Tesla Bot Review
HTML-код
- Опубликовано: 28 сен 2024
- Join my community at JohnCoogan.com (enter your email)
Really amazing stuff from Tesla AI Day. Even though Elon said he would have a prototype by next year, there was still a lot of negativity online saying it was impossible. I think it’s a lot more likely than people understand, but the rest of AI Day was great too. I particularly loved how watchable the technical presentations were, they didn’t dumb anything down, but they still made everything accessible through really good examples. Great stuff!
KEY SOURCES:
Tesla AI Day: • Tesla AI Day 2021
Lex Fridman Breakdown: • Tesla AI Day Highlight...
ABOUT JOHN COOGAN:
I am the co-founder of soylent.com and lucy.co, both of which were funded by Y Combinator (Summer 2012 and Winter 2018).
I've been an entrepreneur for nearly a decade across multiple companies. I've done a lot of work in Silicon Valley, so that's mostly what I talk about. I've raised over 10 rounds of venture capital totaling over $100m in total funding.
I work mostly in tech-enabled consumer packaged goods, meaning I use software to make the best products possible and then deliver them to the widest possible audience. I'm a big fan of machine learning, python programming, and motion graphics.
OTHER VIDEOS:
Peter Thiel’s Big Bet on Space Manufacturing - • How Varda Solved Manuf...
The Only Reason SpaceX Works (Gwynne Shotwell) - • The Only Reason SpaceX...
CONTACT:
You can get in touch with me via Twitter: / johncoogan
Disclaimer: This video is purely my opinion and should not be regarded as a primary source. I am not a financial advisor and this is not a recommendation to buy or sell securities. Always do your own due diligence.
“ship a fully functional humanoid robot” should have been “fully functional prototype”
I'm so glad you issued this correction before I typed an angry comment ranting about how Elon never said anything about shipping a fully functional humanoid robot!
Having a bachelor in robotics. And continuing on a masters i thought right away.... total bullshit. They won't make it in 1 year. I hope they do, but they wont. They're trying to solve an insanely complicated task with way too many variables to do it that fast.
@@FrederikJChristensen Well they've delayed Ai day from August to september so they can reveal it on AI day 2 as confirmed by Elon. So your prediction is not looking good so far.
@@FrederikJChristensen Tesla has the best (Besides Space X) engineers in the world. So regardless Tesla is already in the lead. If it gets delayed by 1-2 years that is fine.
This channel deserves way more subscribers
Love hearing that! The growth has been really strong and I feel like it will continue!
Totally agree. This was one of the best summaries of AI day I’ve seen and I’ve watched the whole thing and other summaries.
Completely agree!
And the video way more likes too!
don't worry
"Most people overestimate what they can do in one year and underestimate what they can do in ten years." -Bill Gates
Great quote.
And there are people who have overestimated what theu can do in decades too
Don't generalize
@@mehul205 them increase the horizon
@@JohnCooganPlus Here is another one: Correlation does not imply Causation!
When John drops a new video, you always know it is going to be a banger 👌
I heard some good argumentation made on the topic of the humanoid form being a good idea. The principle is that, since it will be a general purpose bot, it will have to navigate the world and use the tools that are made for humans. If you think about something as simple as taking grocery, it will have to walk on stairs, grab things made of different shapes, from tall schelves etc... A humanoid bot fits in that environment. I think that bots with different physionomies will be the specialized ones in the future.
This is pretty nonsense. Because a humanoid robot has to be not only “fitting” in the environment. But it also has to do a lot of tasks, which cannot be done with cutting edge technology - and it cannot be done within years.
You want to have a robot which helps you to carry your shopping? The best way is to design an autonomous shopping trolley. It will be also far cheaper than a humanoid robot, which can do several things - but nothing really great!
We have already Siri, Alexa, Google Assistant, Roomba, self driving cars, etc. - basically all robots. Maybe some of the systems can be more multitasking than others. However they are a far better idea than an android. Have you seen the arm carrying capacity of Tesla Bot? 5 pounds?
@@Dominikmj No, because then you will need one specialized robot for every task in your life. That is much more expensive and less convenient.
Elon said they would have a prototype not make delivery by next year
Yes, but lots of people are still skeptical of even that claim. My goal with this video is to show that a prototype is completely possible, and commercial delivery might be possible sooner than people think.
Elon mentioned having a PROTOTYPE ready maybe by next year. This did not imply a fully functional AI robot as this video seems to indicate.
Yeah, it's clearly going to be a big continuum. This year we see a mannequin and dancer in a spandex suit. Next year we see a simple prototype. Within a few years we will be seeing videos on par with Boston Dynamics and then eventually it will ship. Clearly a very long term project. My thinking is, he could release something as good a Boston Dynamics if he really wanted to next year, and if he delivers within 20 years that will still be impressive.
@@JohnCooganPlus hopefully we’ll get to see some of the many pratfalls the bot is sure to have, not unlike the F9 booster landing attempts just less explosive 🧨
Love this one. You probably have the most interesting comment section I've seen on a RUclips Channel. Everyone seems smart. How do you learn so much stuff so quickly?
Just a good stream of information from quality videos, podcasts, and books combined with a strong network of people.
Interesting perspective, thanks for sharing. I'd love to see this thing come to life.
For sure, me too!
Great video again. I’d like to add a a small note: TPU’s are not just training chips, they can be used for both training and inference. In fact, they used to be inference only till around 2017, when Google made them training capable. So TPU’s can be compared to GPU’s but the D1 is purpose built for training. It would be fascinating to find out what specific advantages Tesla saw in designing a dedicated chip vs just adapting their inference chip to enable training a la TPU’s.
Interesting! I didn’t know that. Thanks for sharing. Definitely need to dive in deeper. Such a big shift in the way companies are doing things.
Do you know who’s manufacturing the D1 chip for Tesla? Or where it’s made?
@@JohnCooganPlus I don't think that info is public yet, but I remember them mentioning (during the q&a iirc) that their process node is some number over 10nm, which leads me to think it might be Intel. But that's just a guess. It could be Samsung or TSMC or a smaller fab for all I know.
This Tesla bot was announced to pump the price of the Tesla stock in order for Elon to sell 10% of it. You will never see this as a product, as specialized robots are much more efficient then general use robots like this one, that is harder and more costly to make. No price was announced also if I remember right, the deadline is ridiculous : only redflags of vaporware.
Yeah, I mean where is the Cyber truck? where are the Tesla Semi's? Hyperloop? So many broken promises, all from the guy who in 2020 said corona wasnt gonna be a big deal.
@@HelgastJon yeah. The robot felt like another promise on top of the pile. He will probably end up delivering the cars tho...
How in the world do you not have more subscribers? Keep at it man and you will blow up real big real fast! Thanks for such in depth videos. And also, do you work in the computing industry? Your comments and ways of explaining concepts makes me think you certainly are!
I think jobs was talking about the PowerBook, which shipped in 1991
From Atlas the backflipping robot, to the near human-like expressions of Ameca, the Tesla Bot's unique feature is still unknown to us yet. Exciting!
Maybe they are monitoring boston dynamics very closely!
I'm sure Elon has kicked the tired on acquiring them. Would be interesting to know why he passed. $1.1Bn isn't a lot of money for Tesla, so they must be thinking about things in very different ways.
@@JohnCooganPlus yep! It will be damn exciting!
Elon is actually the real life iron man
Remarkably accurate.
Uhmmmmm
NO
IIRC he said would, "have a prototype next year" not, ship. But either way agree on all else. The bot is cool but was the AI / computational tech that blew me away.
Also, Hyundai just bought Boston Dynamics last year oddly. So 🤔
Yeah, Boston Dynamics has been bought and sold a few times now. He could probably buy it if he really wanted to... but obviously that's not the goal. That example was mostly to prove that humanoid robots are already doing impressive things and Musk has lots of resources. Good point about the "prototype" next year, but I saw lots of press articles about how even that was an "impossibility" when I don't think it's that far fetched.
@@JohnCooganPlus I don't think it's farfetched either, especially since they actually already have the basic level hardware expertise as Elon alluded to. I'm just curious on what Hyundai is planning to do with Boston Dynamics tbh. Wouldn't be surprised to see them try to be a surprise competitor in the robotics / AI space
He didn’t claim he would ship a fully functional robot next year he said prototype. I know a lot have pointed this out to you and I did watch the video. You dont need to put words in people’s mouths your content is great but I think most of us want honest videos not hyperbolic at least the people you should want to watch your stuff want that. Within 3 seconds of the video you loose credibility with anyone who actually pays attention to this stuff.
That wasn’t his target ship date as you said he never gave a ship date period, did you watch AI day?
Yes I watched it lol, the point is that people think it’s impossible when it’s not.
I know I was just trolling on that part regarding if you watched AI day.
Keep it up I’m shocked you don’t have more subs. I’ve enjoyed your videos.
As a Tesla fan I'm just pissed. The stupid robot is a distraction for the utter failure of the Full Self Driving promise.
thanks for this awesome analysis and summary :)
Wasn’t the book Steve described a laptop?
that's a good point, definitely could have been.
Hi can you tell me the name of the first music
It's from Artlist.io, not sure the exact track. But I really like ANBR, the song is probably from that artist.
I think the first purpose of the Tesla bot is to fill in all the gaps left by manufacturing robots. Elon is trying to create his alien dreadnaught to have a factory this is almost entirely automated. There are also many random tasks or certain tasks with a lot of random variation that aren't worth creating a purpose-built robot, like walking packages from a self-driving delivery truck. In this example, there are a ton of random challenges like stairs, doorways, cracked sidewalks, parking lots, dirt paths, snow, pets, and many many more that I can't think of off the top of my head.
Yeah, it will be super interesting to see where it goes. There are obviously a ton of tasks that would be better suited to robots, so maybe the first version is still very tailored to a specific use-case.
Great video. 👍🏻
Love 4:20 😎
I’m almost through all of your videos. You (and your team) having done all of this research, could possibly make you an apt futurist. For the sake of posterity, you should pull an Asimov and film your predictions for 2050/2100!
He had a person dancing on stage to demostrate some obscure robot. Regardless of what he has achieved as CEO, a revolutionary robot being released from someone who said it was a revolutionary idea to have single lane tunnels underground to travel through cities is not someone i put much faith in.
I think the key point about making a humanoid robot, is to create a single robot capable of doing everything. Today everything is made for humans, so to ensure compatibility you need a "human" robot. In theory it can be your chauffeur, your cleaning person, your masseuse and so on. No matter what software they put in a Roomba, it will be difficult for it to scale with the software.
I love dis channel
Apple (AAPL)
As the most valuable public company in the world, Apple has plenty of money to invest in the technologies of the future. In fact, its cash on hand is $204 billion as of last quarter, so it's no surprise the company is using considerable resources to expand its commitment to artificial intelligence.
This year, Apple recruited a prominent Google artificial intelligence expert, Samy Bengio, and is also building out a new campus in the Research Triangle of North Carolina. The campus, which will house jobs in machine learning, AI and software engineering, will create at least 3,000 jobs.
I wish people would focus more on Elon's ambitious plans and what he wants/plans to deliver rather that when he claims he will. I don't understand the skeptics attacking the timelines. It makes me wonder if they're actually skeptical more about the technology itself and start by attacking the ambitious timeframe
Hopefully one day, Elon will release "Master Plan Part Trois" to see what his vision is for the Tesla bot post FSD. People are too nit-picky about his timelines, can't wait for Tesla to continue proving these non-first principle thinkers wrong
Completely agree about people being nit-picky about timelines! Would love to see a Master Plan for this idea.
Sure 😂
Nop
No
i love u... u conten is good
Impactful... Sweat shops... Yes very impactful.
MKBHD wierd take on tesla bot, he believe this wont work because people want specialized Machine with AI, if that is true than why we still bother trying to create AGI?
tesla bot could be the vessel for AGI machine of all trade, master of everything.
imagine tesla bot cook your food in the morning and drive you to office using your car. why buy smart car if tesla bot can make any car with streeing wheel and gas padle a smart car.
Yeah, I think humanoid robots could actually just be an intermediate step. This is a long way away though, we’re clearly going to have both. Highly specialized robots and general purpose robots.
The robot is taller than me
Prototypes don’t ship…
WTF?!!?!?! How did i not even know about this?! Holy shit that Elon!! hjahaha he's an alien!
↑ Comment for algorithm + 👍 'd by the way, i Sub'd too.. Your videos are stoopidly professional, in-depth and well-crafted!
#OldManPaxus Rating: ☆☆☆☆☆
Except musk making way more cool, impactful technology.
Much better in my opinion then iphone ,facebook..
If you want Elon to do something tell him that it is too hard
For sure. He loves to prove the trolls wrong.
You know i have concentration problems and i don't why may be the way you talk, i don't loose much focus listening to you lmao
Huh? Elon said they'd have a working Prototype in 2022 not that it would ship. This is ridiculous.
Nice video
Thanks!
I somebody build TARS for me to buy, I'm pretty sure I'm going to sell my kid to pwn one.
The Tesla bot will not ship next year.
He can't, the engineers he hired maybe.
Take Google’s sister company Waymo, the industry leader in self-driving cars.
Even with extraordinary amounts of time, money, and effort invested, no team could figure out how to have AI solve a real-world problem: navigating our roads with the high degree of reliability needed.
Your video is filled with lots of misinformation
Feel free to explain, if I got something wrong I’m happy to pin a comment explaining things.
😍😍😍🤩🤩🤩🤩
Elon is one of those people who gives me hope for the future.
A future of successful conmen?
It would be cool if Tesla added bitcoin or udst as a payment method
Discovering this channel while it’s under 5k subscribers is like investing in $TSLA in 2012
Love that analogy!
@@JohnCooganPlus Love your engagement! Lol
@@JohnCooganPlus Subbed - Looking forward to more Metaverse content from you!
To be fair, Steve Jobs was essentially talking about the PowerBook which came out more than 10 years before the iPad. He was more like 10 years late and he also got fired from apple during that time which didn’t help.
Ok, you've convinced me - I may have to actually eat my hat
Ahhh, yes. Another jest from the Vaporware King.
I've been on a coogan binge all day. These videos are fantastic!
Great video and presentation! Loved it and enjoyed learning. Thank you 👍🏻
Glad it was helpful!
Thanks John, great video.
I think Elon is betting on a big switch in people buying cars to people buying home robots in the hopefully not to distant future. He is in a great position then with all knowledge acquired about mass scale manufacturing & deep learning. Its remarkble how these big tech companies not only branching out to other technologies but also make them all build on each other or fuse them alltogether. The goal for humanoid though has really surprised me when you think about automation in industry sector as its normally highly specialised to given tasks etc. I can see more of a wholistic approach here hacking our work environment with these robots beeing so versatile and aplicable in most jobs done by humans really. Maybe we will see Tesla Bots working in some Amazon warehouses soon haha.
Definitely, it will be interesting to see where all this goes. It's clearly going to take a long time, but will be very fun to watch.
if we are going with how tesla first made roadster then move onto cheaper mass market model car. The first gen tesla bot would probably be very expensive and doing special high margin job like astronaut before move onto low margin job like amazon warehouse worker. Probably acting as remote control astronaut on space station or moon base. Lot easier and cheaper to sent up tesla bot to harsh space environment than sending living astronaut, don't need to worry about low gravity bone less, air leaking, radiation damage, space danger, etc. Plus the cost of a 200k will be cheap compare to the 90 mil per astronaut ticket, so tesla bot can pay off the ROI extremely quick compare to sending tesla bot to doing warehouse worker where it take multiple year to pay back the initial investment. Similarly tesla bot is very suited to work in toxic or biologically dangerous environment like nuclear reactor or chemical lab or deep sea under water, where the cost of shipping a human there with proper protection is well above tesla bot's cost.
@@keenheat3335 Thats indeed a possibility, NASA already uses humanoids for remote control delegate tasks etc. Maybe thats why Elon is going for humanoids. 90mil per astronaut is propably due to weight issues,
the same applies for robots unfortuantely.
@@metamind095 according to wiki cost per kg on falcon heavy is around 1400 usd. so a100kg (221lb) robot could cost around 140k. So i guess including robot base cost of 200k. probably 350k total. can probably launch a small army of tesla bot for the cost of one astronaut
The problem with him consistently missing deadlines *is* that there is no problem, nothing is holding him accountable because every time he makes a promise his stock just goes up, he's encouraged to lie due to greed. He is awarded infinite integrity despite being late and under delivering on every promise he makes due to the shear media hype. It's a bubble fueled by hoodwinked investors and government subsidies and as much as It should theoretically pop one day he has too much wealth to ever really be stopped.
"Builtding the Teslabot doesn't break the laws of physics, so it's possible!"
This is the kind of optimism we need 🙂
there is an advantage of having a huminoid robot, if it has the body and functions of a human it can use human tools making easier to intograde into a house hold.
No, he said prototype by next year. They need a factory and suppliers. They dont even have a perfected model yet.
Extremely well said and outstanding analysis. The way you tie the important key points of the technology is rare. Your research and delivery are truly gems.
what i got from Elon is that pushing the time line helps you do the job for real..end of story
Yes, in many cases, saying "let's try and do this next year" will make it more likely that you successfully deliver eventually than if you had said "we need 20 years to make this a reality" - people want to work for optimistic people.
@@JohnCooganPlus I agree John.
AI will never be able to replace Gwynne Shotwell. Who is, in my opinion, the only reason why SpaceX works
Hahahaha love it
That dancing Tesla Bot at AI day was so cringe.
He never said ship. Elon said a prototype by 2022
no
no
No.
He did not say that. He said a prototype, with dimensional specs. Functional specs are not mentioned nor is the verbal interface. Tesla people think function in 5 years and available at scale around 2030 or whenever Mars dictates.
Yeah, I agree with all that lol. Did you watch the video?
This will not take place (looking the way it does) for 10+yrs from now!
The thing is we need these types of ai robots to start the Mars mission before humans make their debut on the red planet ...
I think they want these robots so they can goto the moon, because NASA's new space suit won't be ready in time.. if Starship works as expected.
Weeeeeeeeeeeell... this didn't age well.
Apple went to skeumorphics because it needed to attract the masses. There were digital calculators that pretty much look and function like the current iOS calculator way before the iPhone. Skeumorphics have nothing to do with innovation, it's only about accessibility and easing in the less tech-inclined.
as a roboticist, I highly doubt it. even if buys Boston Dynamics he won't be able to pull it off with that form factor, you just can't fit it into a slim for with strong enough power and engines, yet, maybe in a few more years, for sure won't pull it off in that time frame, I think he was just trying to create Hype, I for one am not fooled, and I love Elong Musk and his companies, just saying this is not practically possible yet.
Another simp for Musk. Stop promoting scamers, or you shall be treated as one yourself.
Latest at 2025 they'll overcome Boston Dynamics thanks to their huge AI experience.
What about Apple's laptops? Maybe Jobs was referring to those not the iPad, because it fulfils everything he mentioned in the speech
The D1 chip is a thousand times more powerful then a ps5, new business of selling these chips would work as a new stream of income. Would it ?
BumbleC is awesome 🤖
I need to say to Elon Musk, let him implant my memory in the brain of one of those robots. This body of mine is too damaged.
The iBook shipped in 1999, with optional WiFi, no? So Jobs missed his deadline by about one decade.
hey John thank you again for your wonderful amazing video! :)
I couldn’t finish this vid, if you know, you know!
Only if he could build a convertible tesla
Insta subbed, ludicrous videos!
Correction at 3:30. During AI Day, they said they were using AI for planning and control. Musk also said that you could make a self driving AI better than a human only using traditional software for planning and control, but it would not be as good. One reason mentioned was that doing planning and control without AI in a place like India or South East Asia with very unstructured traffic would be hard to develop and computationally intensive.
ruclips.net/video/j0z4FweCy4M/видео.htmlh20m
ruclips.net/video/j0z4FweCy4M/видео.htmlh34m
I thought they were still doing the planning and control in C++. It seems like they might eventually move planning to an ML model, but it didn't seem like they are doing that already. Am I missing something? Would love to get this corrected in a pinned comment.
I don’t think it’s even a question… of course not
RIDICULOUS ARGUMENTS!
It will ship by 2025
Such a great content
Amazing channel!
You should go on Joe Rogan!
I spent 4 years at Tesla ; I am familiar with trying to hit some of this ambitious deadlines. Its amazing how it does motivate a large team to get alot of things done quickly.
For sure! Humans need stories to get motivated. Everyone wants to be on an epic journey, working toward something bigger. Elon is great at telling those stories. And the most interesting thing is that, telling these (sometimes farfetched) stories is actually the most effective way to make these ideas become reality.
@@JohnCooganPlus I agree, people don't talk about his ability to inspire or more specially tell stories. Especially early on, this was his way to compel people to come work for him.
PepsiBot (looks like pepsiman)
Elon said he would have a prototype by the end of next year...
Yes, and lots of people are saying this is impossible, which I disagree with.
One more important feature during AI day is the high definition map with the representation from AI; this can enable endless use-case for other AI applications. And currently, those map is extremely expensive and prone to being outdated.