Forums

Self driving features complete by end of 2019

Self driving features complete by end of 2019

Elon Musk announced Tesla will be able to achieve feature complete for full autonomous self driving by end of 2019. What do you think? If possible, does this require upgrade to HW 3.0?

Shock | 20 February, 2019

No chance in hell it happens by end of 2019, so don't worry about requiring an upgrade any time soon :)

spuzzz123 | 20 February, 2019

Shock is back and optimistic as ever! Unfortunately I agree with his comment. Too many external forces working against it.

thedrisin | 20 February, 2019

External forces? It is hard to believe that the vehicle will go from a couple of limited beta features with more warnings than instructions, unable to work fully except in the best of circumstances, to FSD. I am skeptical there will be level 4 or 5 autonomous driving by the end of 2019. Would be wonderful, but doubtful.

rdavis | 20 February, 2019

I don’t see level 5 until the government requires all cars on the toad to be autonomous. Level 4 by end of 2019/2020 is possible in hardware/software, but I doubt regulators would allow it for many more years.

I would expect 2023-2025ish for limited use of FSD in set geolocation areas under certain criteria, but many many more years to full level 5.

howard | 20 February, 2019

NO WAY

WantMY | 20 February, 2019

Mr. Musk is master when it comes to moving goal posts. Do not loose sleep over it, lol

sheldon.mike1010 | 20 February, 2019

Yes, we're close, but end of 2019? - not happening. I'm just enjoying what it does now and will be happily surprised when it's capable of more.

Steven.Haver | 20 February, 2019

Don’t take it out of context. He thinks the features will be complete by the end of this year. That means he expects they will have a version of the software that contains all the features that they think are necessary for FSD. It doesn’t mean it will be stable or ready for release yet. Just that the features will be functional.

From there they can reduce the focus on new development and shift their resources to cleaning things up, making the software more efficient, improving the subtleties, handling corner cases, etc. So give that about 6 months to clean up, and then give it about 6 more months of functioning in shadow mode to collect billions of miles of data.

Either way, it’s fun to watch it evolve.

WantMY | 20 February, 2019

@Steven.Haver Following your logic just about any manufacturer can declare they are ready with FSD BS. As it is not possible to validate, because of it is not even legal for years to come. BS is BS, just call it when you see it.

ksusweep | 20 February, 2019

My guess is that the self driving on the highway (at least in the US) will be much better, with massive amount of data been collected, especially from navigation on autopilot. Autonomous driving, with auto land change without human inputs might be possible by end of this year. Not sure if I want to be the first one to try it though.

I don't completely trust the current iteration as is, and I suspect HW 3.0 is required to safely handle the level 4 autonomy - along with other sensor upgrades, perhaps a rear facing radar, and / or more cameras.

billlake2000 | 20 February, 2019

Hell, I already fully drive my own self! And my nomous is auto.

RedPillSucks | 20 February, 2019

@thedrisin Remember that the FSD testing being done at Tesla is most likely happening on selectively few machines with HW3. So I wouldn't make any assumptions on the functionality we have currently vs FSD. Having said that, I agree that it most likely will not happen this year. Even if they're able to complete the features by next year, who know when they'll get regulatory approval and be able to push it out to everyone. Worse yet, if things end up as state by state approval.

chris.pribe | 20 February, 2019

I bought FSD with my Model 3 last August with the estimate that I'd have something useful from it on my car within 2 years.

So, yeah, I think they'll have something amazing in that time frame, though what exactly I don't know; certainly something worth at least what I paid, perhaps much more.

Not much interested in what the naysayers say it can't, won't, or shouldn't do or any particular quibbles about this or that level, timeframe or technology.

I also think the US regulators will be aggressive in supporting autonomous driving. In case folks haven't noticed, we're in a new and rapidly becoming more intense race for future markets enabled by advanced technology.

jithesh | 20 February, 2019

EAP is one thing but FSD is a whole new level.

even if tesla completes FSD by 2019 they still need to get regulatory approval which I think is the trickiest part of this. it's going to take a lot to convince the govt to approve this.

who will be responsible if FSD Tesla involves in a crash due to a FSD error? how will it impact the owner's insurance ? i think there are some legal questions to be answered before we see this on roads

thedrisin | 20 February, 2019

@jithesh. Tesla would have to be liable or it could not be true autonomous vehicles. A bunch of fearures with extensive warnings making the driver responsible would not be autonomous driving by definition. Tesla has to be ready to take on the legal responsibility for it to work.

Brian | 20 February, 2019

I think they are probably a lot closer than most think. They have a couple of videos out where the car does an amazing job navigating on country roads, driving someone to work, merging with traffic and obeying traffic lights. EAP went from keeping a lane and making lane changes to taking exits on it's own. I think that this advance has provided an enormous amount of data that is going to help them make a giant leap towards FSD. Just my opinion.

Brian (The one from Massachusetts)

inconel | 20 February, 2019

Features complete but still requires human supervision. I'd say possible

TAC | 20 February, 2019

I see the same comments everytime anything is posted about FSD most of them about how its not going to happen .... .... ... ..
Let me guess. All the naysayers and negative nellies about it dont have the FSD package pre ordered am i right?

ksusweep | 20 February, 2019

Wishful thinking aside, a question.... perhaps with legal implication, is if 'full self driving' will require minimum of HW 3.0 and perhaps other sensors not equipped in the current production models. This is an excerpt from the Tesla website that was announced at the time when Tesla started to make cars with HW 2.0:

"All Tesla vehicles produced in our factory, including Model 3, have the hardware needed for full self-driving capability at a safety level substantially greater than that of a human driver"

dalesmith1962 | 20 February, 2019

“but I doubt regulators would allow it for many more years.”

Regulators are on record saying that FSD will be authorized as soon as it is validated to work. How long that will take is unclear. It could be as soon as a few months after FSD is available for testing. No one really knows.

ODWms | 20 February, 2019

I’d be blown away if level 4 autonomous driving will be widely available in the next 5 years.

RES IPSA | 20 February, 2019

Nothing happens with widespread Level 4 or 5 autonomous driving until 5G network is up and running.

solstice | 20 February, 2019

I contacted Tesla last year and the rep told me self-driving will be available by 20/20. Need government approval. Just think of all those driving jobs lost.

Madatgascar | 21 February, 2019

Would you take your eyes off the road for 5 seconds with EAP? No. Would you go to sleep behind the wheel next year if Tesla flips the FSD switch and regulators say it’s OK?

I’m sure the feature set they have in place is amazingly capable. Elon has probably been enjoying it for years and wants to share it with humanity. I don’t care how good it is, as long as there are lawyers Tesla will never be able to tell people they can go to sleep. I don’t know why Elon thinks he needs to paint this picture.

M3phan | 21 February, 2019

Here in link below Elon clarifies what he means. It’s not FSD as in we enter car and go to sleep and wake up at destination. Whatever “it” is will still require driver oversight.
https://apple.news/AJUoiiZthQSWa5dawP7yiyw

jithesh | 21 February, 2019

@thedrisin that's what I would expect but in some recent interview I heard Elon saying that even with FSD the last certain % of autonomy sometimes would still need human intervention . If that is true then I guess the human driver would still be responsible. I guess we would have to wait and see how the game goes.

I think the way it will work is that Tesla will complete the FSD features by this year or next and then regulators may approve FSD in next 1-2 years with drivers still being primarily responsible.

I am just not sure at what point will regulators approve the FSD with Tesla being solely responsible. Maybe it just depends on Tesla agreeing to take responsibility or may be its more than that.

Effopec | 21 February, 2019

Just remember that this is quoted in Elon Time. You need to double it to do the conversion to standard time. Now if we would just convert to the Metric system we wouldn't have to make that conversion.

walnotr | 21 February, 2019

Make no mistake people, once you step into any form of conveyance you are taking responsibility for your fate.

HighlandPony | 21 February, 2019

A Delta airplane manufactured by Boeing crashes while flying on autopilot. Who is ultimately responsible? Boeing, Delta, the pilot?

rdavis | 21 February, 2019

@Jithesh,

FSD will be released for known, specific areas and it won't be level 5. Maybe level 4ish. It will require a human to be able to take over at any moment (I wouldn't be surprised if selfparking, pickup would require the attention of the owner via video). I would expect those features to be ready by mid 2020, but I doubt regulators will allow it for several years after. So purely as my buess I'm saying 2023-2025 for some form of LVL4 in restricted know areas. LVL 5 really isn't feasible until you pass laws requiring all cars to be autonomous and/or maybe some infrastructure upgrades. I don't see that for many many years... maybe 10-20, but that's a bit harder to determine.

rdavis | 21 February, 2019

@HighlandPony,

When you get to the point of all cars being self-driving... I'm not sure you'll see insurance in the same way you do know. Mostly because I don't believe you'll see car ownership being the same thing. Maybe more of a autonomous ride share model and/or monthly fee for use of a vehicle that is owned by the developing/manufacturing company like Tesla. If development moves more towards all cars being autonomous, I don't think you'll see the same ownership structure of cars as yo use today.

ajbutler45 | 21 February, 2019

Think I saw somewhere that companies have invested $21 Billion into self driving technology. Seems a lot of companies are very serious about it. One of them will do it and there will be a lot of pressure on regulators to approve it. Tesla has the lead in data collected, now see if they can turn it into functional programing.

thedrisin | 21 February, 2019

FSD is a proprietary name and can be whatever Tesla says it is. If the driver has to be ready to take over at any moment that is SAE level 3 Conditional Automation. Level 4, eyes off, may be a decade away.

TAC | 21 February, 2019

^ needed to be said thank you.

If Elon would chime in here I am sure he would say something along the lines of the FSD title “causing too much confusion”

rdavis | 21 February, 2019

@thedrisin, that's why is said "LVL4ish"

jefjes | 21 February, 2019

Still waiting on lane change without confirmation and smart summon that was supposed to be out this week I thought. Not holding my breath for FSD by the end of this year but when it is ready, I'll have my checkbook ready too.

rdavis | 21 February, 2019

To be clear, my LVL4ish comment is that the technology is there there to perform lvl4, but the regulators and/or infrastructure is not which would require the intervention of a person...

thedrisin | 21 February, 2019

Not just regulators/infrastructure but manufacturer has to accept responsibility as well as liability for their product. The regulators are not going to approve something with a manual plastered with warnings or something with an "-ish" in the description. And if it just some new features that require driver involvement, doesn't matter then.

JustSaying | 21 February, 2019

The only self driving I would like reasonably soon would be to set EAP when I get on the freeway and then wake me up / nudge me when I get to my exit. I will take it from there. A very safe trip. The nuances of FSD might take awhile as we wait working out the problems of the nut behind the other cars steering wheel.

Mike83 | 21 February, 2019

It's great so many have opinions that underestimate EM. Funny that no other company is even close.

M3phan | 21 February, 2019

@Mike83 so true

thedrisin | 21 February, 2019

Close only counts in horseshoes. High level automation needs to be exact, not close. It may be awhile before there is true eyes off automation.

terminator9 | 21 February, 2019

Apple, IBM, Microsoft were all close, Microsoft commercialized PCs successfully first. Apple and IBM made the first PCs. I always liked macs but they didn't go mainstream until very late after Apple was at the verge of bankruptcy and Jobs iPhone saved the company.

There were several people about to take first flight and some also did but Wright Brothers got it right first.

If you/Elons says Tesla would be the first to make a battery/car that will be able to go 1000 miles in one charge, I will totally believe it.

But for this, no so much. No matter how close they are. The current features, although very cool, are like 20/100 in the last 5 years and we are saying it will go 90/100 in the next 10 months. Sorry, that is very hard to believe. Again, I use EAP every day and it's getting better with every update but simply cannot go from 20/100 to 80/100 in 10 months.

Disclaimer again: I am a TESLA investor (long) and am betting on the success of the company. But unfortunately this is not plausible. It's like Trump saying the wall will be finished in the next 4 months. He is the president, he is probably very close, he can get the emergency funds any day, he might build his wall eventually but it is not plausible to believe it will happen in the next 4 months.

terminator9 | 21 February, 2019

Does Vegas have betting for this? :)

Hal Fisher | 21 February, 2019

Musk knows there’s no downside to being optimistic.

bj | 22 February, 2019

@terminator9 - “If you/Elons says Tesla would be the first to make a battery/car that will be able to go 1000 miles in one charge, I will totally believe it.”

USA should convert to metric, then Tesla could be the first to make a battery/car that will be able to go 1000 kilometres in one charge. Much easier!

Shock | 22 February, 2019

"I contacted Tesla last year and the rep told me self-driving will be available by 20/20. Need government approval. Just think of all those driving jobs lost."

Statements like this are just really laughable. BTW Musk promised two years out for FSD...in 2015. https://electrek.co/2015/12/21/tesla-ceo-elon-musk-drops-prediction-full...

He always couches these demonstrably wild claims (because they never, ever pan out and he's made many of them) by vaguely blaming regulators at the same time, as if the tech is almost ready but it may be longer because of regulation.

If anybody here actually thinks they will be able to tap a "come get me" on their phone and their Tesla, in the middle of a rain storm at night, is going to be autonomously navigating a mall parking lot, then taking them home while they read a book, bringing them home on the highway, and surface streets, by 2020, they're not just wrong but negligently naive.

That aside, this notion that these cars will be able to fully navigate city streets--but qualify that the driver needs to be ready to take over at any time!-- in 22 months is patent rubbish. You cannot deploy a system of this sort that's dealing with traffic lights and pedestrians and pretending that a now-marginally attentive driver is going to be suddenly able to pop up out of their daydream and suddenly avert an accident that the system was able to identify as a risk but not avoid itself.

These cars cannot even go down a freshly parked highway in good weather, autonomously changing lanes without requiring constant driver monitoring and y'all think next year they'll be making their way through a city? haha!

ODWms | 22 February, 2019

I love Tesla too. And I’m betting on the long term success of the company with my money. A lot of it. so I need them to be successful. But I think there’s a lot of hopeful naïveté where this subject is concerned.

Madatgascar | 22 February, 2019

I have to agree with Shock on this one. It’s 2019 and the systems still can’t avoid a stopped fire truck with lights a-flashing. Do you think you can go to sleep in 2020?

thedrisin | 22 February, 2019

It would be nice to get rid of the term "FSD." It is not well defined and means something different to everyone and is confusing. Discussion should be in the accepted SAE nomenclature. No "ish." It is or it isn't. If it is not level 4 by definition it is level 3.

Pages