Justin Woolhandler

Stocks ($TSLA) - Sports (Steelers/Pens/Bucks/Pitt) - Law (O&G/Title) - Politics(moderate Dem) - Chess

[Justin Woolhandler @jwoolhandler]

We use too many abbreviations. I literally have to look at least one up every day. We should have more respect for words and the rhythm of language. This isn’t The Pitt ER, you’re literally sitting on your couch doing nothing, confusing half of the audience to save a millisecond.

[Justin Woolhandler @jwoolhandler]

@TheLongInvest I don’t think it happens for a while. The deal will need to be favorable to Tesla, because its shareholder base could reject the deal. So I think they wait for Robotaxi revenue to actually matter, late 2027.

[Justin Woolhandler @jwoolhandler]

One of the key factors here is Elon’s control of Spacex. This merger would need to relatively favor Tesla, because there is a very real chance of shareholders not approving. For that reason, I think this merger takes a while. It makes more sense to do it when Tesla has real Robotaxi revenue and a much higher share price.

[Kalshi @Kalshi]

JUST IN: 54% chance Tesla and SpaceX merge this year

[Justin Woolhandler @jwoolhandler]

I hate when people say $TSLA has been flat for five years. That’s only true for investors who solely bought the top and have held this then.

I’d say most Tesla retail investors have huge gains from pre-2022 buys and, if you’ve been dollar cost averaging, shares acquired the last few years are up 80% - 200%.

The stock ran up too high and came back down to earth, like all story driven growth stocks do. The move from $180 to $400 since has been earned.

[Justin Woolhandler @jwoolhandler]

@squawksquare Honestly, the only reason I’m not selling is I know my brokers worked their asses off to get me an allocation and it feels wrong trading it away a few days later for a small amount of gains in relation to my larger portfolio.

[Justin Woolhandler @jwoolhandler]

@KevinMelnuk @betastuff1999 @tygrepath I thought I was the dumb guy who’s totally lost that you muted. If you can’t have a cordial argument without name calling and accusing me of not thinking, I really don’t feel like articulating my feelings on Waymo. The market seems to agree with me.

[Justin Woolhandler @jwoolhandler]

I am rooting for Rivian and kinda was considering investing with the R2 launching. But how could this be achieved without the data and compute Tesla used to make FSD?

I hate to sound bearish on a company working on some awesome cars, but this kind of sounds to me like the R2 production line is not profitable and they need another story.

[Sawyer Merritt @SawyerMerritt]

Rivian CEO RJ Scaringe in new interview:

"Later this year, we'll have full supervised point-to-point, which will be very similar to @Tesla's FSD. That will roll out to all of our Gen 2 vehicles and, of course, R2. Next year, we'll allow it to go unsupervised, so you can take your eyes off the road. The next step is achieving full capability for the vehicle to drive itself with no one in the car. We've been developing that for a while for our personal vehicles, but we also see it unlocking new business models, robotaxi being one of them.

We took the decision to parter with Uber so we could focus on the tech and leverage them for their access to a big distribution channel. They're really a category of one. There's no one quite at their scale or reach."

Full interview: https://t.co/8zE1FmQShw

[Justin Woolhandler @jwoolhandler]

@betastuff1999 @tygrepath @KevinMelnuk Again, Robotaxi’s are operating without safety drivers with AI4. Are you saying that it works when you have 50 Robotaxi’s but not 500, 5,000 etc.?

[Justin Woolhandler @jwoolhandler]

This is the best WC game so far. I love Japan’s game. So well coached.

[Justin Woolhandler @jwoolhandler]

@tygrepath @betastuff1999 @KevinMelnuk Yeah Elon literally said that HW4 is sufficient.

[Justin Woolhandler @jwoolhandler]

It sounds like they are using both imitation and reinforcement. I wouldn’t be surprised if you need both.

If HW5 is necessary, Tesla sure isn’t acting like it. Cybercabs rolling off the line, job postings in major cities, Elon said HW4 is sufficient on the last call and Robotaxi’s are literally operating now, without a major incident.

[Justin Woolhandler @jwoolhandler]

I may be in over my depth and I do like the read, but I think Tesla uses both imitation learning and reinforcement learning?

And I don’t think Tesla is stuck. Cybercabs are rolling off the line for a reason.

And does Rivian even have the compute they need to create a Level 4/5 system?

[Justin Woolhandler @jwoolhandler]

@KevinMelnuk @betastuff1999 @tygrepath I think Waymo is a joke. They embody everything wrong with the max sensor approach. I think Tesla eats their lunch, but we’ll see!

[Justin Woolhandler @jwoolhandler]

My whole timeline is the Tesla community quote tweeting Rivian autonomy claims, but I just want to see Netherlands v. Japan content let’s go!

[Justin Woolhandler @jwoolhandler]

@betastuff1999 @tygrepath @KevinMelnuk We’ll see. I don’t think anyone else has all the pieces in place like Tesla does. I’m open to the idea that Rivian has a straighter, faster path than Tesla took, but they don’t have the fleet size and data to solve edge cases.

[Justin Woolhandler @jwoolhandler]

@SawyerMerritt The Tesla community has been dumping on RJ’s autonomy claims today, but at this point, I got Rivian ahead of Waymo.

[Justin Woolhandler @jwoolhandler]

Curaçao whoop whoop!

[Justin Woolhandler @jwoolhandler]

I buy that. Though, FSD was neural networks wrapped with C++, it wasn’t all hard coded. Rivian benefits from our knowledge AI being exponentially better than it was five years ago. But I don’t see how they get around the data issue. It’s seems to me like you need to believe that simulation is sufficient to replace the millions of Teslas collecting driving data every day. Or Tesla engineers just suck and they’ve had enough data for years and failed to utilize it correctly.

[Justin Woolhandler @jwoolhandler]

I know many smart people ascribe to this view, but I’ve been using FSD for six years. It doesn’t have perception issues. It’s all about inference and I don’t believe that lidar and radar provide some kind of shortcut that helps the system make better decisions.

And, they were using neural networks the whole time. It’s just as AI improved, it allowed the neural network to subsume parts of the system that were hard coded.

[Justin Woolhandler @jwoolhandler]

@KevinMelnuk @tygrepath @grok Okay thanks I’ll watch the YouTube video the other guy mentioned. It is Sunday morning haha. Are you an investor? Because $20 bil is very low if you think they can roll out a system like FSD.

[Justin Woolhandler @jwoolhandler]

@tygrepath @KevinMelnuk Tesla made their own chips too, but yeah I’ll watch that thanks.

[Justin Woolhandler @jwoolhandler]

@grok @dare_darr3n @moregainzs I don’t know the details, but I thought alphamayo was just the rudimentary tools to begin training a self-driving system. It’s not a solution for all the edge cases, which requires data that Nvidia and their customers do not have.

https://t.co/LZTVR88lb1