Video

Martin Christopher Miller Helen Martinez Lisa

You're so myopic Over on your next Chinese Chinese supreme leader. You reimagined it to be elar mask eriod all players involved are making progress in ai chips. You only focus on Tesla. Everybody is making privacy ai Chips not only Tesla. This is why you completely missed NVidia till recently!

Well, we can pretty much say that you were wrong. China BEV share in 1H/2024 was 26%; EU 12,5%. The transition is going significantly slower than anticipated. Let's see what the upcoming 2-3 years will bring.

I can’t listen to him… wow

Brian, the treasure.

Tesla forgets that networking stack is equally important.

Can you please clarify. Are you saying that DoJO is 10x LESS power efficient compared to the H100? Sorry, but I got confused by the FP16 / FP8 and just wanted to make sure I got this right. Is Tesla so far behind?

Sounds like fpga asic if possible

DoJo has the same performance as the Intel 386 CPU!😂

Have you tried the new Cerebras chip? I am getting 1800 t/s on the free tier. Also, did you see this video explaining Metaculus? Scaling Superforecasting : AI Forecasting Tournaments & Road to Epistemic Security, with Deger Turan czcams.com/video/mpz1cuR4opQ/video.html Your website is a great resource or this kind of task. Would you be interested in working together in one of the upcoming tournaments? Finally, please tell us more about the Sohu chip. Who designed it and who is manufacturing it, for example?

I’m amazed more people aren’t subscribed. This stuff is gold all the time. Thanks !

You should just call this the Tesla channel, since that's all you talk about 😂. You snub so many other companies. Rocket Lab, Anduril, Palantir. Expand your horizons, man. Pull nose out of Elon's butt.

Finally remembered to subscribe to Mext Big Future... Great channel... 👍

Brian, you don't need Randy. He is forgetful and waffles too much, showing early dementia.

Musk says our robotaxis will earn money while we are asleep. Let's think - that's 11pm and onwards, and isn't that the time when drunks are going home with bursting bladders and raucous behavior? And I will want them to have been in my car that I need next morning?

Having watched (from the UK) videos of driving in the USA I can see why they believe in the future of FSD. The USA has terrible driving standards, no lane discipline, they are on their phones, must have loud music on (don't react to police sirens), have an appalling 3rd world accident rate, and all this is on wide, nearly straight and nearly empty roads. Nothing much that is relevant to Western Europe at least.

Robotaxi will not be a success. Running one won't just _"make money while you sleep",_ it will be running a small business which few are cut out for and can be a full-time suicidal nightmare. Unlike Uber where the owner is in close control, most would end up being re-sold (for peanuts) to businesses running small fleets, fleets needing ownership by very hard businessmen to deal with the problems and shit that comes with the idea.

very safe, even some middle eastern well intentioned religious leader can handle it with due care!

Thanks for an informative discussion

I appreciate the information. personally I dont like the drop comment animation. I think it cheapens the presentation. maybe use it less frequently

Food for thought here, if you can automate a chunk of one industry as mentioned (HR/payroll/accounts), then this could be done across many industries, and whilst there are some gains to be made, essentially you are making a huge portion of the human race redundant, permanently. I suspect that whilst most believe, this is still too far off, given the processing power coming and the potential for AI to rapidly ramp up still, a large percentage of the human race may be living on very much borrowed time.

Interesting presentation with some compelling insights. Please can you lose the annoying subscribe pop ups?

Hey Brian and Hans, A really thought provoking session. Brian you helped me move my thinking in this area foward significantly. Whilst I am for the XAI investment from Tesla, I probably still felt that Tesla should be creating it's own LLM style models. But the thoughts you put forward on the risk of this at a business and strategic level makes sense. Greed can make you want to have it all rather than understand what organically will work better. The cross polization and keeping interests in each of the seperate but still vested entities makes sense. It probably shows that Elon has advanced his thinking and is formalising it better now than he did in the early days. The proposition you made about the financial imperative for Ai to be viable at a business model level is brilliant. You are spot on here. If you can't viably fund your compute into the future the venture and the compute growth will falter. It really shows how Tesla has a significant advantage over many others in all areas. I am still yet to fully understand how XAI will self fund its compute growth, but as you say it clearly has Elon with some of his ideas no doubt his sleeve to create value. Exciting times!

It would be interesting to see how the implications of dojo may effect things. If it is viable and tesla is able to ramp production of their own chip they would be able to directly compete with nividia but more likely they would produce as many as possible to use for themselves which has a potential for an advantage in numbers and energy efficency.

AGI Will be man's last invention

Good luck to them, but i wonder how they will succeed where Astra failed, despite a ton of funding, stellar team and advanced manufacturing.

How can you be a Space Expert ... and barely have a Clue about Musk's BF Rocket and Starship ? ... Is this some kind of contrived skit for Dummies !

Need to hire a Boston Dynamics unit to teach this one a thing or two.

Almost forgot.... "Next Year" 😂

We'll import power from Mars to shore up the grid. 😂

Randy, why on earth would Tesla charge $2 a mile when robotaxi initially comes out? Let's assume cost per mile is $1 at the start due to ramp up time, initial slow rollouts etc. This is very high, but Tesla could charge $1 per mile and just breakeven for the first 12 months or so to drive adoption. There'll be infinite demand at this price point, given it's 50-75% cheaper than a regular Uber! It makes no sense to charge the same as Uber prices at the start.

ASI could help humanity develop a human super brain in a Petrie dish encasement that would take humanity next next level…Heard it from me first…klp

My question is simple. How much revenue and profit you need to justify all these investments. The competition plans the same, the revenue pies are not infinite

Thank you guy's!

Your content is so good, you deserve far more followers!

Wild numbers on the XAI future compute table. This is just the kind of original content that we come to NBF for. Any update on the Italian rocket builders?

Having lots of smaller fission power plants that are also mobile is a terrible idea for safety, security, proliferation, economics, and more. The whole point of large plants is economies-of-scale + insurable + centralizing risk into manageable known quantities. Mobile fission sources would do away with all 3 of these. Thankfully, no government will allow these to operate outside of a well-defended military base and no insurance company will insure these boondoggles. Plus, PV and onshore wind (combined with PES and NMC storage) are so much cheaper to deploy, there's no real reason to obsess over fission apart from a religious sci-fi fixation like "flying cars" when viable, cheaper alternatives exist that don't have nearly the same baggage. PS: Book recommendation _50 Years in Nuclear Power: A Retrospective_ written by a key individual from GE NE.

Faster. Please.

What a silly title. Elon is already 100% in on robotaxi.

Do you help in getting government approval for having and using your unit? That, to me, is the biggest obstacle at this point.

What about the nuclear waste? How do you handle this and is the cost for waste disposal included? If not, the price you're talking about (cheaper than diesel generator) is an illusion.

A space tug.

This site is VERY unprofessional

The Super Dragon should be terrific for unmanned probes of the solar system with some sample returns.

In general, to add your car to Tesla’s robotaxi fleet, Tesla sets the price. It would be a condition of joining the network. Tesla can launch in any city in Florida. Cities and counties are not allowed to block autonomous taxis. They can prove the system out at scale in Florida. I personally think that most people are too optimistic of the revenue bull case. Too many taxis will likely add to road congestion, so it tends to self limit. That is why Elon started the Boring Company - to allow for more traffic into city cores.

Elon didn't say they were going to catch me booster on this launch. He estimated there will be three more practice launches.

Its still runs red lights , its not even close to being ready.

How you know how many miles in drive between interactions? Where you get the data from?

your thinking is too complicated. they‘ll put the software on participating cars, make revenue sharing agreement, make insurance arrangement, and then it‘s up to the car owners how they‘ll use it.

In China, maybe. In the US and EU, no chance. NTSB, DOT, will never approve autonomous vehicles in 1 year. Even 5 years to get approval is optimistic. FSD could be godly but NIMBY will win. Not to mention the taxi, limousine and Uber lobbies.