Motorcycle couriers and Deliverooligans

[Neilio]

So, TACC (and presumably by extension Autopilot) has a LONG way to go before it can cope with the motorcycle couriers that plague london.

A drive down Uxbridge road saw me having to switch it off as the car just stopped at everything. motrocycles flying round left right and centre and the poor little M3 just gave up, sat in the road beeping most of the time. Be interesting to see if they can ever write FSD software to cope with it.

 

[MrBadger]

 

[Neilio]

View attachment 597108

Agreed, my point being eventually FSD WILL have to drive on city streets, I can't see how it'll cope at the moment.

 

[MrBadger]

What you have at the moment is not for driving on city streets. So you cannot extrapolate its capabilities now with how it will perform with the future driving on city streets functionality. We may get a better idea of its limitations some time after the AP rewrite.

 

[Type2]

Wasn’t FSD supposed to be “feature complete” at the end of LAST year? In Feb 2019 Elon said:

“ I think we will be feature complete, full self-driving, this year. Meaning the car will be able to find you in a parking lot, pick you up and take you all the way to your destination without an intervention. This year. I would say I am of certain of that, that is not a question mark.”

That may have been for the American market, but even given the limitations of current legislation I would say that is still many years away.

 

[pdk42]

Let's face it, FSD is nowhere near being able to drive the car autonomously, not even on highways. In truth it's a fraudulent sell. It's not even as good as many competitive systems that claim only to do lane keeping and TACC.

 

[Neilio]

What you have at the moment is not for driving on city streets. So you cannot extrapolate its capabilities now with how it will perform with the future driving on city streets functionality. We may get a better idea of its limitations some time after the AP rewrite.

You can certainly see how far it has to come though. The answer is a VERY long way

 

[bhav]

Thank you @VanillaAir_UK for stating what many choose to ignore.

If autopilot and autosteer happen to work nicely on a city street that's basically luck. It's not designed, in its current form, to actually navigate any of those situations.

And no, we shouldn't be feeding data into this neural network by trying it anyway... That's Tesla's job to make it safe before we dangerously go around testing it on streets with so many variables that it's an accidental waiting to happen.

@Neilio the autonomy on city streets many many many years away.

 

[Js1977]

I think you will find that most people, mostly the ones that rarely go into london driving, will find it a challenge.

I think for FSD to work in the city streets will involve it be a levelled playing field for the most part since I dont think the FSD system will ever be able to cope with cheeky drivers/riders that weave/wonder in traffic (more often than not whilst texting on their mobiles), suddenly cut you up, tailgate, barge in and the list goes on as far as the human side of the driving.

As for the Motorcycle couriers deliveroo etc (for the most part) , I agree that they are abysmal and I noticed this the most when I rode into London last (quite some time ago when deliveroo and the likes was starting up). The L platers in particular are very bad since they've had no proper training and for the most part have no self preservation or so it seems.

This has now spread and every time I go into Bournemouth and its the same scenario where all these scooters adopt the same low standard of road manners.

My take is that for these scooters (mostly thats what they use) to be operated commercially, the riders should undergo a full motorbike test.

 

[Durzel]

To be honest I’m glad AP can be used everywhere, but you do have to respect the fact it’s not meant for that, not advertised as supporting it, etc and temper expectations accordingly.

The i3 I had previously had a form of self drive below 30mph, but wouldn’t activate anywhere except motorways! Very annoying as built up traffic exists on single and dual carriageways as well..

 

[Glan gluaisne]

The issue of FSD/Autopilot not being sufficiently mature to operate in most road conditions is valid, but public perception seems to be that it is a lot more capable than it really is. I can pretty much guarantee that when chatting to any non-Tesla owner the topic of Autopilot will come up very quickly. Everyone I've talked with about the car assumes it can drive itself, none seemed the slightest bit aware of the fact that both FSD and Autopilot are extremely limited at the moment.

I'd guess that it's only those like us, that have driven the car for a time and understand all the limitations that currently exist, who know that FSD/Autopilot is a long, long way from being a routine reality here. It's not helped by news articles highlighting self-driving capabilities, that often don't go into any detail about the limitations. For example, I was reading an article about Waymo in Phoenix, Arizona recently, and no mention was made that Waymo will only work within a very tightly mapped zone. It is not autonomous driving in the way that Tesla are trying to implement it, as Waymo have taken out the majority of the variables by just mapping, in great detail, every single road that a Waymo Robotaxi can use. Take a Waymo Robotaxi outside of that mapped zone and it cannot function fully autonomously.

 

[Roy W.]

Wasn’t FSD supposed to be “feature complete” at the end of LAST year?

Here’s a news story from 2016:

Tesla wants new self-driving tech to autonomously road trip from LA to New York

In it Elon was touting a fully autonomous drive between Los Angeles and New York would happen before the end of 2017. It didn’t.

As far as I know, it’s still never happened, but please correct me if I’m wrong.

I’m sure Elon believes what he’s saying at the time. Even though he’s a genius he is also an eternal optimist. The fact that he can persuade people to part with thousands of pounds/dollars for a mythical product really shows how talented he is!

 

[Glan gluaisne]

Perhaps it's Elon Musk's eternal optimism that has pushed so much innovation into every area of business and technology he's been involved in?

The idea of landing rockets on their tail safely, so they could be reused, seemed almost like science fiction ten years ago, yet now it's just routine, business as normal, for Space X (still seems amazing, though). There's little doubt that much of the technology in our cars would never have made it into them without Elon Musk's drive to do things differently.

I think that self-driving is just a great deal more difficult t achieve in real world conditions than anyone, even Elon Musk, could have predicted. It's almost certainly all the "little stuff" that makes it so difficult, rather than the relatively simply stuff, like keeping the car driving within a set lane.

 

[Roy W.]

Perhaps it's Elon Musk's eternal optimism that has pushed so much innovation into every area of business and technology he's been involved in?

The idea of landing rockets on their tail safely, so they could be reused, seemed almost like science fiction ten years ago, yet now it's just routine, business as normal, for Space X (still seems amazing, though). There's little doubt that much of the technology in our cars would never have made it into them without Elon Musk's drive to do things differently.

I think that self-driving is just a great deal more difficult t achieve in real world conditions than anyone, even Elon Musk, could have predicted. It's almost certainly all the "little stuff" that makes it so difficult, rather than the relatively simply stuff, like keeping the car driving within a set lane.

I agree completely. Elon really is a bona fide genius.

When you consider that NASA has been developing rocketry for 62 years, and newcomer Space X has been at it for 18 years, the amazing feats Space X have achieved are mind-boggling. It reminds me of the NASA days of the 1960s when they were racing to get to the moon, and really tried to innovate.

 

[LongRanger]

Honestly, I reckon Elon Musk should get into mag-lev trains and ditch the whole FSD thing. If we want mass-transit with safe journeying, the road network with dumb drivers and supposedly intelligent single-person transit machines is not the way to do it.

Really struggle to see the appeal of fully autonomous driving when it’s a car. It needs someone with the drive and creativity of Elon Musk to really nail mass-transit for medium and long distances overland. It just doesn’t make sense to me that without mass-transit being equally advanced tech we’ll ever avoid FSD phantom braking or swerving to avoid a Leyland coach-built bus with a human involved then crashing into another one coming the opposite direction.

I don’t doubt how incredible it is to get software and cameras “reading” the highly complex scenarios that it can already do (sort of), but fully autonomous accidents / swearing incidents just doesn’t do it for me.

My vote would be focussing on the charging infrastructure, power train and efficiency for cars/vans/trucks and handing over the autonomous bit to keep choreographed travel for public transport.

 

[M3noob]

Here’s a news story from 2016:

Tesla wants new self-driving tech to autonomously road trip from LA to New York

In it Elon was touting a fully autonomous drive between Los Angeles and New York would happen before the end of 2017. It didn’t.

As far as I know, it’s still never happened, but please correct me if I’m wrong.

I’m sure Elon believes what he’s saying at the time. Even though he’s a genius he is also an eternal optimist. The fact that he can persuade people to part with thousands of pounds/dollars for a mythical product really shows how talented he is!

The spirit of Steve Jobs informs him.

 

[NewbieT]

Honestly, I reckon Elon Musk should get into mag-lev trains and ditch the whole FSD thing. If we want mass-transit with safe journeying, the road network with dumb drivers and supposedly intelligent single-person transit machines is not the way to do it.

Really struggle to see the appeal of fully autonomous driving when it’s a car. It needs someone with the drive and creativity of Elon Musk to really nail mass-transit for medium and long distances overland. It just doesn’t make sense to me that without mass-transit being equally advanced tech we’ll ever avoid FSD phantom braking or swerving to avoid a Leyland coach-built bus with a human involved then crashing into another one coming the opposite direction.

I don’t doubt how incredible it is to get software and cameras “reading” the highly complex scenarios that it can already do (sort of), but fully autonomous accidents / swearing incidents just doesn’t do it for me.

My vote would be focussing on the charging infrastructure, power train and efficiency for cars/vans/trucks and handing over the autonomous bit to keep choreographed travel for public transport.

I’m not sure maglev trains will solve UK transport issues (HS2 won’t fix traffic in Brixton)...or were you thinking more of a tram replacement kind off maglev shuttle pod affair?

 

[Glan gluaisne]

We may find that some of the worst city transport problems reduce, as a consequence of changing working patterns. Around 25 years or so ago there was programme to try and shift government jobs out of London and the "greater south east" to other areas, followed a few years later by an initiative to try and encourage people to adopt more flexible working patterns, rather than coming into an office five days a week.

I think that, with the present situation being what it is, working patterns may well permanently change, such that many more people work from home at least some of the time. There are a few people living around here that used to commute into London every day, but who are now working from home most of the time, with maybe a trip into work once every week or two. Be interesting to see whether or not this brings about a long term change in city transport needs.

 

[NewbieT]

I agree I think travel demand will reduce as a result of flexible working & acceptance of VC. It’ll probably buy us another 15 to 20 years until population growth starts to catch up.

I think the real impact is on longer distances though, not local congestion. Long distance is the easy but incredibly expensive problem to solve. I used to commute to London from the North West once a month. HS2 will be great, only with MS Teams adoption that journey is probably now only twice a year (post Covid). HS2 starts to look eyewateringly expensive. Better to accelerate FTTP investment & universal service obligation instead.

But I still drive to the shops & my local office.

I think Hyperloop uses maglev technology but no idea if it would work for short distances.

 

[Glan gluaisne]

For about five years I commuted from South West Scotland to London twice a week, down and back on Tuesday, then again on Thursday. Had to be one of the maddest commutes ever, hour and a half drive to Glasgow Airport early in the morning, 07:15 shuttle to Heathrow, 40 minute tube ride to Whitehall. Then back again in the evening, getting the 19:15 shuttle back from Heathrow to Glasgow.

The mad thing was that I'd see the same faces on the shuttle every day. There were clearly a lot of people who were living within driving distance of Glasgow and commuting to London every day. I was just glad that I wasn't paying for my travel warrant, hate to think what it cost.