Tesla AI Director Shows PROOF Tesla Full Self Driving is Near | Brighter with John Gibbs
ฝัง
- เผยแพร่เมื่อ 20 ก.ย. 2024
- Dr. John Gibbs also known as Dr. Know It All on TH-cam reviews 2 presentations from Tesla's Autopilot Team.
Follow John on TH-cam: / drknowitallknowsitall
Follow John on Twitter: @drknowitall16
Tesla FSD Presentation by Ashok
• Tesla CVPR2023 Workshop
John Gibbs DrKnowItall explanation of Ashok's presentation
• ONE MODEL to Rule Them...
John Gibbs DrKnowItall explanation of Phil Duan's presentation
• Tesla Engineer: Tesla'...
❤️ Thank you to my TH-cam Members! Your support means a lot!
❤️ Pls like, subscribe and comment. It's a simple way to support the channel.
🔹 TESLA Accessories - very good quality materials: Get 5% off
jowua-life.com...
🔹 MagBak Wireless Charger : Get 15% off code: HERBERTONG15
tinyurl.com/4w...
🔹 MagBak Wheel Protector (THIS IS AWESOME!) : Get 15% off code: HERBERTONG15
tinyurl.com/3m...
Let's get brighter!
Follow me on Twitter: @herbertong
Need a Product Consultant? As founder of 3 startups and product expert, I consult with companies helping them: 1) Develop new product innovation using a scientific approach or 2) Determine the most ideal go-to-market strategies.
❤️ PLS CONSIDER SUPPORTING THE CHANNEL?
🔹 Become a TH-cam Channel Member: / @brighterwithherbert
Member-only perk: Most Important Graphs and Tables for EV and Tesla
🔹 MagBak Wireless Charger : Get 15% off use code: HERBERTONG15
tinyurl.com/4w...
🔹 MagBak Wheel Protector (THIS IS AWESOME!) : Get 15% off use code: HERBERTONG15
tinyurl.com/3m...
🔹 Amazon Affiliate link for the US: amzn.to/3TyWbkL
I get a small commission for anything you buy from Amazon in the next few hours at no cost to you.
🔹Reserve a CyberLandr: cyberlandr.com...
Tech First Camper for CyberTruck
🔹Open a brokerage account at Interactive Brokers: ibkr.com/refer...
Disclaimer: All content on this channel is for discussion and illustrative purposes only and should not be construed as professional financial advice or recommendation to buy or sell any securities. Should you need such advice, consult a licensed financial or tax advisor. All views expressed are personal opinion as of date of recording and are subject to change without responsibility to update views. No guarantee is given regarding the accuracy of information on this channel. Neither host or guests can be held responsible for any direct or incidental loss incurred by applying any of the information offered. Author is long TSLA and other stocks at time of original video publish date.
#Tesla #TSLA #stock #ElonMusk #investing #teslabot #technology
I love how all the stars of the Tesla TH-cam community join forces and team up on each others’ channels to support the mission.
@@spanishflea634 yea, good point. Even though the bulls' arguments make sense, we need to hear more from reasonable bears, (not bears like Gordon Johnson who lies and twists things). This will keep us more balanced in our expectations.
Tesla's youtube speaker community are fantastic! Here are 2 of my favorites, thanks for all you do, you solidify my convictions and it fun watching you❤❤❤
The A.I. convergence between the Automotive and the Robotic is mind blowing and a very efficient way to cross develop a shared platform, I am impressed.
Finally John said it!!! This isn't just taking thousands of cases and memorizing them. That would be an example of a deterministic solution. The idea is to TRAIN the "computer" how every thing works by showing it thousands of examples so then it will be able, on it's own, to compete tasks "IT HAS NEVER SEEN" because now it is smart!
THANK you HERBERT and JOHN 🙏
Great chat, of course. Thank you for this. When your mind imploded 24:41, something I wish John had covered was that they can take sections of the real world and staple them together. They can create virtual towns and cities consisting of intersections and stretches of roadways from different parts of the world (but not in the same place). With Dojo, they can simulate driving through real places trillions of times a day.
Herbert is a superb content creator, he brilliantly enhances our knowledge base. Thanks Herbert!
Incredible discussion, and I agree with Herbert, FSD is making incredible progress and will be here sooner than most folks think!
Thanks Herbert, great video as usual !
When the MY stops pulling to the right during a pass of every on ramp (not off ramp) I'll know this is finally working. Not there for sure!
Thanks Herbert and John, I really learned a lot. That was amazing so I took notes and need to go over it again. So strange that what method is good today is no longer good tomorrow. They say if it ain't broken don't fix it. How many more times will the engineers find that what FSD method is best is really a never ending story of new methods. So interesting for sure.
We need to be patiently optimistic but you guys are a great team for bringing the reality of our current situation forth to investors! Great team work and well presented!
This is by far the best discussion and analysis of the new FSD operating systems/computer/dojo interaction. Herbert, you did a great job picking he key points of the Tesla presentation and then John did a fantastic job of explaining in detail what each segment was showing and how it will work in FSD system. I learned so much vs any other discussion I have seen. OUTSTANDING!! Thank you so much. Herbert, you are a terrific facilitator. And your knowledge compliments the guests you interview. Impressive.
THANKS HERBERT 🤗 JOHN,FOR DISCUSSING FSD AND THE PROGRESS 🤔🤯🤯
ESPECIALLY GOOD OF YOU TO KEEP BUSY…WITH ALL THE PROBLEMS YOU ARE FACING 🤗💚💚💚
Thanks Herbert and Dr Gibbs for this video. The more I watch FSD tech content the better I appreciate it even more. Dr Gibbs is a great source of information to teach us just about any subject. Good job. :)
Most fantastic video from both of you!! Never heard better explanations!
Not so much predicting the future but "modeling" it with different possibilities. Like when I drive a car I imagine a car suddenly cutting in front of me or running a red light or a person suddenly stepping out into the crosswalk. "Watch out for the other guy." Our medulla remembers all such prior events and pops them into my mind as fear. That allows me to prepare to hit the brakes or take some other evasive maneuver. If the possible outcome happens, I'm ready for it and can react much faster. Of course, our minds are much slower so we see a snake when hiking that is actually a root, or a spider that is actually just a dark dried flower. Our mind is so cautious of danger, it immediately interprets things in the worst possible way just to be safe. That can make us skittish and do foolish things, but better safe than dead. Fortunately, a computer can analyze things much faster and figure out what they are before having to react.
Modeling reality is exactly the same thing as predicting the future. You set up the parameters for the current situation, run the model, and the resulting data from the model is the prediction of future.
Our minds are wonderful computers but one flaw is when we are reacting to emergencies our minds skip through small details and surprises us that our reactions are slower or over the top if we don’t find that calm middle ground not to hit someone or object etc.
@@hwillia204 Unfortunately, we didn't evolve driving cars or using guns or things like that. We are highly emotional creatures, not really logical. Professionals can train themselves how to deal with extreme situations, accidents, combat, etc. Most drivers lack any capability to do so and impulsively do crazy wrong things or nothing at all in an emergency.
Peering into the mind of a visually generative ai, what humans call imagination.
FSD is very close.
@@TeamAgainstWild Humans do create visual realities in their minds, modeling how they might be using pieces of actual experience or things they have seen in media. So yes, the process is similar. However, human imaginations are backed by and shaped by emotions, dreams and nightmares, things we would like to see and fear to see and we try to learn from them all. AI can only offer reflections of human emotion built into its database. Although I have no doubt that AI won't soon be able to create fantasy images and horror-invoking ones that surpass what humans have created.
Great discussion guys. The future looks brighter and brighter!
Wow ! Did I get brighter ? ABSOLUTELY
Herbert asks the best questions indeed!!
This was great! I went back and watched the presentation by Ashok, then watched this again and it sank in... Wow! Thanks to you both!
TOP CLASS CONTENT. Many thanks for putting this together for us all.
When I get in the car to go to Lowes I have a full visualization of the the ride there to the build ing parking lot. This just happens, it is just there. Is this how this computer works
Great to watch.Thanks. In some ways, Tesla safety systems are already demonstrably better than humans - correcting incorrect pedal presses (like pressing the accelerator instead of the brake), and not just stopping for cross-traffic (someone jumped the lights), but actively swerving to avoid that rogue driver...
SO good guys! Thanks Very much for helping us housewives understand just a little bit of the Tesla magic!
A major problem I see is comes from the old dictum 'Garbage in, garbage out'. The cameras on the Tesla get 'blinded' by the sun, and 'blocked' by darkness so how can the FSD predict correctly the future if it can't see the past in these situations?
I’m blown away. Sounds not like artificial intelligence but artificial consciousness…
The car will slow down for speed bumps and potholes obviously, car-car allows real time interaction muffler in road ahead accident in progress ahead, this will save so many lives!
Thank you so much for explaining the future of Tesla and the self-driving.. I use it as much as I possibly can 95% plus of the time. It amazes me how far it's advanced in the small amount of time. I've had it for roughly 1 1/2 years. It's incredible. This is so amazing and we're seeing it happen right in front of our own eyes, and driving it what an amazing experience. Thank you for sharing!! ✌️ Great video keep them up!!
excellent!!
Patiently waiting for FSD to improve. Nice to handle the small stuff that have been ignored. I know the engineers and computer has bigger fish to fry to resolve FSD. Will be nice
Can you do a talk about car car- communication? Would be awsome!
Fantastic episode!!!!!
Amazing interview must watch for all Tsla holders!! Thank you for explaining the progress of AI in Tesla. 🙋🏻♂️😁👏🏽
This was really good! Thanks.
Whaaaaat! It's like a predictive AI on the fly? 😮 THIS is a game changer folks
Wow! I learned something new today!, what I don’t understand is how it is possible that all the data that is created from learning with those huge supercomputers, is going to fit in a small computer in a car
What concerns me when I drive with FSD, is that I am not sure if it can see far enough ahead to decide what to do.
The car is NOT Dojo. Where does the data reside? How does it get to the vehicle and when?
Is FSD going to slow down when it sees fog ahead? I think another 2 years until high resolution RADAR is integrated to eliminate the 1-2% of the risk.
The software counts photons rather than just seeing as we do, the AI understands what those photons mean i.e. it will understand that it is seeing fog (as it does the smoke in this video) but it will be able to see through it better than we can so will drive "according to the conditions" as we do. But it will be better than us even without radar = safer!
Great discussion, thank you!
Can any video in the world of the physical world now be used to help the Tesla bot software learn what the world is like?
Thanks!
Thank you thank you William!
so considering the analogy to chat GPT used in the video real world so to speak Will this enable a similar situation where after Tesla has solved the problem a subsequent company could easily replicate for much cheaper a la "alpaca" relative to ChatGPT for say $600? perhaps no without the data set and the underlying foundational structure but it's worth considering.
A very good question! I will need to ask James Douma
@@BrighterwithHerbert definitely! look forward to hearing his take
visualisation jitter can certainly be tuned independently of how assured is the perception model
John, is you're idea that by combining ground truth and generative situations to challenge the driving AI in a sort of adversarial reinforcement training? Where the generated derivations of ground truth get harder as the driving AI improves?
Awesome! Thank you both.
I really like your show, but the video quality is super bad.
I live in Sweden and have 10 Gbit/s (that is overkill, but we build our own fiber network, so...) I think 50-100 Mbit/s would be enough to get 1080 or 4K.
What kind of connection do you guys use!?
Around 25 min, the discussion about the generated video is 100% generated. But the generated video could based on 1. the past video. 2. The prompt (Go straight or go right). 3. Also could include the geo information and all the what's about the roads and maps aheads within the "MATRIX".
Herbert, there is a better quality video directly from the “WAD at CVPR” TH-cam channel but I am unable to post a link here.
Awesome - informative - mind-blowing!!!
2029 for superhuman AI is beginning to look like a slam dunk.
Drove a 120 miles today Central pa to Townsend md. And back. I have Been using FSD a year and a halve. I was gee wiz,but the last 6 to 8 months feels every FSD glitch fixed a new or old appears or reappears. Route 83 has nonexistent merge ramps. Despite real time information the system should be utilizing. The buggieness is confusing other drivers appearing to allow a merge and then just being in the way,as the mini ramp ends. This road is heavily used and has remained unchanged for 60 years especially where is FSD is struggling the worst. Confusion… the wheel jerky jerkies and all but stops till I takeover and this is off the same main road but the intersections are expansive and of recent construction. We called this behavior on the farm. A calve looking at a new gate phenomena. 15 takeovers 7 absolutely required and 2 being serious to change my heart rate.
i wouldnt even bother using it if i had to intervene regularly. may as well just drive as sit there waiting.
And the system cannot even read speed limit signs in construction zones or on local roads with non standard sign sizes. It also still blindly follows white lines into a possible crash unless you intervene.
My best understanding of neural networks and FSD is that they are magic.
I am so lost that I can’t even frame a sensible question.
Ditto! But I do know when I’m hearing people who know what they’re talking about.
VERY few of us gets more than a fraction of it, safe to say. I still learn a little more every time I tune into these discussions from relative (to me) experts.
Science fiction writer Arthur C. Clarke formulated his famous Three Laws, of which the third law is the best-known and most widely cited: “Any sufficiently advanced technology is indistinguishable from magic”.
@@darylfoster6133 Well said. 👍🇺🇸
@darylfoster6133 very similar names we have...
This is all well and good with the video projection of actual road scenes for FSD, but why can't they even fix the "Phantom Braking" so we can enjoy the Basic Auto Pilot feature????
Phantom Braking is definitely a very very big safety issue and incidents have happened because of this problem. I think FSD will have the same issues while driving.
So what do you people say about this problem instead of how great AI can improve FSD???
Today I was driving my Model 3 had my hands on the steering wheel can’t remember if FDS was on but all of a sudden on this sunny day in Vancouver my wipers with fluid spraying out cleaned my windshield. It wasn’t really dirty. Has anyone had that experience? Maybe I knocked the stick but don’t think so. Weird!
Until now, I thought Elon was still on Elon time. Now... i think FSD will be this year.
Incidentally, the human brain causes us to "see" moving objects not where they are now, but where our brain predicts the object will be 0.2 second from now. This is because our typical reaction time is 0.2 second, so we need a little edge.
I am still waiting for my Model Y FSD to stop taking every turn lane while driving down a county highway or small town street.
Makes me wonder if this will be based on mapped roads or if it will actually drive correctly on a road it never went down before?
12.4.4 was such a huge regression for my car.
As always, two steps forward, one back (avg. 1.5 forward). Progress requires painstaking patience.
I think you mean 11.4.4? Hell, I can’t keep track anymore, and never use FSD. I’m just waiting to see if it’s ever actually useful.
Yep, that is a weird new behavior.
@@thefish5861 yes 11.4.4 oops...
But could it do frustum culling? The video is nice and steady but the buildings kinds drops into place.
Your either an optimist or pessimistic about your progress when solving a problem. Both Hebert and John are correct in their assessments!
Tesla plays the sandbagging game, there’s so much going on behind closed doors it will blow our minds .
Hope so, but I’ve been hearing this stuff for years and years. It feels like we’re always waiting for something huge that never materializes. Hell, I’ll be satisfied if they ever build that damn pickup truck.
Every time your "Thank you for watching" clip comes up in the middle of your videos, I think it is the end.
I don’t see how every road intersection in the US is modeled like this and how the system computes real time in the car without the super computer. You can’t have all this detail in the on board hardware and software?? Works on the super computer but can’t translate to real time on board functions??
THANKS HERBERT,NOW I’M GETTING MUCH BRIGHTER 🤯😎💚💚💚
I may misunderstood some. But generative Model can predict next few seconds video does not mean it understand the current text without labeling. I think the labeling technology including auto label are here to stay and still be crucial part of training. I don’t see it being dropped
it doesnt jitter...fantastic
You got the AlphaGo story confused with AlphaZero.
And the figures running around wasn‘t OpenAI it was DeepMind.
I would love to have Tesla release an app for Apple Vision Pro or other headsets to see these visualizations.
Maybe Tim will go to Austin and bow before his master!
And we know the future never changes.
Like a future "Time Machine Generator".
Herbert, I am confused. You are raving about Tesla being able to precisely map certain areas. What’s the difference in outcomes compared to LiDAR everybody else is using?
Another question: In the Q2 2023 Update Tesla states "This month, we are taking a step towards faster and cheaper neural net training with the start of production of our Dojo training computer." I'm not sure if I understand the meaning with "start of production" clearly. Also different news sites interprete it differently. Is it meant that Tesla starts builiding the Dojo supercomputer in july '23 or does it mean, Tesla start using the Dojo supercomputer in production in july '23?
The 3D scene shown at the beginning of you presentation is generated with 3D photogrammetry. I have done lots of that professionally. What isn’t clear is how it is being used. What amount of this data does the car get? I’ll bet none. I suspect these datasets are just used for training on Dojo or Nvidia servers. The biggest issue I have with FSD is that there is no localized learning. All learning is done in a general sense and so the software wakes up in a new day every time you leave your driveway. My car refuses to turn left out of my driveway despite having done it a hundred times. There is no way of teaching FSD to back into my garage or know what is the fastest lane at a certain point. Maybe they will layer localized training later.
It is not photogrammetry, but the appearance is similar. If you watch the video from Ashok from WAD 2023 CVPR, he explains that it is similar to NeRF (Neural Radiance Fields) but is actually created by the system using multiple video streams from many Tesla vehicles using what he refers to as a Voxel population algorithm. A voxel is a three dimensional version of a pixel.
Its key important feature is that it determines whether or not the voxel is empty or occupied by a real object, which allows objects to be tracked in real-time. It means that FSD version 12+ in your car will be representing the surrounding environment in this exact way. No official word, but it is doubtful that Hardware 3 has enough computing power to do this and that this is why HW 4 was created. The new DOJO computing center at Tesla has just come online and will do the labeling of objects and street details in real time.
@@tech-utuber2219 Thanks. I must admit that I don’t see how an individual pixel can be 3D. There is no depth information in a pixel so the only way you can calculate a depth is to correlate it with another pixel from another camera or from a small time delta. However, I will check the reference you pointed out.
@@monsenrm he explains that the actual 3-D size of the voxel is variable, so it is not directly the same as a pixel, and I was just explaining what the word means. The representation of other cars for example, can be built up from a smaller number of large 3D blocks, or a more detailed version with much smaller voxel blocks. In Ashok’s presentation video, buildings are constructed from very large pieces, which is fine since they do not move.
@@tech-utuber2219 I get that they have a method for determining a voxel / cube in 3D space. What I don’t understand is how they get depth information. There are 3 ways of getting depth. The first 2 are accurate and the last one is approximate. So you can use LIDAR or photogrammetry to get solid depth measurement. We know they don’t use LIDAR. The third way is to “label” a known object like a car or truck and then calculate depth from its relative position in the FOV. That video showing a detailed street with trees and all sorts of details could not be done with that accuracy using the third method. I still maintain that the detailed video was done using photogrammetry but is only used for training.
@@monsenrm Your depth measurement question has a straightforward answer. Let’s use the photogrammetry notion which we are used to in order to point out one thing which you forgot, which is that a Tesla vehicle with eight cameras on it is a precise 3D camera array, in which all the dimensions are known and are fixed in space with reference to the vehicle itself. The optical cones which project out into the surrounding spatial volume from each sensor (plus lens) are accurately specified so that an almost (not under the car) complete 360°, 3D view from the camera array can be re-created, and all details can be distance and depth measured because of the known optical geometry.
This means that you could use a Tesla vehicle as a photogrammetry capture unit and drop it into any place on earth and re-create a 3-D model of the surrounding spatial details with an accuracy which is only limited by the visibility and resolution of the camera array. Since they built the vehicle body and placed all the sensors on it and specified all the lenses for the sensors, the optical measurement accuracy is built-in. A normal photogrammetry capture, in which we use a handheld camera or a flying drone with a camera and move about the object of interest in order to capture multiple images at different locations, uses only a SINGLE camera. Think of a Tesla vehicle as a driving drone instead of a flying drone with multiple cameras instead of just one. So far, this defines object placement with respect to the vehicle, so you only need to add the GPS map location of the vehicle to get an absolute world position instead of a relative one.
Ashok describes their approach as a hybrid of NeRF as opposed to straight photogrammetry which we are used to when creating a quasi-3D mesh of a static object or scene to import into a game engine such as “Unreal” or “Unity”. In photogrammetry, we take multiple digital photos and the “stitching” code will map the 2D photos into a mesh. I write quasi-3D, because in a game engine such as “Unreal”, this mesh is a bunch of surface triangles with color textures mapped onto it specified in U & V coordinates so that Lumen (or ray tracing) can light the mesh object with photo realistic results.
Instead of working directly with a surface mesh, their custom code stack creates a voxel occupancy representation, which can be queried as to whether the voxel is an empty cube of air, or part of an object that must be tracked with respect to the moving vehicle. The irony is that he specifically states that mapping out the lanes and road markings are much more challenging (he said obnoxious) than your question due to uncertainty and complexity. As he points out, different people can look at the same street layout, and not agree as to details of the street lanes. They create a polyline representation of the street geometry which can be tokenized (not unlike word tokens in chatGPT) and need to do it in tens of milliseconds so that FSD can function in real-time. He also said that the labeling of street data is done off-line.
It is not clear to me how much computation is done by the Hardware 4 computer system in the vehicle versus “DOJO”. It appears that HW-3 is maxed out with this FSD task versus HW-4, hence the upgrade. Still missing is the general ability to read all street signage and identification of emergency vehicles and school buses for example. In addition, I don’t see a way to do full autonomy without also understanding and predicting human gestures and behavior from law-enforcement officers, emergency first responders, and children. Optimus the robot needs this even more.
So for cruise and waymo, if smoke and papers and falling trees were thrown at them. Will they be able to figure out the next step of action?
But the cars don’t hac]Ve the hardware they need for FSD to come out of beta. These cameras don’t view what’s in front of my from bumper. How am I ever going to trust it when it doesn’t even know if I am safe in my garage
Can you imagine all the Street View road data that Google must have, or potentially have?
How does FST handle snowy roads, blizzards?
The was a video out months back of a drive along a snowy tree lined road, just banked up snow each side, the car drove it without any problem.
I have seen about 5 companies that are at this point in their development. It seems dramatic only to people who are only following tesla, and not the whole industry.
How many of those are car makers, and build their own hardware?
I have FSD visualization, and when a car is right behind me it flips between being on screen and disappearing every second. Not convincing at all ...
The moment that we realize what this new world model means it's like a dam breaking. The rush of ideas and the understanding of how much this world model opens up a universe of possibilities that never existed before makes your brain freeze from the overload. It will take a long time for us to even dream what this means. Meanwhile the computers running in the background will be learning our world as never before. This model is like a baby being born and discovering the world around it. It will not just be thought the world It lives in it will DISCOVER the world it lives in. That is a different periderm of understanding.
Buying $YSLA every other week… DCA my way to fortune… 🎉🎉🎉
Regulation aside, this sounds like a better proposition for applying to other countries.
48 min mark, real time coms to EVERY equipped vehicle land sea air subter
I come from a family that practice Martial Arts. I got into big trouble with my Mom after I told her what I did. I was messing around and sent a burst of Chi energy to my son who was playing video games. He thought I was behind him messing with his hair so he slap the back of his head. He turns around and sees me about 40 feet away. My son said do it again and I did. My son said wow that's weird. When my Mom found out, she said that dangerous and don't ever do again.
So I was wondering if my body is like a computer that has electrical signals. I wonder if I can play against another computer and we can play this game of sending Chi. Would this be possible for the computer to figure a way to do the same.
So what I'm understanding is that all of the work Tesla has done on self driving has been tossed out and now they are starting from scratch with a whole new method?
There is going to be no other company in the world like Tesla. By this time the rest of the companies realize they are so far behind that they can't even figure out how to proceed forward. It's astounding how fast Tesla is innovating the AI technology. The rest of the competition is rapidly falling to the wayside and don't even know what's going on. This is getting very very interesting.
If this eventually were to be licensed out.. How can it work with other automaker's system and make it compatible as they r not running on tesla hardware. How would it interpret and affects it's action?
I loved seeing Herbert’s mind blown. I watched the footage before watching brighter and didn’t understand the generating future until watching this version. This time, my mind was blown, but before Herbert’s was! Ha.
Does the future image depend on the past speed of the vehicle?
I have no doubts that it will happen and with AI it will happen much faster!
I'm hoping. I'm about to retire and put most my liquid assets in TSLA some 2 years ago, only to see the stock take a sh*t. I have a strong conviction but when you get older, it's even tougher to handle.
A computer that can learn by learning without human intervention and a learning no human will understand.
This is big news!
With core expertise in battery electric vehicles, AI, power electronics, and robotics, Tesla is capable of developing novel transformer architectures....As in the Transformer movies! 😋
10-4-2024. FSD comes out of Beta. Wide release to fleet. First Robo Taxi ride in Q1 2025.
Now say that in the voice of Sarah Conner
Stil believe we are a couple hardware iterations away?
Guys Guys, Tesla has Gamified Reality ! That is the most significant achievement here. It is VR backwards, Gamified Reality, so whatever the camera system "Sees" is translated into a digital version, Gamified Reality. The Resolution will continue to improve especially with Dojo processing at higher frequencies.
Amazing show. Thank you. One question - why is Dr know it alls video such poor quality when Herberts is all clear?
As near as fusion energy...
I spent a life time in Martial Arts and Qigong allows me to play fight with an equal opponent. We close our eyes or blind fold ourselves and I can predict his moves before my opponents react. Sounds crazy but this is the level I have reached.
Most people will think you ARE talking crazy and dismiss your comment because they cannot comprehend thus accept what you know as they do not have your lifetime of experience. Get into Tai-Chi and ask your instructor about energy exercises! ;-)