Видео

Robotaxis will be released right away now following the same paradigm as Waymo .. starting with a limited number of cities and likely a limited number of routes. This will be done to prove the system and to limit liability. As the system improves and confidence builds it will roll out to more areas. No need to fear this technology .. it is already working quite well and it will continue to improve. Remember the system only recently started to use AI and full-stack architecture and only recently began using a supercomputer to aide learning ... these AI systems learns fast.

😊Thank you guys 🎉 I really enjoyed your perfect blend of thoughts and insights ‼️

When has Tesla ever sandbagged and released anything early? The hyper bulls will always scream from the rooftops...Model Y release was on time. Ok, that's one time example of something on-time, but with a vehicle that uses 70% identical parts with the Model 3 and almost identical manufacturing lines. Elon's strategy of publicly setting outlandishly aggressive timelines (in hopes of motivating his teams) has officially backfired. Tesla/Elon is truly becoming the company the cried wolf. I would love to see some of these projects developed strictly in the background with a surprise unveil of a product or service that is truly ready to go-to-market. that day. I hope they'll switch to that strategy with Optimus, because telling the public and shareholders that Optimus will be sold to customers for actual meaning work by the end of next year is borderline irresponsible. The delays with optimus will make FSD look like a cake walk!

FSD will always be ‘coming next year’ - that last % is exponentially harder and they may need another re architecture to achieve it

The unveil was delayed from 8/8 to 10/10 for one reason....FSD improvements were not as far along as they hoped!! The very sad news is that it's still the case. Several months ago they were talking full point releases every few weeks, with the assumption they'd be at V13 by unveil. At this point it's almost inevitable that they'll be another announcement within the next year or two that they'll need to again fundamentally change the architecture and/or Hardware 5 (or 6) will actually be needed to reach true Level 4. I don't foresee any sort of full robotaxi rollout for at least a few years, possibly several. If they don't also unveil the new Model 2 for retail sale on 10/10 (and then when you add in Elon's complete submersion into the insane political quagmire), I'll be seriously concerned for the future of Tesla. That would mean they are putting almost all their eggs in the FSD basket with absolutely no guarantee of success, despite what all the uber-bulls are endlessly promising.

“ we should all be a little bit more cautious” Larry speaking wise words with the absence of hopium

Great discussion

It is not a compute issue. LLM based AI’s hallucinate. Improvements are needed in the architecture and training of AI before we can have errorless autonomous driving.

It is all smoke and mirrors. Did you really think, after years of missed promisses to get something on a certain date ? I got bored with this debate after 10mins

Tesla needs someone who tells the truth. Obviously Elon doesn't fill the role.

Trying to decide when launch FSD. 12.3.6 is not good enough to justify $106/mo. I am not convinced that hardware 3 will ever be a full FSD product.

When a stock or company becomes a religion, it's time to take a step back and poke your head out of the echo chamber you've been in for the past months and years. I have every faith in Musk teams across the board, but there will be some hard roads ahead when various entities and powers flex their muscles against one of the most brilliant humans of all time and possibly delay many of the plans Elon has. It's very easy to get caught up in the hype and excitement.

When customers buy FSD, they aren't expecting to become part of a Beta development process. They are expecting the full product. It will require human-level awareness and nobody can say when or if that will happen. It will be another huge flop and one of the things that helps propels Tesla into bankruptcy.

Imagine every company can enhance/fix the map of their car park from tesla gathered data ( e.g walmart, mcdonalds...) then tesla makes the data public so nobody has to do it again. All things like smart summon and banish are fixed. That way odd car parks are fixed by big multinationals. You then map your house one and then all the vague bits are covered. Perhaps including gated communities. You do it once and its done.

Good work Hans

TEsla is not compute contrained, more compute is nice, but really hard technicals are the problem.

Agree with Larry 100%. There is zero chance they demo feature complete on 10/10. 12.5.2.1 can’t handle a blinking red light for Christ sake. Absolute best case is feature complete in California.

I agree with Nico on this one. Tesla FSD team has done enough software releases to know that the real world is not identifying enough bugs to be worth the delay in release schedules.

I have a feeling that they discovered v13 is way better and fully focussing on that version instead of v12 that's why we don't see any new rollout. V13 would also be robotaxi.

Most of the FSD videos are recorded in good weather. I'm wondering if it is a fair weather trained model and not so good in bad weather? I have seen a few bad weather videos and they weren't as impressive. We need more bad weather videos.

Larry - you missed the mark on Wednesday. The FED did reduce rates by 50 points.

The answer is simple. They will not do a 12.5/6 fleet rollout until all the features are consolidated in a stable version. No rocket science here.

I have FSD 12.5.2.1 (AI3). It is pretty good most of the time, but still has some issues here and there. I've had some good drives and not so good drives with it, and I imagine that Tesla would like to iron out these issues before wide release. For me, the intervention count is actually probably more than with 12.3.6, but it's smoother overall. It just occasionally does some weird/unpredictable things.

I thought Larry said, "I'm not going to sell my eggs." It would take a lot of eggs to trade for a Cybertruck. 🙂

clown car company hyping clown car software..rollout slow as FSD is garbage beyond the most simple of situations..we get Tesla ASS rather than real improvement. Tesla at 6 mph or less seems great..clown car software...ps tesla insults customers constantly.."necessary interventions" i make about 30 a day due to garbage software...i want my money back..eventually Tesla will admit HW3 Hw4 Hw5 will not achieve touts of Musk autonomous driving..show is fillled with idiot pumpers particularly the clown at 5:30 comparing FSD fail with washing a car..good grief fluffers.

While Waymo continues to expand its driver less taxis Tesla lags. Typical.

10/10 is happening because they have no choice now . FSD is not near ready to be unsupervised . It still can't navigate school zones and still makes bad lane choices , can't handle rain. I use it all the time but it is not ready to be unsupervised . The only way we see a robotaxi unsupervised in the next two years is if there is other HW built into it . I do believe we will see it on the road but there will be someone sitting behind the wheel (physically or remotely) to supervise the drives.

I agree that the next upgrade is MAJOR and therefore it won't be released until 10/10

What you are saying makes me think they already have solved FSD and are sandbagging up to the event.

over the years Ive caught Larry being intellectually deceptive. I've caught him using notes from other podcasts, misusing definitions, and so on. I think he's a con, or at least masking a shallow intellect

Everyone keeps thinking Tesla is going to explode with robo taxis. Wouldn't that require the non existent FSD? Also, the robo taxi would have to essentially eclipse the entire current car market to hit projected valuations. Complete BS.

I doubt I'll get v12.5 until well after 10/10.

I also suspect all Tesla resources are being poured into 10/10 to the detriment of the v12.5 rollout.

2022 MYLR here, still waiting for v12.5. :(

Listening to Larry, I speculate the compute constraint is partially caused by the use of the resources to confirm the robo taxi build use of FSD.

The purpose of the roll-out of the FSD is to gather data for it's improvement! It's not a priority to make an all encompassing roll-out yet. It's very possible that they don't expect to do a general roll-out until all the currently buit Compute is on-line.

I think they're paranoid of the media. if FSD even makes a too wide U-turn it's going to end up on CNN as an epic failure.

I enjoy your content but I feel the need to Confront you on your spelling of “Conront”.

conront?^^

Norway is offering tons of incentives to push EVs. And they are paying the incentives by selling oil. hmm...anything wrong?

That could be a real problem...

"Underhyping" and Tesla in one sentence. wow

Has anyone reported on the electrification of construction grade vehicles (earth movers, backhoes, dump trucks, etc) and other forms of transport (eg busses)?

Tesla feeding on the government tit even after all these years. Elmo tells his RUclips pumpers to start pumping and the obedient poodles obey him.

Imagine the positive impact on Tesla if Elon invested in an EQ coach. EQ is Emotional Intelligence. Elon is super smart but lacking social skills and can't control his urges to promote disinformation and hate speech. Elon and Trump are in the same boat.....they think they know everything about everything and don't take advice. The Jeff Bazos / Elon comparison is apples and oranges. People don't talk about Jeff Bazos because Bazos doesn't promote disinformation and hate.

LOL. I'm sure Tesla Robotaxi will perpetually be on the roads "next year" 😂

Elons next milestone is 1000 ppm CO2 in earths atmosphere in the future - that was his sneak peek for plan 3.

Hey guys it'd be so nice if you'd talk cars and get out of politics for crying out loud we're Americans first.

When you discuss brand loyalty, it helps your credibility to get facts correct. Tom Brady never "left his pregnant wife" 》 They were never married. 》 They found out that she was pregnant after the breakup. Facts Matter.

As we continue to deny the facts, the Chinese will eat our lunch.