That's me, just squeaked into the old-school group Nov. '21 (Model Y). I'm an Intel agent!
Until now, most of the Ryzen advantages related to gaming and maybe streaming, not big priorities for me. But "immersive visualizations" could be interesting.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Elon: FSD Beta tweets
- Thread starter Terminator857
- Start date
Until now, most of the Ryzen advantages related to gaming and maybe streaming, not big priorities for me. But "immersive visualizations" could be interesting.
Until the cheaper car and the factories are really ready for volume production, it would be a poor business strategy to announce it.
If the Robotaxi version is unveiled as a significantly pricier version, or as a commercial fleet vehicle only*, Tesla avoids most of the Osborne effect.
* I'll be highly disappointed if FSD Unsupervised is tied to commercially-operated hardware only, but it could happen. As much as I want my existing car to fulfill its Autonomous destiny, I'd still be willing to upgrade to a newer car to get it. It doesn't have to be the low-cost Robotaxi; something like an HW5 3 or Y is OK and the sooner the better l. But not being allowed to operate it personally, at all, would be bad.
Yeah. I’m cool with updates my intel MCU can’t handle to not be available for my legacy model 3. Like gaming etc, certain streaming… but “visualizations” or other things I don’t trust Tesla enough to why suddenly they only work for cars made within the last ~2 yrs and 4 months…
If robotaxis and FSD are priority #1 and balls to walls…. Why in the world did Elon waste precious resources and engineering on the Cybertruck ?
Yeah, I think in hindsight it's pretty clear that Cybertruck was a bad idea all around. It doesn't really live up to its hype and it took way too many resources away from other more-important goals (like Model 2 + Robotaxi, further battery developments, software investment in FSD, etc). It was never going to be mass-market anyways. IMHO, the new Roadster goes in the same bucket. At least that one is capable of meeting the hype expectations, but it will never be mass market and doesn't help with any bigger plans. Tesla really needs to focus on what matters: evolving the battery and propulsion systems for greater efficiency and longevity, making mass-market models, reducing mfg costs in general, and delivering on FSD software.
I know we always say "never say never," but at this point I am comfortable in saying there will never be a Tesla Robotaxi - not with anything resembling their current FSD software and hardware. And, yes, I have it and drive using it frequently (almost daily).
agree - especially on the hardware part. "cameras only" without Radar, Lidar, night vision has just too many limitations at night, in bad weather, sun blinding camera sensors etc. you want a driverless L5 vehicle to see *more* than a human and not *less* - and yes, the current cameras with their resolution are nowhere near human eyes.
I never bought into it in the first place. But I also don't follow Musk on any social media so the only time I ever heard of "robotaxi" was here on TMC. I reached out to a few Tesla owners in my company and asked them if they heard of "robotaxi" from Tesla - nope. They had no idea what I was talking about.
JupiterMan
Active Member
it was widely covered for all to see by Tesla in their "Autonomy Day" back in 2019.
It is not TMC gossip.
I'm not saying it is. The "common man" has no idea. Just us enthusiasts.
Knightshade
Well-Known Member
Anyone who bought FSD back in the day would've seen it if they read the fine print, where Tesla told us using FSD for ride sharing commercialy was only ok on the Tesla Network, details of which are coming next year, and have been since like 2017.
Interestingly, the Wall Street Journal had a podcast two days ago about the robotaxi.
Could Tesla’s Robotaxi Finally Be Coming? - Tech News Briefing - WSJ Podcasts
For almost a decade, Tesla CEO Elon Musk has been promising that fully driverless vehicles are coming soon. Now, Musk says he plans to reveal the company’s first robotaxi in August. WSJ columnist Tim Higgins tells host Alex Ossola why Musk says Tesla has an advantage over its competitors… and...
www.wsj.com
DanCar
Active Member
Off topic: Tweeter gives long thread on what it takes to improve FSD, with some technical references. Lots of interesting replies. Summary: More compute, bigger models, more training data.
One of the interesting replies:
Quote: The learned model is most likely in the 100s of MBs and easily run even with a 100x growth of learned parameters.
Speed of the inference model is key and the architecture needs to execute on the order of 40 to 100ms per photon to control output cycle.
One of the interesting replies:
Quote: The learned model is most likely in the 100s of MBs and easily run even with a 100x growth of learned parameters.
Speed of the inference model is key and the architecture needs to execute on the order of 40 to 100ms per photon to control output cycle.
Similar threads
