Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

There was never a plan

This site may earn commission on affiliate links.
Speaking of the competition from the other auto OEMs... :p


It's not like every single other auto OEM has been selling their own ADAS package for years and Elon is being mocked for being woefully behind. It's the exact opposite.

Tesla intends to keep the tech lead and FSDbeta demonstrates this to customers.

Oh hey, look--it's the non-owner FUDster that's been on TMC for 9+ years and has 10k+ posts that are probably all negative.

Meh.
The only problem with FSD is the bullshit from Musk while engineers have tried to do the hard work of figuring out the hardware and software requirements.

This implementation and bitching about the safety score it's just a sideshow.
 
  • Like
Reactions: rxlawdude
When Telsa was the only game in town, it made sense to give Elon the benefit of the doubt. Was he trolling us or just wildly optimistic? It doesn't matter now. There are serious players in the market, yes, grown men (and women) who don't fixate on farts and Jack-in-the-box. People who can manage a product release. I'll always have a soft spot for the guy and an infitite respect for what he's accomplished, but the cowboy days are over. Tesla had better shape up or get left behind.
 
This is not a rant or complaint about the way Tesla has handled the FSD rollout - there are plenty of threads on that topic.

What I find fascinating is that, despite the fact that Tesla has been working on FSD since AP 2.0 was released exactly 5 years ago, it is clear that no one gave any serious thought to how they might gradually roll out significant new FSD capabilities to customers. It is clear that Elon literally started making it up as he went along in a series of tweets this year. There was never a plan.

The reason I find this so fascinating is that they are solving infinitely more complex problems on a daily basis. Watch any Karparthy presentation and you'll realize what astounding progress they have made over the years. Imagine the amount of talent required to come as far as they have.

And yet the chaos and confusion around the FSD rollout has been comical. When the price of FSD was reduced several years ago, Tesla hurriedly published a blog promising early FSD access to early purchasers. They seemed to forget this promise as soon as they made it. Elon then just offers up on Twitter for anyone interested to 'ask us' if they want access to the beta, then makes up the idea of a 'button' after their service centers get swamped with requests. He then has to stick to his plans for a 'button', while his team scrambled to figure out how to build and release this feature. And then they realize they need some kind of method for assessing beta testers, so they decided to repurpose the Safety Score system built for their insurance program.

In contrast, it would take a single mid-level product manager/project manager to plan a sensible, incremental, and customer-centric rollout of FSD features. Some examples:
  • Converge the production and FSD beta code base first - with most of the navigate on autopilot features disabled. This would give all FSD customers a greatly improved 'basic autopilot' experience. Then gradually enable features - for example, auto steer on unmarked roads, or 'no confirmation' traffic light control.
  • Recognize early on that any rollout of the 'full' FSD beta would need to be gradual (something they are *finally* beginning to plan for now). Develop a gradual rollout plan that takes into account date of FSD purchase, along with other considerations (such as Safety Score). Clearly communicate to customers via e-mail, in-app notifications, etc. Perhaps even give them an idea of where they are in line?
  • Stick with the plan that you communicate to customers. Don't make it up via Twitter and then change the plan every few days. I'm not saying they need to hit specific dates, as that is very difficult when dealing with an engineering challenge of this magnitude. But be consistent with *how* you're going to release these features to your customers.
Again, not a rant. I don't agree with the haters who say that Elon is a charlatan and is selling vaporware. It's clear to me that they have something very, very real here, and they clearly want to get to a wide release. But why they would put so much effort into engineering and innovation, and so little into basic planning and communication, is baffling to me.
Good post.

IMO, because…. They had a vision and broad objective. I’d imagine there have been dozens of plans along the way. And when they hit an insurmountable roadblock, the plan was redone or the then head of AP was fired. However, they have persisted.

If this could be planned out with predictability, someone would have done it. But you cannot always plan doing something that’s never been done.

in my day, we called this design by prototype. Drives some people crazy, but I love it.

There’s a huge pot of gold for whomever does get it right. I suspect it will continue to be like this til they get there. Or the regs stop them.
 
Hyundai/Kia/Genesis is making very good EVs, but their ADAS (Ioniq 5 in this case) needs a lot of work. It's like Ford's system where it disengages on slight turns. They might be using Mobileye as well.

See from 3:38:00



Like who ?

In consumer FSD space there is no one anywhere close to making city FSD a possibility. All legacy manufacturers are too afraid to take risks.

We've all been hearing that "the competition is coming" for years.

Yeah, not only is the competition not catching up, it's falling further behind.

Go try and order a new GM with supercruise today.

You can't.
 
Ford is using an older version of MobilEye.

Incidentally, this video gives really good insight why Tesla is way ahead of MobilEye. Many people think MobilEye is ahead or on par with Tesla because their videos show perfect runs in tough locations. But there's more than meets the eye....

Start at the 3:00 mark (and you prolly want to watch this one on 2x playback :D )

 
When the price of FSD was reduced several years ago, Tesla hurriedly published a blog promising early FSD access to early purchasers. They seemed to forget this promise as soon as they made it.


NO. Do not let them off that easily. They didn't forget. They TOOK DOWN THE POST and pretended they never said any of it. There is a wide gulf between negligence and dishonesty, and while Tesla has had their share of both, this falls under the latter umbrella.
 
Hyundai/Kia/Genesis is making very good EVs, but their ADAS (Ioniq 5 in this case) needs a lot of work. It's like Ford's system where it disengages on slight turns. They might be using Mobileye as well.

See from 3:38:00





We've all been hearing that "the competition is coming" for years.

One thing to note about Fords Blue Cruise is that it disengages on slight turns on PURPOSE.

It's not that the system isn't capable of those turns, but they powers to be at Ford didn't want to risk it.

The issue with L2 is the better it gets the more complacent the driver is going to get. The more complacent the more likely an accident. So the easiest solution to that is to simply not allow it to do much. This is the route a lot of car companies take, and the route the Europeans took in limiting what AP could do.

I haven't had a chance to test drive a vehicle with blue cruise, but I will once they officially release it.
 
There was a plan, but not everyone likes the plan.

1 year, no accidents. Plan is working. FSD beta is improving very quickly.
This is some next level gaslighting / 1984 stuff. There is zero evidence that Tesla's plan was to go a year with 71 beta testers and no accidents before going wider. In fact, it's trivial to see that Elon has been trying to release it for a year.

"We have, hopefully, a wide release by the end of this year."
--Elon, Q3 2020 conference call


Plus, what good is a "year without accidents" when that year includes two major re-writes? Why not just start the calendar in 2016 and say it's been 5 years!?

Even you know the reason it hasn't been released is it's not good enough (by whatever measure, safety, functionality, PR, etc). Not that they were waiting for a metric of 1 year, no accidents (which may not even be true based on your definition of accident) as some part of grand plan.

You want there to be a plan, because you know it it looks really bad to try and release self driving cars without one. But that was the only plan. Release self driving cars. Yell at your engineers, you post that it's coming soon, you engage the reality distortion field, you let people post videos (that you later claim are NDA violations) to try and make it as fast as possible. But that's not really a plan.
 
Last edited:
I'll just lay out here what I believe is the hard on-the-ground truth about the FSD timeline debate:

- it was oversold in the beginning, with no visible progress. many people probably assumed Tesla had this working but wasn't showing it.
- in the past year, it's become obvious they only really, really just started tackling city streets, making FSD finally feature-complete
- but even with city streets, still far and away from the lofty early promises.
- those early promises with early hardware may have been viable in a sterile, traffic-free road system, or in cooperation with other identical Teslas, at least.

That is to say: there have been just major advancements in the past year alone towards finally making FSD a reality. Thus, stretching comparisons back more than a year into the past are kinda sloppy - as the rate of growth in the past year is far higher than that of 2016-2020. FSD's development can practically be isolated to a vacuum of the past 1 year, when development on it really, honestly seems to have started in earnest.

Just my thoughts on it.
 
I agree completely with everything you've said. My guess is that this event might be relived in college management classes over the years as a case study in what NOT to do. First lesson: Don't run the company information releases via the CEO on Twitter.
First: get an actual CEO/COO to run the company and keep the "ideas person" locked away
 
Tesla ADVERTISED and SOLD city streets autosteer as "coming this year" in 2019 and 2020.
Would that not be fraud if they were not working on in "in earnest"?
There was no visible development until about the past year (late 2020) when FSD Beta finally started going onto real cars.

I already had a separate argument for their messaging and advertising. My point is solely on the "extrapolating the rate of progress from 2016 out to infinity" being flawed, when the real progress really started happening very recently.
 
First: get an actual CEO/COO to run the company and keep the "ideas person" locked away
If you don't have a real leader running the company, someone that actually inspires people, you end up with a company just like everyone else: full of clock-punchers and PowerPoints.

For as clunky as Tesla is, their "ideas person" is the brightest star in the industry and is the reason they're shattering production records and lighting the auto industry on fire.

(oh boy, look, now I'm an Elon fanboy again. I was a hater just 5 minutes ago on Twitter, lol)