. Courtesy of their AI strategy and technology; including the in house developed chips.

[ Parent | on HN ]
Re:Tesla’s Dojo Microarchitecture by friendzis on Tuesday September 06, 2022 @02:11AM (#32733532)
> Tesla is converging on good enough

This is a big problem, though. Any ADAS is a safety critical system and while "good enough" can help push volumes, it is extremely bad from safety perspective. If you really want to read on this u/adamjosephcook/ has some awesome writing going into far deeper detail than I could explaining this particular issue.

[ Parent | on HN ]
Re:Tesla’s Dojo Microarchitecture by mavhc on Tuesday September 13, 2022 @10:44AM (#32824933)
In terms of most lives saved, I'd bet Tesla's AI is winning, all those people they automatically stopped from driving into rivers etc
[ Parent | on HN ]