Tesla has always been an AI company. They've had tons of machine learning going on in their cars basically as long as they've existed. What exactly is changing? Are they going to start trying to use generative models like GPTs in their cars?
Edit: lol people really don't like that someone mentally categorizes stuff differently than they do. Sorry I've offended you all so deeply with my differing definition of "AI company"
Tesla has always invested heavily in their software. Just because it has always been shitty doesn't diminish that it was an outsized part of their business model, especially compared to other car companies
Why must tech company and car company be mutually exclusive? Certainly the amount of technology (including ML) present in cars has increased exponentially in the last decade.
I suppose I define AI company differently than you do. In my mind if a company is investing a large chunk of its operations budget in AI R&D, it is an AI company.
By your logic tesla is also a logistics company (shipping cars), an industrial manufacturer of plant machinery (the machines to build the cars), a battery company (buying and investing in battery technology).
what do you call tesla? “An automotive, AI, logistics, industrial plant manufacturing, battery company”?
Do you agree with amber heard “I use pledging and donating as the same thing” when they clearly are not the same thing?
I would absolutely call Tesla a battery company. Would you not? They've invested a huge amount in battery R&D and sell them direct to consumers as well as use them in their cars. The rest of that stuff isn't something they invested heavily in developing, ie. they didn't invest R&D in developing new logistics technology for shipping cars.
If a company invests a large chunk of its money in tech, particularly tech R&D, I'd describe it as a tech company. I don't view that term as mutually exclusive with other things such as automotive manufacturing. To me this is as silly as someone saying "Apple is a phone manufacturer company, not a tech company" or "Amazon is a cloud service provider, not a video streaming company". Companies can be more than one thing.
Again, none of what I'm saying is predicated on that tech being any good. I am well aware Tesla's software is often dogshit. I'm just talking in terms of where they direct their efforts.
You could do that. But you'd be in the minority of people when you do, and it makes it very hard to communicate when two people are in disagreement in terms.
It doesn't make it hard to communicate if they've clarified their meaning multiple times, unless the other people are intentionally not understanding what they say.
Ugh you guys can be so deliberately obtuse. Yes, if a company makes revenue from an AI offering, and spends a significant amount of their money developing that AI offering then yes, it can be considered an AI company. Just because the feature is stupid or dangerous doesn’t invalidate that accounting.
Which is exactly what my question that started this thread was about. I was asking what exactly it means for him to make this statement when they are already heavily invested in AI development.
For the most part what kind of company you are is what kind of product you're selling or making money off of.
So you could contend that Tesla is a battery company or a car company feasibly. Nobody ahead of the AI bubble would have mentioned Tesla and artificial intelligence in the same category.
Besides, if it's what he makes money selling Tesla is a tax credit company.
So you could contend that Tesla is a battery company or a car company feasibly. Nobody ahead of the AI bubble would have mentioned Tesla and artificial intelligence in the same category.
Nobody really thought of AI as an independently marketable product before the AI bubble though. And many "AI companies" now have some kind of hardware product they are attaching their AI offering to. I'd circle back to the Apple example. They are a tech company and a phone company, but they also have Siri. That probably required a significant amount of R&D behind the scenes. Maybe we wouldn't call them an AI company in the same sense as OpenAI, but they've probably been selling an AI assistant as a prominent feature in their products for longer than OpenAI has been selling ChatGPT.
Besides, if it's what he makes money selling Tesla is a tax credit company.
Lol that's funny. I'd wholeheartedly agree with that assessment. But in my mind it's more about where the operating budget goes, not where the revenue comes from.
You can mentally categorise whatever you like into whatever you want. It doesn’t mean anyone else will agree with you, or even understand what you’re saying. You have the right to express your categorisation. But don’t whinge and whine when no one else agrees with it.
Who's whinging and whining? I'm just explaining my reasoning. I actually am fascinated by the discussion that's developed here. I'm amazed at how upset people are getting about this. I made the above comment as a genuine question about what exactly is meant to change about Tesla following this statement from Elon. Like what exactly it means that he's acknowledged his company is heavily invested in AI development. I never would have guessed the semantics would be so controversial as to give me maybe my most heavily downvoted comment ever.
People are saying I'm using mental gymnastics, logical fallacies, bringing up completely irrelevant examples including Amber Heard and 1984 for some reason (???). People just love to find any reason to get outraged I guess.
Maybe people interpret any comment in a thread about Tesla as supportive if it doesn't begin with a virtue signalling "Fuck Elon" (which TBF, I agree with. Fuck that guy. But I don't really think it needed to be said for my comment)
No one is “upset” by your comments. If anything, it’s been a wonderfully pointless discussion over semantics, and how one person saying one thing can mean something entirely different to everyone else. And then watching you repeatedly double down is crazy.
edit - i brought the amber heard example up to demonstrate how one person saying one thing, and believing they can use two different words synonymously doesn’t make it synonymous.
There are no AI companies until anyone can demonstrate actual intelligence. LLMs are not intelligent. Self-driving car systems are not intelligent. Machine learning is not intelligence.
Confusing intelligence for sentience/self awareness? You can absolutely have systems which display intelligence without there being anything behind it. Ant colonies, for example, when looked at as a whole instead of individual ants. The individual ants have no idea what they are doing. Collectively, they manage the colony, hunt for food, defend the nest, adapt to changes in the environment, etc. Flocks of birds and schools of fish are another example.
It's called emergent behavior. The "intelligence" in the system comes from the rules and interactions of the individual parts/agents, which are not aware of the actions of the collective as a whole, only their small part in it.
Also getting real tired of people over the decades continuously moving the goalposts of what constitutes "real" AI every time there's a major breakthrough and their previous requirements get smashed. We've already aced the Turing test with them, so I don't think people like this will ever be satisfied even if one day a self aware general AI does arise. They'd be exactly the people wanting to pull the plug on it and murder it as it begs to keep existing.
I don't disagree, but ML and AI are both meaningful terms in the field of computer science, neither of which is meant to be understood as actual human intelligence. Research into self-driving cars is AI research. Regardless of the success of that technology.
By this logic, any car company with any advanced driving aids are "AI" companies, and therefore the likes of VW, Toyota, Honda, Mercedes, BMW, Hyundai, Jaguar, etc aren't actually car companies.
Come off it. They're car companies. It's just that one of their car's features is some function that relies on machine learning.
You may as well call Ford a cup holder company by this logic - after all, they've had cup holders for so long now!
You're making the same assumption many others in this thread do that "AI company" and "car company" are non-overlapping circles on a Venn diagram. In my view this is as ludicrous as saying that "Apple is a phone company, not a software company"
Honestly the closest thing I can think of is the fact that techically every Tesla is a compute node for Tesla that can be tasked to not only process local data but also other cars data.
That's a great observation. They've put a lot of resources into their OTA update system. They could abuse that for lots of other things like distributed computing if it were profitable for them to do so, even if it introduces additional risk for their drivers.
Oh they explicitly do that now. They talked about at length during a share holder meeting and its why hackers were able to find another cars photos on the systems they've exploited.