It's amusing that this was at the Warner back lot, which is out in Burbank. Not Paramount or Universal, which are both in Hollywood. It would have been embarrassing to do this at Paramount or Universal, because those are in Waymo's service area. People would have arrived at the event in real self driving taxis.
They claim that unsupervised autonomy in existing cars will arrive in California and Texas next year (with an easy bogieman that it will depend on regulatory approval), but no details as to what exactly this would mean.<p>It’s possible that they might be able to get an Level 3 product out similar to offerings by the likes of Daimler, Cadillac, and Ford - where on certain highways under certain conditions you don’t have to pay much attention but still must be available to take over relatively quickly if the conditions change. That seems the most likely route, although all other systems I believe rely on vision+radar or vision+lidar fusion. Those approaches have a lot more broad industry experience and quantifiable benefits in safety, but it’s possible Tesla has compelling data on the performance of its vision system, especially during daylight hours.<p>I’m honestly not sure how they could ship what they are implying - basically FSD as it is today but without anyone in the driver’s seat. That would imply they are (nearly) comfortable with it driving 10’s to 100’s of millions of miles between fatal accidents without any intervention. Either that or they are willing to ship and know it’s less safe than an average driver. That’s ignoring non-fatal accident rates.<p>There are some middle ground options where UFSD would have a larger set of conditions it can operate “unsupervised”, say in good weather and possibly daytime, and maybe only on some types of roads. But the edge cases where it transitions out of those conditions can be brutal and not easy to address. It’s relatively easy to say “just pull over and make the driver take over”, but especially on highways or heavy traffic that can take a while.
The most interesting thing to me by far was the lack of a steering wheel on the Robocab.<p>Without manual controls, vague promises ("puffery"?) about autonomy won't drive vehicle sales as they do today across all Tesla's models. The Robocab as shown literally cannot function (or make a dime of revenue) until they've fully solved autonomy and have convinced regulators of the same.
Here's an idea - take the Robocab car design, strip out all the FSD/autopilot stuff, put in a steering wheel and dash and sell it starting at the end of the year for ~$25k.<p>It would sell +++
Anybody know why the autonomous taxi isn't just a model 3?<p>I don't see the point in a purpose built two seater with no steering wheel or pedals and I don't know why regulators would approve an autonomous car with no way to manually override it.
Given current Tesla FSD drives like a drunk teenager even in good conditions, I don't see how this is anything but a scam. It's the perpetual "Teslas will be able to fully self-drive in 2 years".
I must say the show seemed a bit lame and the market seems to agree - TSLA down 8% at present.<p>The robotaxis did 5 mph on cleared streets and so seemed much less impressive than Waymos which can deal with real pedestrians and do 30 mph+<p>They only had two seats which is not how you'd make a commercial robotaxi and so probably just made for show to try to impress investors.<p>The robots serving drinks etc seem to have been remotely controlled by humans.<p>It's not that impressive when Waymo are driving people around in real life and also various robotaxis in China eg. <a href="https://youtu.be/izLfWY4c0Ko" rel="nofollow">https://youtu.be/izLfWY4c0Ko</a><p>I've always defended Musk saying robotaxis will be here in a year or two over the last decade because new tech is hard but they now look a bit in danger of being left behind by the competition.
The RoboTaxi looks neat, but I don't get why it only seats 2 rather than just updating the Model 3? What's the utility of an entirely new production line for a car that is less flexible than the existing model that's built at scale?
Classic bait and switch setup. Selling people the future in 2 years time with little to nothing to show for today. Timelines will slip, plans will change, software won‘t be ready, prices will change. And that doesn’t even account for where Waymo is in two years from now, or what the Chinese EV industry is able to pull off until then.<p>Even this underwhelming event was originally announced on short notice to prop up perception when sales looked bad in April [1], delayed by two‘ish months, and then didn‘t even start on time. Oh, and implementing the robo taxi was a two-months project back then [1]. It‘s a ruse, folks.<p>[1] <a href="https://fortune.com/2024/07/16/elon-musk-tesla-robotaxi-delay-august-reveal-design/" rel="nofollow">https://fortune.com/2024/07/16/elon-musk-tesla-robotaxi-dela...</a>
Robotaxi & Optumus robot still looks to be years and years away. However, I'm pretty fond of the idea of having distributed compute nodes that performs computations when idle, the idea of reclaiming parking lots, the induction charging and vacuum bots looks pretty sleek as well.
It's important to remember when this event was announced. It was announced in April, when stock was plunging following a Reuters report that a cheap Model 2 was cancelled.<p>Musk called them liars and announced on the spot the event that we witnessed today (which was postponed from the initial 8/8 date...).<p>That's what fraud looks like folks.
One thing confuses me... Why invest so much into automony on streets/roads when most (all?) US trains/trams/busses still have operators?<p>Elon said himself that busses have 5x the operating cost of robo-taxis (1$ vs 20¢ per hour), but failed to mention that a typical bus carries 25x people. Presumably the cost delta is due to payroll for a human operators & janitors, so if busses already have 5x max throughout per $ in a worst-case comparison, why not put more energy into driving down costs on the more efficient modality?<p>I imagine it would be an easier problem to solve too, as you could could define constraints on your operating environment. Dedicated travel lanes, rail, and preprogrammed routes seem like they'd massively reduce the complexity of the problem.<p>I could also envision a decent financial argument for moving from a customer base of general-consumers to one of cities & local govt. If a 5x operating cost reduction is feasible, I could see "public transit automation" making for a pretty compelling capital project.
Who is the customer for this? Sub $30k appeals to regular people. But why would regular people want to own a robotaxi? I want robotaxis to exist so I don't have to own a car. I want someone else to own it, and I pay to use it once in a while.<p>If this works, wouldn't just Tesla operate as a taxi company?
The big promise of the Robotaxi was that every Tesla would be one, that your Tesla would earn you money while you weren't using it. This was obviously unlikely to ever happen, but has been the promise right up to this event, and something Musk has been very vocal about.<p>Despite having very few details in this presentation, the one detail that is clear is that existing Teslas won't be taxiing anyone, and Tesla will be the operator reaping the benefits. That's a significant under-delivery, especially for the average Tesla retail investor who believes in the mission and is driving their stock price.
I find these half measure autopilots way more stressful than actually just driving.<p>Level 2-3 FSD is worse experience in my opinion.<p>1) driving as human — if I see brake lights, I apply brakes: see my turn, turn the wheel.<p>2) Tesla “driving” - if I see brake lights, I have to evaluate “did FSD See those lights, is it applying brakes” EVERY TIME. Because I need to pay attention. THEN I may apply the brakes and turn wheel (and if you use FSD a LOT, those skills will atrophy). it needs human intervention about once a day. But you never know when that will be.<p>We either need FSD or humans driving. Shared dynamically adhoc responsibility for the car is way way worse.<p>The CyberCab at least improves on that by removing the steering wheel, so when it makes mistakes you just along for the ride.<p>I am a curmudgeon though; I don’t even use cruise control, and the radar following cruise control gives me the same hereby jeebies “is it braking??” Problem
What if there are humans acting as these Tesla Robots in this product demo or the Tesla Robots are just remotely operated by humans to fake this demo?.<p>Very skeptical of this whole presentation, even if Elon and Tesla are overpromising on timelines.
The Robovan—as an actual van one could buy today—would sell. Especially in Asia, versus Toyota Alphards. Alas, seems it's more likely to get used as point-to-point transit inside closed spaces (parks and convention centers and ... perhaps the Vegas Loop).
It's ironic that the example he gives for driving across LA already has a fast train connection:<p>> People that live in LA, I mean try to get from Pasadena to El Segundo during rush hour. You can fly to another city faster than you get to crosstown LA. And you have to drive the whole way.<p>You can take the Metro A line from Pasadena, then transfer to the C line to get to El Segundo. No driving necessary. Musk sells cars, so of course he has a massive incentive to say more cars are the solution to peoples' transit woes. But it seems like throwing more cars at the problem will simply make traffic worse, and from my experience living in Chicago, the best solution to avoid traffic (and parking!) is to take an alternative mode of transit that can bypass it (e.g. train, bike, electric scooter).
With regards to "can this ever actually be done?" I think it can*, but also it's got all the problems of <i>both</i> a software project <i>and</i> fundamental research: the difficulty is unknown, the timeline is a wild guess.<p>That said, no steering wheel? Finally, he's met my long-standing requirement for what counts as "genuinely self driving".<p>* humans don't need lidar — we clearly benefit from all the extra sensors or we wouldn't even have rear view mirrors let alone parking sensors, but <i>technically</i> it's believable.
After predicting that, in the long term, humanoid robots will cost less than a car, Mr. Musk added: "It'll take up a minute to get to the long term".<p>Please stop trolling, Elon.
Because of the word taxi in the name it gives off the initial impression that this is a product for fleet use and not something you own personally. But I don’t think that impression is going to explain the larger reality. if it’s inexpensive and possible to exist in this form, this is what you’d expect people to buy as a personally owned vehicle specifically for commuting.
I'm confused, cause all these comments seem to just be plain lies? Have anyone actually driven in a tesla using FSD? It's freaking awesome. I literally drive to and from work everyday and pretty much never have to put my hands on the wheel unless i want to go faster.
They are going face major regulatory hurdles. Safety concerns are huge—both real and perceived—and governments, especially in places like the US and Europe, are still far from trusting fully autonomous vehicles. Add in fragmented regulations across different states and countries, plus potential public skepticism, and it's clear this won't be in place any time soon.<p>Labor opposition is also going to be a nightmare. Just as Uber faced resistance from taxi unions and legislators sympathetic to workers who saw their jobs threatened, Tesla will likely face significant opposition from drivers in the gig economy, who rely on ride-hailing platforms like Uber and Lyft. Governments may be pressured to protect those jobs, especially in regions where automation is seen as a threat to employment.
The Verge has good coverage.<p>[1] <a href="https://www.theverge.com/2024/10/10/24267132/tesla-robotaxi-we-robot-autonomous-fsd-elon-musk" rel="nofollow">https://www.theverge.com/2024/10/10/24267132/tesla-robotaxi-...</a>
The race to full self driving is not going to be close. Tesla has no competition.<p>The anti-Tesla sentiment here is hilarious to watch though. All because this place is largely leftist, and their CEO is on the right.
Nobody talking about those gimmicky butterfly doors? I would bet some money that they will not make it into mass production. Not for a car priced unter $100.000
This is not a serious proposal. It's a cool little sports car looking design. It looks very future-y. It's the kind of thing someone would step out of in a sci-fi movie. That's how you know it's not an actual well-thought-out design. This isn't a taxi, it's the torment nexus in car form.<p>Taxis are not sexy, taxis are utility vehicles. Taxis carry people and goods. Taxis carry more than one person. Taxis carry disabled people. Taxis carry the elderly. If you want to know what a purpose-built taxi looks like, look at the London cabs, the JPN Taxi, or the NV200s. If you're going to build a fully custom platform for a taxi, make a goddamn taxi. Actually put in the effort to design something built for the purpose, not just something that you think looks cool.<p>(And if the design didn't give it away - inductive charging? Really?)
This is all well and good, but it's not possible to deploy these kinds of technologies at scale in American cities (well, most American cities). They'll get torn to bits. The ability to actually move to an autonomous transportation future is downstream from citizens actually behaving in a way that allows this. Harsh, but true. Maybe in Singapore.
Oh hey I get to copy paste a list of times elon said FSD would come out next year![1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16]<p>[1]<a href="https://fortune.com/2015/12/21/elon-musk-interview/" rel="nofollow">https://fortune.com/2015/12/21/elon-musk-interview/</a><p>[2]<a href="https://x.com/elonmusk/status/686279251293777920" rel="nofollow">https://x.com/elonmusk/status/686279251293777920</a><p>[3]<a href="https://www.theguardian.com/technology/2016/jun/02/self-driv" rel="nofollow">https://www.theguardian.com/technology/2016/jun/02/self-driv</a>...<p>[4]<a href="https://www.nbcnews.com/business/autos/driverless-tesla-will" rel="nofollow">https://www.nbcnews.com/business/autos/driverless-tesla-will</a>...<p>[5]<a href="https://x.com/elonmusk/status/823727035088416768" rel="nofollow">https://x.com/elonmusk/status/823727035088416768</a><p>[6]<a href="https://www.ted.com/talks/elon_musk_the_future_we_re_buildin" rel="nofollow">https://www.ted.com/talks/elon_musk_the_future_we_re_buildin</a>...<p>[7]<a href="https://x.com/elonmusk/status/866482406160609280" rel="nofollow">https://x.com/elonmusk/status/866482406160609280</a><p>[8]<a href="https://x.com/elonmusk/status/1063123659290595328" rel="nofollow">https://x.com/elonmusk/status/1063123659290595328</a><p>[9]<a href="https://web.archive.org/web/20190220051410/https://www.ark-i" rel="nofollow">https://web.archive.org/web/20190220051410/https://www.ark-i</a>...<p>[10]<a href="https://www.youtube.com/watch?v=dEv99vxKjVI" rel="nofollow">https://www.youtube.com/watch?v=dEv99vxKjVI</a><p>[11]<a href="https://www.youtube.com/watch?v=Ucp0TTmvqOE" rel="nofollow">https://www.youtube.com/watch?v=Ucp0TTmvqOE</a><p>[12]<a href="https://web.archive.org/web/20200709130939/https://www.youtu" rel="nofollow">https://web.archive.org/web/20200709130939/https://www.youtu</a>...<p>[13]<a href="https://www.youtube.com/watch?v=AF2HXId2Xhg" rel="nofollow">https://www.youtube.com/watch?v=AF2HXId2Xhg</a><p>[14]<a href="https://www.businessinsider.com/elon-musk-interview-axel-spr" rel="nofollow">https://www.businessinsider.com/elon-musk-interview-axel-spr</a>...<p>[15]<a href="https://www.cnet.com/roadshow/news/elon-musk-full-self-drivi" rel="nofollow">https://www.cnet.com/roadshow/news/elon-musk-full-self-drivi</a>...<p>[16]<a href="https://www.youtube.com/watch?v=Vwq_PhtvLwo" rel="nofollow">https://www.youtube.com/watch?v=Vwq_PhtvLwo</a><p>(Or simply, <a href="https://motherfrunker.ca/fsd/" rel="nofollow">https://motherfrunker.ca/fsd/</a>)
So this should be available next year … like the full self driving cars since 2012.<p>This is just another to have something in the pipeline to keep shareholders at least interested.
It seems Musk tries really hard to keep the stock price from collapsing.
what the world really needs is a robobicycle and according to this video we are nearly there :-)<p><a href="https://www.youtube.com/watch?v=0F1H1V5MxtY" rel="nofollow">https://www.youtube.com/watch?v=0F1H1V5MxtY</a>
The event was objectively smoke in the eyes:<p>- days ago I was reading an article stating that Tesla didn't apply yet for the license to operate autonomous vehicles in the streets. Competitor had their licenses in 8 months or more. I think this is a critical factor to respect the deadline of December 2025;<p>- Just another risible demo in a controlled environment, a movie set: no real life scenario able to demonstrate the effectiveness of the robotaxi technology. This for something should be in exercise before 2025 ends is and indicator that the tech is not ready. "Only cameras" approach IMHO just won't work.<p>- the presentation images suggested that robotaxis should substitute public transport. This not only is bullshit for a number of reasons, it also can influence public transport politics like with the other bullshit technology called Hyperloop that was accepted as the future of transportation by short sighted administrators;<p>- Wireless recharge : oh please ! Apart technical consideration could raise doubt on the smartness of that approach, they really are saying they can create a network of wireless recharge points before the robotaxi go in exercise next year ?<p>- Robobus : wow ! What about electric tram ? I see a pattern here , step by step, they are demonstrating that the real electric revolution is electric public transport, not that electric robot limousine that are a viable and cost effective public transport alternative only in the mind of a megalomaniac billionaire !<p>- Men, you don't believe this is bullshit ? let's take appointment here December 31st , 2025, to discuss the status of the robotaxi.
Why is this a two-seater? A sedan with no driver can easier fit 5 people. Building one for only two is.. why? There are also no steering controls, this this vehicle can't ever be dual-purpose. Even something where steering/gas could be removable attachments, would be better. Allowing a possibility of being driven by a driver should the need arise.<p>But really? A two-seater?
I'm not a Musk hater but this vision of the future just doesn't seem that interesting. I'm from Switzerland, I much rather have trains and well timed buses.<p>Buses already exists and paying drivers isn't actually that big a problem. Society can easily pay for that, if you take into account the reduction in investment and cost you have from other road use.<p>Musk promise of 'less parking lots' can already be easily achieved with technology from 1960. Its not an engineering problem, its social problem.<p>Trains, trams, buses, bikes and walking is far cheaper and more efficient in every measurable way then fancy robotaxis even if they worked, witch they don't really.<p>Can you imagine the horror of a large city where most people each use an individual vehicle? That just dystopia.<p>An small autonomous bus has some uses but at best its a small part of a much larger transport system.<p>The US being so obsessed with robotaxis is just a consequence of 70 years of horrible road design and land usage and city planning.
> Mr Musk's prediction that production would begin some time "before 2027"<p>So, no big reveal after all? Just another press conference where he once again says "it's coming soon"?<p>I didn't think he was going to unveil anything actually new, but I did honestly think he'd sing a new song about it all. I guess I overestimated him.
About the Robotaxi: I really, really wanted to impressed by what they will demo. However, I thought it was more marketing, with the product being "just two years away" as always. The demo was in a controlled environment so I doubt its real-life capability. I guess I will believe it when I see it on roads. Disappointed.
I think the products all look great. The timeline, who knows. They probably will manage to get at least the Robotaxi out as a buzzer beater December 2026 release of a handful, at a minimum.<p>Unsupervised FSD timeline? Elon says 2025. I suspect they can do that, if you define it as Level 3, where you just need to take over within, say, 10-30 seconds. Probably starting on highway. Beyond that, hard to say. Recent FSD versions are impressive, but there is still a lot to do to reduce disengagement rates. But their strategy seems like it can be made to work.<p>How soon will it take to get it good enough for Robotaxi/Robovan? Well, they can basically have said no sooner than two years (for the late 2026 Robotaxi release). I doubt unsupervised FSD will work in a fully autonomous taxi mode on existing vehicles until the very end of that window.<p>Will the Tesla AI3 computers support Unsupervised FSD, sounds like possibly not, based on Elon's suspiciously noncommittal response to a guy yelling in the crowd. They can probably squeeze out of past promises on this in a couple ways if it won't work on AI3 cars. Presumably they'll want to minimize cost impact of making good. I think they'll first try to get as many FSD users to upgrade to a new AI4 (or later) vehicle by incentivizing them with free FSD transfer. For the rest, they could offer even bigger discounts on upgrading to a new car, and then finally bite the bullet and do some kind of FSD retrofit or simply refund. Or just piss everyone off and wait for the class action. But I suspect they'll resolve it sort of amicably while wanting to sweep it under the rug.<p>They could also never be able to make autonomy work, but I think that's doubtful at this point. Waymo demonstrates you can do it with geographical, time, and other constraints, limited release, and a different hardware stack. I think Tesla's vision-only approach has been fairly well validated on the existing hardware. The remaining issues do not seem like they require significantly improved sensors to operate in at least 90% or so of conditions, which is more than sufficient to run a taxi service.
For me the surprising thing is that the stock fell "just" 8% after this event. No sane investor should believe anything CEO says, given he is running multiple companies, is busy with upcoming election, as well as propagating conspiracy theories and hate speech on his platform.<p>After a decade of lies, this is another nail in the coffin and I honestly can't believe the marketcap of this company hasn't been significantly affected by something that was supposed to be a big reveal of those promises.
I'm disappointed that it wasn't a single occupant, three-wheeled, ultra-basic microcar limited to 50 km/h to gather and scatter people from public transport stops, solving the last-mile problem.
Overpromised and underdelivered, but Elon did make a strong case for a very positive autonomous-driving future. Unfortunately for Tesla, Waymo is at least 5 years closer to delivering that future.
This is honestly quite painful to see. Makes me all the happier to live in Switzerland and be able to use great trains and public transport everyday. (It's definitely not cheap though but the comfort is amazing)
You know what screams "democratizate transportation"?<p>99% single occupancy vehicles (well, that's democratic, I'll grant them that) controlled by a corporation known for a million abuses and basically no oversight, everything managed by faceless software.<p>Let me tell you an Uber story.<p>I've been an Uber customer across multiple countries for many years, and one day, I had a layover in London. So I tried to book an Uber in London, but it failed a few times for technical reasons (so their problem).<p>After a bunch of attempts, I received a notification my account was banned.<p>I contacted support, they gave me the runaround for 30 minutes or more, in the end their response was the standard BigCorp canned answer for fraud: "we can't tell you why your actions are suspicious" (with the subtext: if you're an innocent victim of our scans, tough luck, but we can't tell you because statistically if we tell you what you did wrong, real fraudsters will learn and defraud us even more).<p>So now I can't use Uber in the UK (I imagine if I try to circumvent the ban with another account, I risk that getting banned, too, and who knows what else, as I have to put my credit card in their app).<p>Now imagine if you want to plan your life around stuff like Tesla Robotaxi and they ban you. What's your recourse?<p>With your own motorbike/car, you need to commit serious crimes to lose your license forever. With public transportation, as long as you pay the ticket, nobody can ban you for life. And I don't think anyone can take away your bike/ebike/scooter/escooter :-)
I think their "robotaxi" is kinda pathetic. But goddamn... That bus is one of the single most beautiful vehicles they have ever designed. I absolutely love it.
(Fan of Art Deco and streamliner locomotive design)
Elon Musk Has Been Promising Self-Driving Cars For 10 Years [Update - We Are Now On Year 11] [1]<p><a href="https://jalopnik.com/elon-musk-tesla-self-driving-cars-anniversary-autopilot-1850432357" rel="nofollow">https://jalopnik.com/elon-musk-tesla-self-driving-cars-anniv...</a> [1]
HN is packed now with Elon Musk haters after the mass firings at Twitter. Anything related to him or his companies has long lists of negative comments. Bitter grapes?<p>Even if Elon cured cancer, there would be haters for his politics and business approach.
Hey Tesla, where's the Roadster 2 you announced in 2017 and took $50,000 deposits for?<p>Oh, that's right, it was a stunt to boost the stock price, not a real product you intended to sell. Just like this.
Why are the comments on this topic so consistently low quality? It's not even cynicism. It's just snark. It seems like most commenters aren't addressing the topic but rather are venting about their feelings.
Man I really have to stop reading hackernews comments, its a total bummer.<p>His companies are making rockets, autonomous humanoid robots and autonomous cars, I will cut my left pinky finger to work on any of those, and this is my capslock finger, so I will have to switch away from emacs..<p>With the worldlabs[1] work maybe they unlock models that do have even better spatial world model and can be used label data even better and faster, or create even better synthetic data so it can unlock FSD even sooner.<p>The fact that LLMs work means there is structure in language that is beyond our understanding, and yet the transformer can discover it and program itself to solve for it. I think that the stupendous amount of compute that is going to be released in 2025 will make it possible to train labelers that can do temporal labeling much better than humans and than the current models, and synthesize and perturb data to train really really good transformers that will outperform 90th percentile humans.<p>Why do you guys think that it can not happen? Maybe its not Tesla that does it, but I certainly think amazing tech is coming.<p>Hopefully it will not be just autonomous humanoid robots with guns paid by the military and patrolling the borders :astonished face:<p>[1]: <a href="https://www.youtube.com/watch?v=vIXfYFB7aBI" rel="nofollow">https://www.youtube.com/watch?v=vIXfYFB7aBI</a> (Fei-Fei Li and Justin Johnson)