MTL_HABS1909
Active Member
Gotta love plausible deniability
-
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
LOL My first thought was SpaceX for sure but on 2nd thought, I'm not so sure.
Does literally going to the Moon make any business sense? I can think of lots of possibilities but I doubt any of them really make sense for where humanity is right now. Maybe he does mean TSLA.
anybody else think this sounds like the removal of the home button and of the headphone jack? We don't know what's best for us until we get it?
Well NASA is planning to build a moon base this decade. SpaceX might be helping them with that, thus planning some moon missions soon?
Knightshade
Well-Known Member
Or they are someone who actually understands that AI is not magic.
The best AI in the universe still can't make vision see through a solid object to what is beyond it like radar can, or make vision see through fog and snow that blocks visual wavelengths but not radar wavelengths.
They aren't removing the radar because AI has magically made the cameras capable of X-ray vision- it's because it's too hard to continue trying to integrate with the new 4D BEV vision system and they feel it's not worth the extra effort for the marginal additional info. See Prunesquallors reposted elon remarks on that.
That might well be the absolute right path to go down.
But we know for a fact FSD has gone down several wrong paths before (including in 2016 when Tesla and Elon loudly announced Radar was awesome and should be a primary sensor specifically because of its ability to provide data vision physically can not provide
They've now decided THAT was a wrong path they went down and are trying another thing.
Teslas willingness to do that is a major reason for their pace of innovation.... rather than, say, Waymo having spent 10+ years now going down the WE ARE SURE HYPERLOCALIZED MAPS AND LIDAR IS THE WAY and still only operating RTs in one tiny geofenced area.
But it also means you can't automagically assume each new path is FOR SURE THE RIGHEST BEST ONE, because it might not be.
Tesla is fine accepting that possibility- if it turns out wrong, they'll try something else.....
Some superbulls are... less fine accepting that possibility than Tesla themselves though.
James Douma already covered this in a video with Dave Lee.... where he told us the same thing I've said here.
LIDAR adds no value to a vision system.
Radar absolutely does add safety and value as a secondary sensor in conditions of low visibility and things like bouncing under cars but should only be used in the occasional situation like that- not as a primary sensor which should be vision.
But the goal is many times safer than a human, not just "any" safer right?
I literally posted a video of it preventing a Tesla from crashing into another car at speed on a highway because the radar saw the stationary vehicle 2 cars ahead the cameras could not.
That's not theoretical like your math- it actually happened.
It's likely it has happened many times, though not every time did the driver save the video and spread it around the internet.
Now does it happen often "enough" to be worth preventing? Some would argue 'at all' is worth preventing- especially since it's using a sensor already on the car.
But both the suggested reasons for the change I cited from the other thread- And Elons own words in the tweets Prunesquallor quoted- essentially boil down to integrating radar with the new 4D/360 vision system is REALLY hard, and they don't consider the # of cases where it might make the car safer worth delaying further development to make radar integration work properly with it.
There's every chance that math is totally correct... if getting a much-safer-than-humans FSD out sooner means forgoing a theoretically-even-safer-but-they-have-no-idea-how-long-it-would-take-to-make-work FSD that keeps radar that might well be a tradeoff worth doing.... and if research on radar 4D integration gets better in the future they can always revisit.
Anyway I'd again point out there's an entire section of the forums specifically FSD related if folks genuinely want to get deeper into the weeds of why this isn't a case of "Cameras just do everything radar can thanks to AI!" which isn't factually correct versus why this still might be the CORRECT decision even though it's removing data the cameras can't fully replace.
Whole forum for it, with multiple threads getting into it, here:
AI, Autopilot, & Autonomous/FSD
Discussion about AI, Tesla Bot, Tesla Autopilot (AP), the promise of Full Self Driving (FSD), as well as other Autonomous Vehicles.
teslamotorsclub.com
It's weird you'd quote my post where I specifically say having millions of additional people experience the cars adds a lot of value and then claim I said it was a dumb idea.
The text you quote cites 2 potentially significant value-adds for the company- both the exposure at all, and the fact it claims Tesla will be building a larger people-mover vehicle (the 16 person vehicle) yet somehow the only part you appear to have actually read was pointing out going 35 mph through a single-lane tunnel wasn't especially futuristic.
Man, 100% wrong twice in the same post. At least you're consistent
Last edited:
So...when did naked shorting start occurring? Would a new, reputable head of the SEC, that does enforcement of the rules, change the act of this behavior in the market?
Edit: Also, why does the Wikipedia page for Naked Short Selling have no historical artifacts after 2014?
Edit: Also, why does the Wikipedia page for Naked Short Selling have no historical artifacts after 2014?
Last edited by a moderator:
Or out the back window whilst driving by SEC offices?
I was joking about LIDAR, but thanks for dropping knowledge on Douma radar etcOr they are someone who actually understands that AI is not magic.
The best AI in the universe still can't make vision see through a solid object to what is beyond it like radar can, or make vision see through fog and snow that blocks visual wavelengths but not radar wavelengths.
They aren't removing the radar because AI has magically made the cameras capable of X-ray vision- it's because it's too hard to continue trying to integrate with the new 4D BEV vision system and they feel it's not worth the extra effort for the marginal additional info. See Prunesquallors reposted elon remarks on that.
That might well be the absolute right path to go down.
But we know for a fact FSD has gone down several wrong paths before (including in 2016 when Tesla and Elon loudly announced Radar was awesome and should be a primary sensor specifically because of its ability to provide data vision physically can not provide
They've now decided THAT was a wrong path they went down and are trying another thing.
Teslas willingness to do that is a major reason for their pace of innovation.... rather than, say, Waymo having spent 10+ years now going down the WE ARE SURE HYPERLOCALIZED MAPS AND LIDAR IS THE WAY and still only operating RTs in one tiny geofenced area.
But it also means you can't automagically assume each new path is FOR SURE THE RIGHEST BEST ONE, because it might not be.
Tesla is fine accepting that possibility- if it turns out wrong, they'll try something else.....
Some superbulls are... less fine accepting that possibility than Tesla themselves though.
James Douma already covered this in a video with Dave Lee.... where he told us the same thing I've said here.
LIDAR adds no value to a vision system.
Radar absolutely does add safety and value as a secondary sensor in conditions of low visibility and things like bouncing under cars but should only be used in the occasional situation like that- not as a primary sensor which should be vision.
But the goal is many times safer than a human, not just "any" safer right?
I literally posted a video of it preventing a Tesla from crashing into another car at speed on a highway because the radar saw the stationary vehicle 2 cars ahead the cameras could not.
That's not theoretical like your math- it actually happened.
It's likely it has happened many times, though not every time did the driver save the video and spread it around the internet.
Now does it happen often "enough" to be worth preventing? Some would argue 'at all' is worth preventing- especially since it's using a sensor already on the car.
But both the suggested reasons for the change I cited from the other thread- And Elons own words in the tweets Prunesquallor quoted- essentially boil down to integrating radar with the new 4D/360 vision system is REALLY hard, and they don't consider the # of cases where it might make the car safer worth delaying further development to make radar integration work properly with it.
There's every chance that math is totally correct... if getting a much-safer-than-humans FSD out sooner means forgoing a theoretically-even-safer-but-they-have-no-idea-how-long-it-would-take-to-make-work FSD that keeps radar that might well be a tradeoff worth doing.... and if research on radar 4D integration gets better in the future they can always revisit.
Anyway I'd again point out there's an entire section of the forums specifically FSD related if folks genuinely want to get deeper into the weeds of why this isn't a case of "Cameras just do everything radar can thanks to AI!" which isn't factually correct versus why this still might be the CORRECT decision even though it's removing data the cameras can't fully replace.
Whole forum for it, with multiple threads getting into it, here:
AI, Autopilot, & Autonomous/FSD
Discussion about AI, Tesla Bot, Tesla Autopilot (AP), the promise of Full Self Driving (FSD), as well as other Autonomous Vehicles.
It's weird you'd quote my post where I specifically say having millions of additional people experience the cars adds a lot of value and then claim I said it was a dumb idea.
The text you quote cites 2 potentially significant value-adds for the company- both the exposure at all, and the fact it claims Tesla will be building a larger people-mover vehicle (the 16 person vehicle) yet somehow the only part you appear to have actually read was pointing out going 35 mph through a single-lane tunnel wasn't especially futuristic.
Man, 100% wrong twice in the same post. At least you're consistent
lafrisbee
Active Member
AND, as a human driver I was taught to off-center my car to the car in front of me so I can see a brake light of the car in front of my car, and to base my decisions more off of the situation in front of the car in front of me than the brake lights of the car in front of me. I have also learned through my life to gauge the driving style off the condition of the car and driver in front of me before I can witness their style. Such as proper running condition of the vehicle and if it has a "Hillary is my co-pilot" Bumper sticker. My truck has a "My Dog is Smarter Than your Honor Student" bumper sticker, make of that what you will.
I wonder how much FSD already does as to positioning the vehicle to optimize the sensors. I am sure it does to some extent. But does FSD read bumper stickers? Or recognize to give a car some additional space when they throw a beer can out from the driver's side?
SunCatcher
Member
OK, hold on here. He didn't say "going to the moon", he said "going to moon". But he didn't say who he was going to moon. I'm confused.
Surely Elon wouldn't moon us. So he must be planning to moon them. Right?
Honesty, I bet if someone who actually has twitter followers some of these taunts, I bet Elon would retweet. I’m sure he has lots of German Twitter followers that don’t like being second place.
zzziii
Member
Here is a quite interesting 18 min video from CNBC on the work that JB Straubel is doing with Redwood Materials. It covers quite a few insights from the company and has details from customers/partners (Panasonic and Amazon) and mentions some the expansion plans.
Inside Redwood Materials, former Tesla CTO’s effort to recycle batteries for rare components
Krugerrand
Meow
Certain topics, like the eight we’re experiencing now, do - on the surface - seem mundane at best. However, the participants reveal much about themselves, least of which their knowledge base and how deep down the rabbit hole they will go embracing their ignorance. It’s actually quite fascinating human behavior and it does help to reveal who can be ignored without risk of missing some key Tesla investing information, but more importantly who should not bring children into this world.
Conversely of course, it also reveals those with intact grey matter of intelligent, sensible, logical et al function. That’s also helpful.
Elon's stating he's going "to moon", there's no "the" in the text. I guess you all know what that implies
Had to LOL at this:
Krugerrand
Meow
Isn’t moon supposed to be about proof of concept; a base built as a test run for Mars, also a refueling and supply depot, an opportunity for people to try out space before full commitment and a revenue stream in terms of ‘take a trip to the moon/vacation on the moon’.
Krugerrand
Meow
I have a sock for you. It’s even clean.Or they are someone who actually understands that AI is not magic.
The best AI in the universe still can't make vision see through a solid object to what is beyond it like radar can, or make vision see through fog and snow that blocks visual wavelengths but not radar wavelengths.
They aren't removing the radar because AI has magically made the cameras capable of X-ray vision- it's because it's too hard to continue trying to integrate with the new 4D BEV vision system and they feel it's not worth the extra effort for the marginal additional info. See Prunesquallors reposted elon remarks on that.
That might well be the absolute right path to go down.
But we know for a fact FSD has gone down several wrong paths before (including in 2016 when Tesla and Elon loudly announced Radar was awesome and should be a primary sensor specifically because of its ability to provide data vision physically can not provide
They've now decided THAT was a wrong path they went down and are trying another thing.
Teslas willingness to do that is a major reason for their pace of innovation.... rather than, say, Waymo having spent 10+ years now going down the WE ARE SURE HYPERLOCALIZED MAPS AND LIDAR IS THE WAY and still only operating RTs in one tiny geofenced area.
But it also means you can't automagically assume each new path is FOR SURE THE RIGHEST BEST ONE, because it might not be.
Tesla is fine accepting that possibility- if it turns out wrong, they'll try something else.....
Some superbulls are... less fine accepting that possibility than Tesla themselves though.
James Douma already covered this in a video with Dave Lee.... where he told us the same thing I've said here.
LIDAR adds no value to a vision system.
Radar absolutely does add safety and value as a secondary sensor in conditions of low visibility and things like bouncing under cars but should only be used in the occasional situation like that- not as a primary sensor which should be vision.
But the goal is many times safer than a human, not just "any" safer right?
I literally posted a video of it preventing a Tesla from crashing into another car at speed on a highway because the radar saw the stationary vehicle 2 cars ahead the cameras could not.
That's not theoretical like your math- it actually happened.
It's likely it has happened many times, though not every time did the driver save the video and spread it around the internet.
Now does it happen often "enough" to be worth preventing? Some would argue 'at all' is worth preventing- especially since it's using a sensor already on the car.
But both the suggested reasons for the change I cited from the other thread- And Elons own words in the tweets Prunesquallor quoted- essentially boil down to integrating radar with the new 4D/360 vision system is REALLY hard, and they don't consider the # of cases where it might make the car safer worth delaying further development to make radar integration work properly with it.
There's every chance that math is totally correct... if getting a much-safer-than-humans FSD out sooner means forgoing a theoretically-even-safer-but-they-have-no-idea-how-long-it-would-take-to-make-work FSD that keeps radar that might well be a tradeoff worth doing.... and if research on radar 4D integration gets better in the future they can always revisit.
Anyway I'd again point out there's an entire section of the forums specifically FSD related if folks genuinely want to get deeper into the weeds of why this isn't a case of "Cameras just do everything radar can thanks to AI!" which isn't factually correct versus why this still might be the CORRECT decision even though it's removing data the cameras can't fully replace.
Whole forum for it, with multiple threads getting into it, here:
AI, Autopilot, & Autonomous/FSD
Discussion about AI, Tesla Bot, Tesla Autopilot (AP), the promise of Full Self Driving (FSD), as well as other Autonomous Vehicles.
It's weird you'd quote my post where I specifically say having millions of additional people experience the cars adds a lot of value and then claim I said it was a dumb idea.
The text you quote cites 2 potentially significant value-adds for the company- both the exposure at all, and the fact it claims Tesla will be building a larger people-mover vehicle (the 16 person vehicle) yet somehow the only part you appear to have actually read was pointing out going 35 mph through a single-lane tunnel wasn't especially futuristic.
Man, 100% wrong twice in the same post. At least you're consistent
The $128,000 question (inflation): Will he take a hint?
May be Elon was referring to Radar maximum capacity when he mention "we were stuck in local maximum", may be there is no way to improve false positive with Radar.Or they are someone who actually understands that AI is not magic.
The best AI in the universe still can't make vision see through a solid object to what is beyond it like radar can, or make vision see through fog and snow that blocks visual wavelengths but not radar wavelengths.
They aren't removing the radar because AI has magically made the cameras capable of X-ray vision- it's because it's too hard to continue trying to integrate with the new 4D BEV vision system and they feel it's not worth the extra effort for the marginal additional info. See Prunesquallors reposted elon remarks on that.
That might well be the absolute right path to go down.
But we know for a fact FSD has gone down several wrong paths before (including in 2016 when Tesla and Elon loudly announced Radar was awesome and should be a primary sensor specifically because of its ability to provide data vision physically can not provide
They've now decided THAT was a wrong path they went down and are trying another thing.
Teslas willingness to do that is a major reason for their pace of innovation.... rather than, say, Waymo having spent 10+ years now going down the WE ARE SURE HYPERLOCALIZED MAPS AND LIDAR IS THE WAY and still only operating RTs in one tiny geofenced area.
But it also means you can't automagically assume each new path is FOR SURE THE RIGHEST BEST ONE, because it might not be.
Tesla is fine accepting that possibility- if it turns out wrong, they'll try something else.....
Some superbulls are... less fine accepting that possibility than Tesla themselves though.
James Douma already covered this in a video with Dave Lee.... where he told us the same thing I've said here.
LIDAR adds no value to a vision system.
Radar absolutely does add safety and value as a secondary sensor in conditions of low visibility and things like bouncing under cars but should only be used in the occasional situation like that- not as a primary sensor which should be vision.
But the goal is many times safer than a human, not just "any" safer right?
I literally posted a video of it preventing a Tesla from crashing into another car at speed on a highway because the radar saw the stationary vehicle 2 cars ahead the cameras could not.
That's not theoretical like your math- it actually happened.
It's likely it has happened many times, though not every time did the driver save the video and spread it around the internet.
Now does it happen often "enough" to be worth preventing? Some would argue 'at all' is worth preventing- especially since it's using a sensor already on the car.
But both the suggested reasons for the change I cited from the other thread- And Elons own words in the tweets Prunesquallor quoted- essentially boil down to integrating radar with the new 4D/360 vision system is REALLY hard, and they don't consider the # of cases where it might make the car safer worth delaying further development to make radar integration work properly with it.
There's every chance that math is totally correct... if getting a much-safer-than-humans FSD out sooner means forgoing a theoretically-even-safer-but-they-have-no-idea-how-long-it-would-take-to-make-work FSD that keeps radar that might well be a tradeoff worth doing.... and if research on radar 4D integration gets better in the future they can always revisit.
Anyway I'd again point out there's an entire section of the forums specifically FSD related if folks genuinely want to get deeper into the weeds of why this isn't a case of "Cameras just do everything radar can thanks to AI!" which isn't factually correct versus why this still might be the CORRECT decision even though it's removing data the cameras can't fully replace.
Whole forum for it, with multiple threads getting into it, here:
AI, Autopilot, & Autonomous/FSD
Discussion about AI, Tesla Bot, Tesla Autopilot (AP), the promise of Full Self Driving (FSD), as well as other Autonomous Vehicles.
It's weird you'd quote my post where I specifically say having millions of additional people experience the cars adds a lot of value and then claim I said it was a dumb idea.
The text you quote cites 2 potentially significant value-adds for the company- both the exposure at all, and the fact it claims Tesla will be building a larger people-mover vehicle (the 16 person vehicle) yet somehow the only part you appear to have actually read was pointing out going 35 mph through a single-lane tunnel wasn't especially futuristic.
Man, 100% wrong twice in the same post. At least you're consistent
asburgers
Sell order in at $8008.5
Similar threads
- Locked
- Locked
- Locked
