Apple Inc. is canceling a decade-long effort to build an electric car, according to people with knowledge of the matter, abandoning one of the most ambitious projects in the history of the company.
Well yeah they were achieving absolutely nothing in self-driving vehicles so I suppose they can transfer those skills to achieve absolutely nothing in AI
If you’re only just now entering the AI market you’re not going anywhere
Not sure if 100% a joke or just partly but Volkswagen was probably providing the mechanical engineers who were retrofitting Lexus vehicles of the project as they had for the already-deployed driverless vans.
See NYT & MacReports
Stop I’m unironically excited for this distant possibility.
I’m already impressed by how much the UI for cars have been separatedfromm mechanical systems.
Electric cars have a lot more tunability from a software view too. Clearly there is a plenty of real world between the chips and where the rubber meets the road too.
No one can become a skilled ML/AI dev overnight. That will still take a year or two or more of working with it daily. If you transition to a new field you basically become a junior dev all over again for a while. Domain knowledge is a big part of being a good programmer.
Except news always was that Apple was pushing pretty hard into self-driving vehicles, which would use much the same AI learning systems as you need for generative AI.
Chances are very good the AI technologies they were working on involved developing a GAN, then the knowledge and experience of creating a GAN is fully transferable.
It’s the skills of the developers I’m talking about transferring — not the source code or neural net output.
Once upon a time, stoves had a dial you set, and it was basically a resistor and some wires. Today, a stove has a computer built in it that operates the entire thing.
While the computer in a modern oven is simple - it is an illustration that more, and more of what we have is computerized. When you add in reinforcement learning algorithms to adjust factors like say, If the fridge is aware of what time you generally open the fridge it can opt to kick on the heat pump a little before that to bring the temperature down and avoid running while it is open. This could save pennies of electricity in a year. But more importantly - could lead to less duty cycles on the condesor that could cause a fridge to say instead of lasting 10 years, last 12 years.
If you are starting a car company today, what you have to be thinking about is a reality where we move to “Humans don’t drive, the cars drive you” - I mean even a manual control situation could have the AI actually being a watcher in effect we “Let” people drive, but if the AI detects an unobserved obstacle etc it immediately takes over and adjusts. Well: You need to build that - and that, is AI.
If a company isn’t thinking about AI, and makes anything but basic appliances - they are likely on a limited time window because at some point Autonomous cars WILL be good enough, and the safety consideration will make both people, and governments, along with insurance companies to eliminate human driven vehicles.
Apple isn’t looking next year, or a year after. They are looking 5 to 10 years out and they don’t see a path where they can effectively compete in the car industry and make the profits they are after. However, if they can solve the AI driving problem - they don’t NEED to make a car, they can sell the brains and system that drives the car.
Cleaning the coils would make the condensers last longer than the fraction of a percent of duty cycle that cooler ambient air during the time the doors open would.
Actual good compressors would be even better.
Adding computers to everything is a marketing gimmick. We already know how to make good stuff that lasts a long time, it’s just too expensive for underpaid customers to afford a 25% price premium on quality appliances so they slap a ten dollar iot kit on what would have been bottom shelf junk thirty years ago and call it a day.
If your car development team can be transferred to AI developement you weren’t building much of a car.
This is referring to the team working on the self driving functionality.
Still, self driving and generative AI are very different. Just because they fall into the same big “AI” bucket doesn’t mean it’s the same.
The skills are similar enough to be transferable
We’ll see
Well yeah they were achieving absolutely nothing in self-driving vehicles so I suppose they can transfer those skills to achieve absolutely nothing in AI
If you’re only just now entering the AI market you’re not going anywhere
Not sure if 100% a joke or just partly but Volkswagen was probably providing the mechanical engineers who were retrofitting Lexus vehicles of the project as they had for the already-deployed driverless vans. See NYT & MacReports
Skilled developers can easily transition to another field of software engineering.
Even in 2024, there’s still a lot of non-software (automotive) engineering involved in building a car – even an electric one.
SDA - software defined automobile
Highway infrastructure as code
Stop I’m unironically excited for this distant possibility.
I’m already impressed by how much the UI for cars have been separatedfromm mechanical systems.
Electric cars have a lot more tunability from a software view too. Clearly there is a plenty of real world between the chips and where the rubber meets the road too.
So I get a free car and the owner gets to keep the original? Like hell I wouldn’t.
For real!
Yeah but a car is mostly made of engines and bolts and wheels and stuff like that, you know.
No one can become a skilled ML/AI dev overnight. That will still take a year or two or more of working with it daily. If you transition to a new field you basically become a junior dev all over again for a while. Domain knowledge is a big part of being a good programmer.
Except news always was that Apple was pushing pretty hard into self-driving vehicles, which would use much the same AI learning systems as you need for generative AI.
Ehh, that’s a bit of a stretch. They’re very different technologies with only limited overlap.
Unless Apple had invented artificial general intelligence and not told anyone then the technology isn’t transferable.
Chances are very good the AI technologies they were working on involved developing a GAN, then the knowledge and experience of creating a GAN is fully transferable.
It’s the skills of the developers I’m talking about transferring — not the source code or neural net output.
One question for you smarty pants: is their car real, or artificial?
Well, at this point it is theoretical.
Once upon a time, stoves had a dial you set, and it was basically a resistor and some wires. Today, a stove has a computer built in it that operates the entire thing.
While the computer in a modern oven is simple - it is an illustration that more, and more of what we have is computerized. When you add in reinforcement learning algorithms to adjust factors like say, If the fridge is aware of what time you generally open the fridge it can opt to kick on the heat pump a little before that to bring the temperature down and avoid running while it is open. This could save pennies of electricity in a year. But more importantly - could lead to less duty cycles on the condesor that could cause a fridge to say instead of lasting 10 years, last 12 years.
If you are starting a car company today, what you have to be thinking about is a reality where we move to “Humans don’t drive, the cars drive you” - I mean even a manual control situation could have the AI actually being a watcher in effect we “Let” people drive, but if the AI detects an unobserved obstacle etc it immediately takes over and adjusts. Well: You need to build that - and that, is AI.
If a company isn’t thinking about AI, and makes anything but basic appliances - they are likely on a limited time window because at some point Autonomous cars WILL be good enough, and the safety consideration will make both people, and governments, along with insurance companies to eliminate human driven vehicles.
Apple isn’t looking next year, or a year after. They are looking 5 to 10 years out and they don’t see a path where they can effectively compete in the car industry and make the profits they are after. However, if they can solve the AI driving problem - they don’t NEED to make a car, they can sell the brains and system that drives the car.
Cleaning the coils would make the condensers last longer than the fraction of a percent of duty cycle that cooler ambient air during the time the doors open would.
Actual good compressors would be even better.
Adding computers to everything is a marketing gimmick. We already know how to make good stuff that lasts a long time, it’s just too expensive for underpaid customers to afford a 25% price premium on quality appliances so they slap a ten dollar iot kit on what would have been bottom shelf junk thirty years ago and call it a day.