News Tesla Books TSMC to Produce 4nm SoCs for Cars: Report

[Admin]

TSMC might produce chips for Tesla's Full Self-Driving in Arizona.

Tesla Books TSMC to Produce 4nm SoCs for Cars: Report : Read more

 

[bit_user]

Perhaps the most surprising thing about the report is that Tesla decided to use one of TSMC's N4 (4nm-class) fabrication technologies instead of N5A (5nm-class) manufacturing process specifically developed with automobiles in mind.

Perhaps it's for their humanoid Optimus robots:

The Tesla Robot: Here’s What We Know

The prototype of the Tesla robot, Optimus, was unveiled at the company’s AI Day in September 2022. According to Tesla’s CEO, Elon Musk, Optimus will be mass produced and cost less than a car to purchase.

builtin.com

 

[Bazzy 505]

Perhaps it's for their humanoid Optimus robots:

The Tesla Robot: Here’s What We Know

The prototype of the Tesla robot, Optimus, was unveiled at the company’s AI Day in September 2022. According to Tesla’s CEO, Elon Musk, Optimus will be mass produced and cost less than a car to purchase.

builtin.com

who know with that snowflake Elon, 2 years is a long time for likes of him,
maybe is planning to transfer his mind into a "chappie bot" lol

 

[bit_user]

who know with that snowflake Elon, 2 years is a long time for likes of him,
maybe is planning to transfer his mind into a "chappie bot" lol

Yeah, maybe that's why he's cracking the whip, over at Neural Link:

Exclusive: Musk’s Neuralink faces federal probe, employee backlash over animal tests

Elon Musk’s Neuralink, a medical device company, is under federal investigation for potential animal-welfare violations amid internal staff complaints that its animal testing is being rushed, causing needless suffering and deaths.

www.reuters.com

He needs to get all the pieces in place to transfer his consciousness, before the dementia sets in (assuming it hasn't started, already).

 

[JohnMD1022]

Or, it could be Tesla phone related.

 

[bit_user]

Or, it could be Tesla phone related.

Hadn't heard about that. Makes even less sense to me than the robots.

 

[svan71]

Funny thing is I never once wished for a self-driving car. Is a self wiping ass next?

 

[Deleted member 14196]

Yeah, I think we can pretty much forget about self driving cars for quite a long time. They clearly are not ready for prime time and still infants in testing. And his tesla is doing relying solely on cameras and visual, I think that’s suicide. His refusal to use LiDAR is ridiculous.

 

[kjfatl]

Funny thing is I never once wished for a self-driving car. Is a self wiping ass next?

When I saw the car on the other side of the highway spin out in the rain, cross the median and and pass in front of my vehicle just before I t-boned it, I wished the other driver had a self driving vehicle. 140MPH in a 50MPH zone in the rain doesn't cut it. Fortunately we lived.

 

[kjfatl]

Yeah, I think we can pretty much forget about self driving cars for quite a long time. They clearly are not ready for prime time and still infants in testing. And his tesla is doing relying solely on cameras and visual, I think that’s suicide. His refusal to use LiDAR is ridiculous.

I tend to agree that LIDAR would make the system safer, but I don't know any human driver who has more than 2 eyes and 2 ears and we do OK. Only about 50K deaths per year. It won't be long until driver assistance systems cut that number in half.

 

[bit_user]

Funny thing is I never once wished for a self-driving car.

It did seem like the industry got ahead of demand, here. I had a similar attitude as you, until I thought about how self-driving could provide mobility to people who can't drive. Think of the elderly or people who have some disability that makes driving difficult or impossible. I'm still quite ambivalent towards it.

Other societal impacts:

• There's going to be more car-sharing and less ownership.
• Roads will be even more crowded, as parents send the car to shuttle around their kids on trips they, themselves weren't always willing to make.
• Driving should eventually become even safer than today.
• It will eventually become more costly and inconvenient to drive a manually-driven vehicle.

 

[bit_user]

When I saw the car on the other side of the highway spin out in the rain, cross the median and and pass in front of my vehicle just before I t-boned it, I wished the other driver had a self driving vehicle.

A lot of times, this is simply due to people driving with too little tread. Self-driving wouldn't necessarily solve that. Traction control does help with hydroplaning - it can keep you pointed in the right direction, but you may still go off the road.

140MPH in a 50MPH zone in the rain doesn't cut it.

You mean relative velocity or is that really how fast they were going? It's hard to go 140 in a street car, in spite of what the supposed top speed is. It's insane to do it in the rain. Was 140 some kind of estimate or is that what the data logger in their car said? If they had traction control disabled, maybe that's just how fast their wheels spun after breaking traction with the ground? What kind of car was it?

I tend to agree that LIDAR would make the system safer, but I don't know any human driver who has more than 2 eyes and 2 ears and we do OK.

You also have an accelerometer, in effect. Mirrors and a head that can move multiplies the effective number of "eyes". And your brain is still a lot more powerful than self-driving computers.

In spite of all of that, probably every couple thousand driven miles - if not more frequently - I'll see something and won't quite be able to figure out what it is for at least long enough to think about it. Self-driving cars definitely need more than just cameras - they need radar to know when something is actually there.

 

[kjfatl]

Or, it could be Tesla phone related.

It did seem like the industry got ahead of demand, here. I had a similar attitude as you, until I thought about how self-driving could provide mobility to people who can't drive. Think of the elderly or people who have some disability that makes driving difficult or impossible. I'm still quite ambivalent towards it.

Other societal impacts:

• There's going to be more car-sharing and less ownership.
• Roads will be even more crowded, as parents send the car to shuttle around their kids on trips they, themselves weren't always willing to make.
• Driving should eventually become even safer than today.
• It will eventually become more costly and inconvenient to drive a manually-driven vehicle.

It is clear that once the transition takes off, insurance rates for manually-driven vehicles will skyrocket when it becomes clear that drivers of these vehicles are responsible for most accidents. The automobile insurance industry will drastically changed since accident liability will be the responsibility of the vehicle manufacturers covered by vehicle usage fees.

 

[Deleted member 14196]

I’m not buying it. Teslas are too frequently involved in accidents to this day and the data is showing their self driving is getting worse. Don’t expect self driving cars for at least 100 years.

The national Highway and safety transportation wants to put an end to Teslas automatic car driving because it’s so problematic and nobody else has one that’s any better

 

[bit_user]

I’m not buying it. Teslas are too frequently involved in accidents to this day

But these are very early days, in the era of self-driving. The compute power they have and the sensors they're using are both at the bottom end of the scale.

the data is showing their self driving is getting worse. Don’t expect self driving cars for at least 100 years.

Is it actually getting worse, or are more people just using it more often? I'd be pretty surprized, if the rate of self-driving accidents was actually going up per mile driven.

The national Highway and safety transportation wants to put an end to Teslas automatic car driving because it’s so problematic

Yes, Tesla jumped the gun on this, and it was irresponsible of them. That's a Tesla/Musk issue, IMO.

nobody else has one that’s any better

Nobody else has rolled it out like Tesla did, that's for sure. Whether it's because they're worse or just more cautious... I rather think the latter.

 

[Deleted member 14196]

GM is having major issues with Cruise for instance. And there’s a study showing that Tesla auto driving’s involvement in accidents. I’ll try to find that and post it here. At the cost of that $15,000 for self driving I doubt that it’s being used by that many more people. You cannot assume that everybody buys that feature because the car is so damn expensive. An extra $15,000 slap in the face might stop somebody from wanting that feature. A lot of Youtubers who are testing them don’t like it and won’t pay for it.

in any case, I am taking a wait-and-see approach. I’m not jumping on the EV bandwagon anytime soon. And I don’t think I would ever opt for self driving because I don’t trust something that can’t understand and can’t think. Let alone something that doesn’t have radar and relies on cameras for vision.

for full automated self driving, you’re going to need intelligent roads that help cars communicate with each other as well as having radar or LiDAR, and other visual systems working together. Next to impossible in huge countries that have lots of roads to convert. I’m going to say it’s impossible, and I’ve seen nothing to indicate that it’s possible anytime in the near future. Hence my estimate for 100 years.

 

[Deleted member 14196]

And here’s another interesting fact MIT researchers, finding that driver distraction goes way up during automatic driving another reason that I hate it.

Too many stupid people out think that the systems are level five autonomous and they’re barely level two

Tesla Autopilot results in decreased driver attention, new study finds

A new study based on MIT Advanced Vehicle Technology data found that Tesla Autopilot results in a noticeable decrease in...

electrek.co

Investigation into Tesla crashes involving 'Autopilot' feature gets more serious | Fortune

A probe from the National Highway Traffic Safety Administration now covers 830,000 vehicles.

fortune.com

Tesla Autopilot Crashes: With at Least a Dozen Dead, “Who’s at Fault, Man or Machine?”

A self-driving Tesla recently killed two people in China, and with many other reports of autopilot failure, the scrutiny they face increases

impakter.com

already stated that this technology is in its infancy, but I don’t think they should be able to test it live. This should be highly regulated extremely highly regulated and tested for decades before it’s released on to the public. This is an absolute travesty.

 

[bit_user]

Too many stupid people out think that the systems are level five autonomous and they’re barely level two

That's why I think it was irresponsible for Tesla to launch theirs, without better reliability and more safeguards to prevent it being misused.

tested for decades before it’s released on to the public.

It should really only take about 1 year of testing, which should ensure it can be tested in all weather conditions. However, there should be some standard DoT testing regime that's required for self-driving features to be rolled out. Developing that testing regime will likely be a multi-year effort.

for full automated self driving, you’re going to need intelligent roads that help cars communicate with each other

I'm not convinced of that. Driving is non-trivial, but not that hard of a task. The addition of radar could easily make self-driving safer than manual.

Hence my estimate for 100 years.

I'd put it closer to 10. The amount of compute power in self-driving SoCs is increasing exponentially. The first couple of generations weren't fast enough, but I think we're getting there.

 

[Deleted member 14196]

Let’s hope you’re right.

 

[kjfatl]

Self driving will probably reach the majority of vehicles a piece at a time. Lane keeping assistance is an example. It won't be too hard to add a feature that prevents a vehicle from steering across the medial of a highway or driving the wrong way on an interstate highway. Another area of safety improvement is following distance. The vehicle can refuse to tailgate the vehicle in front of it unless communication is established between the vehicles so applying brakes to the first vehicle automatically applies brakes to the one following. Lots of small steps will get us to the goal of safer vehicles.