Tesla, TSLA & the Investment World: the Perpetual Investors' Roundtable

[woodisgood]

Elon seems rather confident in the vision-only approach.

https://twitter.com/x/status/1380606694816800768

(Twitter user) Owen Sparks: "Does this mean you can remove the radar from production, or will it still be included as a back up?:

Elon Musk: "Remove".

Wow. I mean we know they’re confident in vision, but are there really no scenarios where radar adds to the abilities of the car?

That’s a rhetorical question, obviously answered by the removal of radar.

If I’m the CEO of a competing self-driving company and I hear that Tesla’s vision approach is going so well they’ll be dropping radar from their cars, do I:
A) scoff as usual
B) step down

 

[Todd Burch]

Elon seems rather confident in the vision-only approach.

https://twitter.com/x/status/1380606694816800768

(Twitter user) Owen Sparks: "Does this mean you can remove the radar from production, or will it still be included as a back up?:

Elon Musk: "Remove".

This is what I was suggesting might happen a few weeks ago when Elon mentioned decoupling vision from radar. Drop another $100 (or however much it is) from the cost of the car.

 

[StealthP3D]

I don't buy any puts or calls, I'm a long so these short-term movements don't mean much to me. Having said that, call me old-fashioned for thinking that a stock should be valued on the merits of the company and not on the whims of a bunch of rich douchebags on Wall Street.

I totally agree. However, what kills me is the number of people who know this, that the share price is unpredictable and sometimes manipulated to prices they don't think are suitable, yet they still complain and whine about it. I learned at an early age not to complain and whine about things I had no control over, especially to complain to myself. Feeling entitled is very damaging to your returns.

Instead, I use what I do know, and knowing what I don't know, to my advantage. I avoid labeling the disconnects between my perception and the reality in the market as good or bad, right or wrong, it is what it is and I have no control over it. When it get's to the point when I think I can see what will happen next with a high degree of certainty, I take action. If I make a bad trade I don't get mad at the market for being "unfair" or manipulated, I file it away to use next time. It is what it is, neither good or bad. We have a choice to participate in the market or not, for that reason I can't complain about the market when it doesn't do what I think it should do.

Over time, I have learned the market is quite predictable in the long-term (good companies go up) and quite unpredictable in the short-term (most of the time). No one knows where the market is headed with any certainty most of the time. Always use what you know about the market to your advantage and know what you don't know. This is why over 95% of my gains are due to buying and holding. And I think I'm a better trader than the average retail participant, it's just not worth it most of the time to try to guess these things. The key is to know when the odds are well within your favor and when they are not. If it's close to a "push", I don't want to spend the time and mental energy. That's most of the time in terms of short-term speculative trades.

 

[Todd Burch]

Wow. I mean we know they’re confident in vision, but are there really no scenarios where radar adds to the abilities of the car?

That’s a rhetorical question, obviously answered by the removal of radar.

If I’m the CEO of a competing self-driving company and I hear that Tesla’s vision approach is going so well they’ll be dropping radar from their cars, do I:
A) scoff as usual
B) step down

Other than the radar bouncing under the car in front of you, I think if anything it adds to the confusion of the car's perception. I believe it's the primary cause of phantom braking.

And the "see the car in front of the car in front of you" probably already is getting trained for FSD. So yeah, for 99.9% of cases, the radar seems to be dead weight and a hindrance.

 

[Singuy]

This is what I was suggesting might happen a few weeks ago when Elon mentioned decoupling vision from radar. Drop another $100 (or however much it is) from the cost of the car.

Have they tested this during crazy FL rain storms because I probably need a Radar when I can't see nothing out of my windshield unless I drive way under the speed limit. It's like a flood coming from the sky. Fills up my pool in like 15 mins where it would take a hose 3 hours

 

[LiveLong&Profit]

Well damn Elon IS super confident if they end up actually removing the radar hardware from the cars

Elon is known for his optimism. It is a great strength.
You have to have an optimistic in order to achieve things that are extremely difficult. Otherwise you give up too early. Or worse: You never dare to try.
On the other hand: There is a reason why 'Elon Time' is a thing...

 

[Todd Burch]

Have they tested this during crazy FL rain storms because I probably need a Radar when I can't see nothing out of my windshield unless I drive way under the speed limit. It's like a flood coming from the sky. Fills up my pool in like 15 mins where it would take a hose 3 hours

In a crazy FL rainstorm, would YOU drive way under the speed limit if you couldn't see out of your windshield? The answer's yes, so why shouldn't it be yes for a self-driving car?

 

[tstolze]

Have they tested this during crazy FL rain storms because I probably need a Radar when I can't see nothing out of my windshield unless I drive way under the speed limit. It's like a flood coming from the sky. Fills up my pool in like 15 mins where it would take a hose 3 hours

Anytime I have looked at what a camera records in rain/fog it could see much better than my eyes could, this will be interesting to see if they do remove it from future builds.

 

[Todd Burch]

Anytime I have looked at what a camera records in rain/fog it could see much better than my eyes could, this will be interesting to see if they do remove it from future builds.

My guess is radar will stick around for another 6 months to a year, then disappear once Tesla has enough data to support what they suspect.

Edit: Ultrasonics will remain for situations where obstructions are very close (e.g. in a garage) and vision cannot provide useful ranging information.

 

[Singuy]

In a crazy FL rainstorm, would YOU drive way under the speed limit if you couldn't see out of your windshield? The answer's yes, so why shouldn't it be yes for a self-driving car?

It should, but I haven't seen Tesla adjust speed according to obstruction besides slamming on the brake. During a crazy rain storm you only get a brief second of clarity as wipers are on full blast even at low speed. And one feature everyone is hoping Tesla can implement but have yet to do so is having a memory bank of what it just saw since humans use this ability to navigate out of this situation when we know a second ago the car is still in x spot until we are blinded by the rain. Will the car react accordingly? I would think a radar would help validate the driving experience smoother.

Just feel like if Radar provides superhuman abilities like seeing through things humans have a hard time seeing, then they should keep it if Tesla wants to be safer than a human.

 

[gabeincal]

My guess is radar will stick around for another 6 months to a year, then disappear once Tesla has enough data to support what they suspect.

We might find that the new S&X don't have it any longer...?

 

[Todd Burch]

We might find that the new S&X don't have it any longer...?

I'd be surprised if they ditch it this soon. I think they'll want to gather significant statistical evidence that the radar isn't helpful before pulling it from production vehicles.

 

[Todd Burch]

It should, but I haven't seen Tesla adjust speed according to obstruction besides slamming on the brake. During a crazy rain storm you only get a brief second of clarity as wipers are on full blast even at low speed. And one feature everyone is hoping Tesla can implement but have yet to do so is having a memory bank of what it just saw since humans use this ability to navigate out of this situation when we know a second ago the car is still in x spot until we are blinded by the rain. Will the car react accordingly? I would think a radar would help validate the driving experience smoother.

Slowing down for obstructed vision would be one of the easier tasks that FSD will need to handle.

 

[Knightshade]

Elon seems rather confident in the vision-only approach.

https://twitter.com/x/status/1380606694816800768

(Twitter user) Owen Sparks: "Does this mean you can remove the radar from production, or will it still be included as a back up?:

Elon Musk: "Remove".

That just seems nuts to me.


LIDAR is useless because it can't provide any data you can't already get with vision.

Radar absolutely can provide data, in several different ways, that vision can not provide including not just the bounce under cars thing but things like being able to see through fog and other visual obscurations.


(Yes I know humans drive without radar, but they drive without ultrasonics and cameras in the back of their head too- the point of this is to make a system that drives way BETTER than humans- lidar doesn't add anything to that goal, but radar does)

 

[Singuy]

Slowing down for obstructed vision would be one of the easier tasks that FSD will need to handle.

except this is a thick layer of water, aka pseudo obstruction.

Just feel like if this is only being tested in Cali or FL right now where we don't get them 3pm down pours until summer months only, how can they be sure radar is useless?

 

[st_lopes]

Does Tesla offer a post delivery software range unlock / upgrade on these cars post delivery the way they used to with the 60D to 75D Model S a few years ago ? ( or are these cars delivered with materially fewer cells )?

If this is legally permissible, that would deliver even more high margin revenue to Tesla in the future, result in a lower sales tax outlay for the consumer in jurisdictions where “luxury taxes” are applied based the purchase price of the car.

ie in British Columbia, Canada - the sales tax on a new or used car can range from 12-25/26%. But if you buy accessories or servicing; you would pay 12% - if it is done after the car was delivered.

Electrek running a piece that TSLA is now selling a software locked option M3 SR in Canada (was previously "off-menu" and was the only vehicle that qualified for the Canadian Federal Rebate of $5k). That article gets rerun by Benzinga inferring that there's a demand issue at the SR price point of about $53k CAD. What both have missed is that the SR+ has actually qualified for the Federal rebate for several months now, even though its MSRP is $53k and per the local service centers are flying off the lot as quickly as they can make them.

How, you ask, does the SR+ access the rebate when the MSRP threshold for the rebate was $45k CAD?

Well the SR+ is the same car as SR. It's just a higher trim. The rules of the program are that as long as the BASE trim is <45k, any higher trims can qualify up to $55k CAD. So, by "offering" a software locked SR that you may never actually sell any of, you unlock the $5k rebate for your higher trim vehicle. Magic. Making the option available through configuration was likely just a way to satisfy the program administrators that the software-locked model does in fact exist and is easily available for sale (as opposed to the off-menu, ask a sales rep, some know about it some don't approach) and equally solidifies that they are the same vehicle, ensuring the SR+ can access the rebate as well.

@Tes La Ferrari at the time when the software-locked model was available only "off-menu" and was the only car that could qualify for incentives, you COULD pay for an OTA upgrade that would effectively make the car an SR. Likely still is the case now, but the cost of the upgrade will likely make it so that the net cost of vehicle will be the same as just buying the SR+ with incentive outright. As for luxury tax, Quebec waives luxury tax on EVs up to $75k. So, I've never had to look in to that.

I tried to haggle my way in to doing the same on the M-Y to try to access the Federal and Provincial rebates, but there was no way they were going to software-lock it to come in at $44,999, even if I promised I would pay for on OTA upgrade immediately. Which I inferred meant that the M-3 SR was a car they could still make margin on even if the buyer never bought the upgrade, whereas an M-Y at $44,999 wasn't going to happen.

 

[Todd Burch]

except this is a thick layer of water, aka pseudo obstruction.

Yes, but they already use vision to control the wipers, so they know when it's raining. The logic can basically be

"ooh, I'm starting to see raindrops. Start rain event. If recognizable features are not discernable or confidence drops below a certain threshold, put on hazards, slow until you go above the threshold, or pull over/stop as necessary."

It's easily within the realm of the technology Tesla has already demonstrated IMO.

 

[Singuy]

Yes, but they already use vision to control the wipers, so they know when it's raining. The logic can basically be

"ooh, I'm starting to see raindrops. Start rain event. If recognizable features are not discernable or confidence drops below a certain threshold, put on hazards, slow until you go above the threshold, or pull over/stop as necessary."

It's easily within the realm of the technology Tesla has already demonstrated IMO.

Just saying it's not safe for a human to drive during a down pour like that and we do see some responsible people pull over. The majority of the people keeps on going at low speeds despite having about 10 ft of visibility at most and it's mostly from tail lights. I feel like radar here would provide additional level of safety even going at low speeds.

 

[capster]

Elon seems rather confident in the vision-only approach.

https://twitter.com/x/status/1380606694816800768

(Twitter user) Owen Sparks: "Does this mean you can remove the radar from production, or will it still be included as a back up?

Elon Musk: "Remove".

Perhaps what the active sensor proponents forget is that vision only is not, strictly speaking, a purely passive solution in this context because... wait for it... vehicles have headlights.

 

[ThisStockGood]