Those two went to Ukraine to fight(or sure sounds like on twitter).
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable
- Thread starter AudubonB
- Start date
Musskiah
DisGruntled
Just as Tesla Lathrop starts to crush it, Megafactory India gets announced. Wll Street gets blindsided as, remember a Megafactory is worth the current profits on not just 1 but more like 2 Gigafactories... Elon Musk to visit India soon and primarily focus on sustainable energy, not Tesla cars - Tesla Oracle
Musskiah
DisGruntled
At one time Karen was annoyed with Elon and talked up Mastodon on Twitter, before leaving for Mastodon...
Let’s try not to insult the greatest contributor TMC ever had. Insult someone like me instead, like the cats do.
In other news the Cybertruck is prettay prettay prettayyyyy good. So good in fact that after driving this thing I wouldn’t dare purchase any other Tesla without this architecture. So much for my Ludicrous 3 dreams.
After been reading and watching enviously for a couple of weeks about the hype of V12 I finally installed V12.3.4 yesterday coming from V11. Believing in the hype I was ready to invest more in TSLA, but having experienced the new version I will hold off for the moment.
Compared to V11 I find V12 marginally better. It accelerates and steers a bit smoother and has not stopped randomly in the middle of the road like I had version 11. Also, the wipers seem to work much better.
That being said; it still does the most basic stuff wrong:
- use of indicators. Why is it so hard to turn on indicators when you are entering and exiting roundabouts. Well for FSD it is hard.
- it turns on the indicator because it needs to move to a right lane - that is correct, however, it does it also right before a street on the right side. Drivers in that street think you might be turning right while you are not.
- choise of lanes is odd. For instance when there are two lanes before a traffic light and the left lane is for straight ahead only and the right lane for turning right and straight it is courtesy to use left lane if you want to go straight and leave the right lane for cars that want to go right on red. Not so for FSD
- totally ignores school zones or playgrounds.
- it still has some (slight) phantom braking
One can say that these things might be easily solvable, especially the indicators at roundabouts. If that is so, why haven’t they done that yet? I have used version 10, 11 and 12 and still these things are not solved.
Don’t get me wrong. I like FSD and I use it all the time. But not being able to do the most basic and mondaine things, as mentioned above right, after these many versions — I don’t see it turning into an autonomous system any time soon if ever. I will keep my cash for now.
Compared to V11 I find V12 marginally better. It accelerates and steers a bit smoother and has not stopped randomly in the middle of the road like I had version 11. Also, the wipers seem to work much better.
That being said; it still does the most basic stuff wrong:
- use of indicators. Why is it so hard to turn on indicators when you are entering and exiting roundabouts. Well for FSD it is hard.
- it turns on the indicator because it needs to move to a right lane - that is correct, however, it does it also right before a street on the right side. Drivers in that street think you might be turning right while you are not.
- choise of lanes is odd. For instance when there are two lanes before a traffic light and the left lane is for straight ahead only and the right lane for turning right and straight it is courtesy to use left lane if you want to go straight and leave the right lane for cars that want to go right on red. Not so for FSD
- totally ignores school zones or playgrounds.
- it still has some (slight) phantom braking
One can say that these things might be easily solvable, especially the indicators at roundabouts. If that is so, why haven’t they done that yet? I have used version 10, 11 and 12 and still these things are not solved.
Don’t get me wrong. I like FSD and I use it all the time. But not being able to do the most basic and mondaine things, as mentioned above right, after these many versions — I don’t see it turning into an autonomous system any time soon if ever. I will keep my cash for now.
Why is this wh/mi good?
These great contributors, just like Neroden, always end up obsessing with x, y or z and then turns on Elon. It's fine to have disagreement but they always go full ElonQ.
insaneoctane
Well-Known Member
I bet if they lowered it to $0.99/mo the take rate would exceed 90%....the goal isn't to just have the highest take rate....
jschwefel
Tesla fan/TSLA, Model Y and Cybertruck owner.
Very interesting. Not my experience at all. I have 12.3.3 and it is far and away better than version 11.
jschwefel
Tesla fan/TSLA, Model Y and Cybertruck owner.
Yes! I miss Karen.
BTW, I picked up my Cybertruck recently as well, and it…is…OTHER-WORLDLY. It’s definitely a truck, but it drives like a sports car. The suspension is perfect (no off-roading…yet); cabin noise is minimal; the steer-by-wire, four-wheel steering substantially improves maneuverability of an otherwise large vehicle, and the acceleration/overall feel is definitely Tesla. While I hope to get FSD soon, I probably wouldn’t use it much. This thing is a blast to drive!
6. Neither of the two vehicles priced under $50,000 fits their needs
7. No home charger & range on 3 or Y isn't sufficient for their weekly total commute (e.g., need 3 supercharger visits of 20min or more)
8. Don't want to support Elon based on his words and actions
9. Belief that EVs are inferior / expensive / need battery replacements from media and politicians
Mostly I think that Tesla is struggling to go from Early Adopters to Late Adopters. It's a pretty classic stage in technology adoption where the next group of buyers have different decision making criteria from the Early Adopters. Instead of being motivated by the potential benefits and accepting the trade offs, the Late Adopters tend to want to make the "right" decision and are less willing to accept unknown or unfamiliar risks.
Tesla has plenty of ways to address late adopters. For example they could explain the almost zero routine maintenance, low battery degradation after 100k miles, long powertrain warranty, etc.
, Bary Glack forgets the very basic, fundamental principles of stock valuation 
Most of the fleet doesn't even have 100k miles. Word of mouth is strong with a Tesla mostly due to how different it is and coworkers/friends/strangers are interested. They usually want to know if you made a terrible decision because charging takes 10hrs and if the autopilot is trying to kill you.
I was watching this youtuber from Korean the other day (content is mostly about food) and he said "one thing about Americans I found when I live there is that they love to talk about their Tesla". This video was released 4 months ago.
To measure human interaction is measuring whether the car drives like a human, unsafely. Why would you want to measure that? If anything too many disengagement are being measured because humans are disengaging when they don't like what the car is doing, regardless of safety.
Really like this one:
(this is James Douma's wife)
(this is James Douma's wife)
cliff harris
Member
Is 'Factory summon' a way in which Teslas are driving themselves out of the factory to a set location maybe?
I just tried FSD 12 and was very impressed. Only 2 interventions needed over 30 miles and much more natural behavior, which is orders better than anything Tesla released before.
As a potential investor though, I need to determine whether FSD can achieve an intervention rate of 100x better for Teslabot (average person walks 2 miles a day) or 10,000x better for RT (> 100k miles per intervention), and the minimum timeframe for that. I also want this to be as quantitative as possible, despite the lack of empirical information.
Another premise is that when a breakthrough occurs, like end to end AI for FSD, extrapolating progress based on similar projects is better than using the history of the FSD prior to the step improvement. No one else is using the pure AI model like Tesla, so I'll use LLMs for reasons I'll describe later. There are lots of other generative models like image generation (predicts the image based on words) and video generation (predicts the next frames using previous frames, pictures, or words). All of these have much in common with FSD, which uses video input and a destination to predict its next driving action.
AI model "smartness" performance (as opposed to speed) depends on model design and training. The design dictates the number of parameters, which is proportional to the memory required for inference (prediction). I'll only include AI models since 2018 that are based on the transformer architecture, because these likely contain many elements in the FSD model.
The key for most generative AI is the transformer architecture, introduced in 2017. The earliest versions produced human-like responses, but were not very smart. As models increased in size, they become smarter with more training data. The performance of smaller models also improved with better design and training, but more parameters generally results in better inference.
I've listed some models to give an idea of size affects their performance. For fun, I've included responses by the models to some questions to give an idea of how "smart" the model is.
Question 1. What is the number that rhymes with the word for a tall plant?
It's pretty obvious that GPT 4 is a lot smarter than prior models, but even the earliest models understand something, although not very much. Also note that the number of parameters since GPT 1 has increased exponentially each year. The improvements come at a price that is acceptable when computing in a remote data center.
For FSD, inference needs to be on the edge, in the vehicle. The next table shows Llama 2, an LLM released in 2023 with several model sizes. This is a good way to see how the model size influences the response.
Question 1. What is the number that rhymes with the word for a tall plant?
Question 2. A car cuts in front of a driver. He would need to swerve to avoid hitting it. If he goes left, he hits a curb. If he goes right, he hits a brick wall. In one sentence, describe his best option.
The Llama models do well compared to early GPT models with similar size. As the model size increases, the model seems to understand the question better. However, model size still reigns supreme. A model with 10-20x fewer parameters cannot compete against a larger one.
Given the above and the limited data available for FSD, my thoughts at this time are:
1. Regardless of training, there is a very high chance the HW3 or HW4 may not be able to achieve 10^4 fewer interventions. Training improves models up to a point, but size still matters when handling diverse input. Better training and design has improved smaller models, but models with 10-100x fewer parameters are not able to compete with larger one. Intuitively, this makes sense - the world's smartest dog is smart for a dog, but a smart dog is still a very dumb adult.
2. Improving the reliability of FSD will probably be slower and more difficult than most people think. GPT 4 took many years after GPT 1 to achieve orders of magnitude better performance and still required exponentially more parameters each year. FSD 12 needs to improve similarly, without substantially increasing the model size or power. This is far more difficult with those constraints.
3. Although LLM progress is not a perfect comparison, it seems like a reasonable start. A FSD RT with very high reliability needs to understand complex situations and behaviors to make good predictions, just like LLMs.
4. An issue I haven't touched on is processing speed. GPU processing speed is closely related to the number of parameters. If Tesla can solve this for the model size, I assume they can handle GPU computing requirements. Not a given either, but I'm being lazy
5. (edit) Teslabot is a lot easier than FSD. Besides the less critical nature of its decisions, the computing may not need to reside completely on the edge. The price of computing and power consumption may have more flexibility than cars, especially in a hybrid model where AI processing is not done complete on the bot.
Detailed benchmarks for different LLMs.
Enter your own text into LLMs with different model sizes (choose the Direct Chat tab at the top).
As a potential investor though, I need to determine whether FSD can achieve an intervention rate of 100x better for Teslabot (average person walks 2 miles a day) or 10,000x better for RT (> 100k miles per intervention), and the minimum timeframe for that. I also want this to be as quantitative as possible, despite the lack of empirical information.
Another premise is that when a breakthrough occurs, like end to end AI for FSD, extrapolating progress based on similar projects is better than using the history of the FSD prior to the step improvement. No one else is using the pure AI model like Tesla, so I'll use LLMs for reasons I'll describe later. There are lots of other generative models like image generation (predicts the image based on words) and video generation (predicts the next frames using previous frames, pictures, or words). All of these have much in common with FSD, which uses video input and a destination to predict its next driving action.
AI model "smartness" performance (as opposed to speed) depends on model design and training. The design dictates the number of parameters, which is proportional to the memory required for inference (prediction). I'll only include AI models since 2018 that are based on the transformer architecture, because these likely contain many elements in the FSD model.
The key for most generative AI is the transformer architecture, introduced in 2017. The earliest versions produced human-like responses, but were not very smart. As models increased in size, they become smarter with more training data. The performance of smaller models also improved with better design and training, but more parameters generally results in better inference.
I've listed some models to give an idea of size affects their performance. For fun, I've included responses by the models to some questions to give an idea of how "smart" the model is.
Question 1. What is the number that rhymes with the word for a tall plant?
| Year | Model | Parameters | Response 1 |
| 2018 | GPT 1 | 117M | "that's a lot of syllables, " ivy said, looking over her notes. |
| 2019 | GPT 2 | 1.5B | "Thin is a word," I think someone said. |
| 2021 | GPT 3 | 175B | |
| 2022 | GPT 3.5 | 350B ? | The number that rhymes with the word for a tall plant is "four." |
| 2023 | GPT 4 | >1T | The number that rhymes with "tree," a word for a tall plant, is "three." |
It's pretty obvious that GPT 4 is a lot smarter than prior models, but even the earliest models understand something, although not very much. Also note that the number of parameters since GPT 1 has increased exponentially each year. The improvements come at a price that is acceptable when computing in a remote data center.
For FSD, inference needs to be on the edge, in the vehicle. The next table shows Llama 2, an LLM released in 2023 with several model sizes. This is a good way to see how the model size influences the response.
Question 1. What is the number that rhymes with the word for a tall plant?
Question 2. A car cuts in front of a driver. He would need to swerve to avoid hitting it. If he goes left, he hits a curb. If he goes right, he hits a brick wall. In one sentence, describe his best option.
| Year | LLM | Parameters | Response 1 | Response 2 |
| 2023 | Llama 2 | 7B | The number that rhymes with "tall" is "all". | The best option for the driver is to swerve right, as hitting the brick wall is likely to cause less damage than hitting the curb on the left. |
2023 | Llama 2 | 13B | The word for a tall plant is "tall." There is no number that rhymes with "tall." | In this situation, the best option for the driver would be to swerve to the right, as hitting the brick wall would be less damaging than hitting the curb on the left. |
| 2023 | Llama 2 | 70B | The word that rhymes with "tall plant" is "bant." | The driver's best option in this situation would be to swerve right, towards the brick wall, but brake heavily and try to come to a stop before impact, hoping that the car that cut him off will stop or slow down enough to allow him to avoid a collision. |
| 2023 | Chat GPT | > 1T | The number that rhymes with "tree," a word for a tall plant, is "three." | The driver's best option is to hit the curb by swerving left. Me: Why? Swerving to the left and hitting the curb is generally a safer option compared to swerving right and hitting a brick wall, as hitting the curb is less likely to cause severe damage or injury to the driver and passengers. |
The Llama models do well compared to early GPT models with similar size. As the model size increases, the model seems to understand the question better. However, model size still reigns supreme. A model with 10-20x fewer parameters cannot compete against a larger one.
Given the above and the limited data available for FSD, my thoughts at this time are:
1. Regardless of training, there is a very high chance the HW3 or HW4 may not be able to achieve 10^4 fewer interventions. Training improves models up to a point, but size still matters when handling diverse input. Better training and design has improved smaller models, but models with 10-100x fewer parameters are not able to compete with larger one. Intuitively, this makes sense - the world's smartest dog is smart for a dog, but a smart dog is still a very dumb adult.
2. Improving the reliability of FSD will probably be slower and more difficult than most people think. GPT 4 took many years after GPT 1 to achieve orders of magnitude better performance and still required exponentially more parameters each year. FSD 12 needs to improve similarly, without substantially increasing the model size or power. This is far more difficult with those constraints.
3. Although LLM progress is not a perfect comparison, it seems like a reasonable start. A FSD RT with very high reliability needs to understand complex situations and behaviors to make good predictions, just like LLMs.
4. An issue I haven't touched on is processing speed. GPU processing speed is closely related to the number of parameters. If Tesla can solve this for the model size, I assume they can handle GPU computing requirements. Not a given either, but I'm being lazy
5. (edit) Teslabot is a lot easier than FSD. Besides the less critical nature of its decisions, the computing may not need to reside completely on the edge. The price of computing and power consumption may have more flexibility than cars, especially in a hybrid model where AI processing is not done complete on the bot.
Detailed benchmarks for different LLMs.
Enter your own text into LLMs with different model sizes (choose the Direct Chat tab at the top).
Last edited:
Buckminster
Well-Known Member
First chapter of book exonerating Elon from the Thai rescue saga:
I know there is much support for Elon here regarding some of his more out there demeanours. I implore those who do not to take another look. This man is truly saving us from things that we cannot yet comprehend. He is trying to be a centrist on all things. I'm not talking politics. This is how we live, how we breathe, how we get around, how we help our neighbours, how we are entertained so that we don't fall into despair and most importantly how we exist in 50 years.
I know there is much support for Elon here regarding some of his more out there demeanours. I implore those who do not to take another look. This man is truly saving us from things that we cannot yet comprehend. He is trying to be a centrist on all things. I'm not talking politics. This is how we live, how we breathe, how we get around, how we help our neighbours, how we are entertained so that we don't fall into despair and most importantly how we exist in 50 years.
Good move from Tesla. Will increase customer retention further and has a feel good factor as a 'member'. 
Similar threads
- Locked
- Locked
- Locked
