There is a difference between technical capability and usability. Posts like this highlight the former in a way that suggests the latter is true today. It obviously isn’t, and anyone using FSD today knows it isn’t. Someone who hasn’t ever tried FSD and reads an official post like this from Tesla is mislead. Everyone, including whoever posted this at Tesla, knows that FSD supervised cannot drive you 25,000 miles without touching the steering wheel unless you are willing to let it break the law, take wrong turns, park in the wrong place, etc. It is an incredible statement. It proves that the system is very safe and can cook unlike anything else of its kind. FSD supervised is the most incredible technology I’ve ever used, and I am 100% certain it would save many lives today if adopted at-scale. But why mislead the average person into thinking they can get in a Tesla and replicate this today in a way that is actually practical? There is a reason almost nobody else has a streak of over a few hundred miles. Idk man. Just feels weird to me seeing Tesla post stuff like this. There is still more work to be done before official D swinging like this should occur.

I haven’t driven this specific spot in probably 2 years. It was a problem on V12, but I’m genuinely surprised to see that it’s still a problem on V14.3.7. The car tries to stage to make a right turn on the restricted paint, when staging lane is the one the pickup is in. Really surprised to see FSD still missing basic stuff like this. It usually treats these things correctly, but don’t understand why it wouldn’t see this here.

HUGE NEWS: Orgain plant-based organic protein is currently 8 dollars off per tub at Costco. Orgain pairs fantastically with Fairlife Whole Milk. If you weren’t already aware, Fairlife whole milk isn’t ordinary milk — it’s lactose free, and packs a 50% greater protein punch than a standard gallon. Guess what? That isn’t even it. Fairlife Whole Milk also contains 50% LESS sugar than standard milk. Great for the muscles, easy on the waist.

I’m generally a super calm and patient guy, but if a group of slow walkers line the sidewalk and I have no way to pass, you’ve got about 10 seconds before I totally black out and turn completely insane

Me: “that’ll be 21.19” Customer: “Do you take plastic?” Me: “Ah, no we actually don’t. Only metal. You can just leave everything on the counter. Get back out there and grind, king.” Customer: “…..”

Did anyone who got this new FSD V14.3.7 second update notice that they changed the “install update” UI? The clock icon is black now, and the install pop-up is different. Must’ve snuck it in with the first 14.3.7 update.

I drank way too much caffeine every day this week, and Coach Claude caught it in the data 😬 Nothing gets by this bastard, I swear. Also uhh, this is one of the most valuable tools I’ve ever used.

This FSD clip highlights two things that remain the biggest question marks to me with self driving: taking cues from the tiny details, and lack of memory. Humans are so good at this. We can see the faintest “only left turn” paint in the distance and know how to stage the car. Or a tiny little overhead sign. I’m less concerned about this for FSD and think it’s probably a matter of increasing parameter count, but not sure if they’re going to be able to get proper resolution with HW4 In addition to this, we also have memory. When you’re pulling out of a parking spot in some small private lot that is one-way, you’re usually angling the car out correctly because you remember which way you’re supposed to pull out. If you have no memory, oftentimes the only possible cue is a small arrow painted on the tarmac. Sometimes this isn’t even available. In scenarios like this, I don’t see how even the most intelligent AI can perform correctly without memory. There are definitely other moments where human memory is 100% relied on for correct vehicle operation. Maybe the system just needs to be intelligent enough to hold context from one drive to the next when it recognizes scenarios like this? Idfk. There’s probably a way to do it without human-level memory, but seems hard. I think in terms of safe driving, Tesla has already passed human capability and I have no concerns here. But when it comes to autonomous fleets at massive scale without being disruptive, I wonder whether they’ll be able to do it with HW4

I don’t think anyone, including David, would argue that in order to achieve this streak, you have to let the car make safe mistakes, and you have to massively inconvenience yourself/sit through things consumers will never sit through. But that’s not really the point. I don’t think David is trying to use this to make an argument that FSD is fully autonomous and anyone can do what he has done/would even want to. What it does do though is show that the vehicle can safely operate itself for many thousands of miles, and that almost every remaining issue that would have the average person disengage isn’t safety related. It’s definitely a strong statement for how safe FSD has become. But yeah, you can’t get to 25,000 miles or probably even 5,000 without multiple moments of major inconvenience. The product isn’t there yet for this sort of thing to be common, but it’s profound that it’s even possible at all today

This video is actually so fucking good it feels staged. But if you’ve ever had a run-in with Pipsqueak, you know that only him, in his relentless narcissism, could produce such an iconic video naturally. The entire thing is gold, from the first words til the very last

Agentic AI has been the most liberating experience I've ever had with productivity technology in my life. It has changed everything for me. I have automated every part of my job I hated, and I have far more time to do the things that I actually love -- the reason I'm actually doing what I'm doing. No more ball and chain. Once you're in it, the high is insane, and you just keep thinking of more and more things you're able to do as you get deeper down the rabbithole. Automate this, automate that. The vibe is: "oh, now I have a massive soup of data growing and growing, and a genie sitting on top of it who can suck it all in and grant me unlimited wishes." I have done this for myself using Claude Code, but Palantir can take the little tiny world I've built for my own job and scale it to the entire enterprise. The power is insane. Anyone who adopts this stuff will absolutely run circles around anyone who doesn't. It really is existential. That's how big of a deal you realize it is after you've gotten a glimpse of it.