Lucid and Mercedes EQS in the wild

[MoniputerLM]

In my experience, even very capable and well-rounded mechanical engineers do not understand that the software contained in a complex electromechanical product (car) can require 10x more effort to "complete" than the mechanical part.

Peter Rawlinson should have known better, or hired people that did! Also, IMHO the basic software design seems flawed and I think that it is being re-written which may be why there have not been any recent updates. Again, poor management.

 

[hydbob]

Peter Rawlinson should have known better, or hired people that did! Also, IMHO the basic software design seems flawed and I think that it is being re-written which may be why there have not been any recent updates. Again, poor management.

Someone said that their whole software team got sacked sometime during the process. Can't remember who said it though, it was on these forums

 

[joec]

Peter Rawlinson should have known better, or hired people that did! Also, IMHO the basic software design seems flawed and I think that it is being re-written which may be why there have not been any recent updates. Again, poor management.

Knowing what I know about software development, I can say if there are serious underlying architectural issues that need resolving (due to say, being rushed into showing off features for the executive team or the marketing team before production, high turnaround in staff, etc.) the correction can take way more time than you might think. What looks good in a demo and what works for end customers are usually two very different things.

That technical debt adds up quick, and it always bites you in the butt at the least convenient time.

Given the wake from sleep delays I've heard discussed here, that may very well be where Lucid is right now. It certainly feels as if there are some fundamentals that aren't quite sitting right.

Some bugs can be fixed in a few seconds. Some take months. And you never know which ones are which until you dig into them. And sometimes you can't even address the user-facing stuff because the foundation is so borked you need to rip out and rewrite a ton of code just to get back to a functioning system.

A lack of regular updates feels like they aren't doing anything, but it could mean they are doing way more work than if there were quick bug fix patches every week. The hope is on the other end of that work you have a system that can then be tweaked and updated cleanly and quickly in the future.

Should it have been "done right the first time"? Sure. But in many years working in this field, I've never seen it. Not because software managers are idiots (some are, some aren't) but because software teams are often handed different sets of priorities before and after shipping, and those priorities often don't complement each other.

Then again, if Lucid let their software developers set the priorities, they'd have a perfect, bug-free, solid code base that works perfectly in their first shipping car—in 2040.

 

[MoniputerLM]

Knowing what I know about software development, I can say if there are serious underlying architectural issues that need resolving (due to say, being rushed into showing off features for the executive team or the marketing team before production, high turnaround in staff, etc.) the correction can take way more time than you might think. What looks good in a demo and what works for end customers are usually two very different things.

That technical debt adds up quick, and it always bites you in the butt at the least convenient time.

Given the wake from sleep delays I've heard discussed here, that may very well be where Lucid is right now. It certainly feels as if there are some fundamentals that aren't quite sitting right.

Some bugs can be fixed in a few seconds. Some take months. And you never know which ones are which until you dig into them. And sometimes you can't even address the user-facing stuff because the foundation is so borked you need to rip out and rewrite a ton of code just to get back to a functioning system.

A lack of regular updates feels like they aren't doing anything, but it could mean they are doing way more work than if there were quick bug fix patches every week. The hope is on the other end of that work you have a system that can then be tweaked and updated cleanly and quickly in the future.

Should it have been "done right the first time"? Sure. But in many years working in this field, I've never seen it. Not because software managers are idiots (some are, some aren't) but because software teams are often handed different sets of priorities before and after shipping, and those priorities often don't complement each other.

Then again, if Lucid let their software developers set the priorities, they'd have a perfect, bug-free, solid code base that works perfectly in their first shipping car—in 2040.

I have developed applications software and device software and it can be done correctly.

It this point, all we can hope is that upper management knows that they screwed up and will fix it but it will take time.

 

[LCIDdreams]

In my experience, even very capable and well-rounded mechanical engineers do not understand that the software contained in a complex electromechanical product (car) can require 10x more effort to "complete" than the mechanical part.

Yes I doubt Peter Rawlinson would have had that much to do with software pre-Lucid, other than interfacing with the software engineers on issues related to drivetrain, brakes, stability co trip etc. “It’s just software, it can be fixed in an update” is a little overused in the auto industry IMO. How many years did it take Microsoft to deliver a version of Windows that didn’t BSOD several times a week? How long did it take Apple to develop a version of ITunes that is actually decent to use (never)?

 

[Bunnylebowski]

On the topic of this thread, driving back home from dinner there was a lot of traffic on a two way street near my house. I passed a silver EQS slowly, and several people on the street nearly snapped their neck looking at my car. Nobody noticed the EQS. The looks was not my top priority in a car, but it tells you something about the vision behind Lucid and that people are super fascinated by this car.

 

[MoniputerLM]

Yes I doubt Peter Rawlinson would have had that much to do with software pre-Lucid, other than interfacing with the software engineers on issues related to drivetrain, brakes, stability co trip etc. “It’s just software, it can be fixed in an update” is a little overused in the auto industry IMO. How many years did it take Microsoft to deliver a version of Windows that didn’t BSOD several times a week? How long did it take Apple to develop a version of ITunes that is actually decent to use (never)?

Sorry, Lucid does not get a pass for crap base software. If the self driving takes awhile, then OK, but, basic car functions, no excuse.

 

[borski]

Sorry, Lucid does not get a pass for crap base software. If the self driving takes awhile, then OK, but, basic car functions, no excuse.

Nobody is giving it a pass, but that doesn’t mean getting it right the first time is easy, either.

 

[MoniputerLM]

Nobody is giving it a pass, but that doesn’t mean getting it right the first time is easy, either.

I know, but the software in our cars is a year behind, IMHO.

Yes, I bitch about this a lot, but Lucid got so much right in the car so the really bad software just detracts from an otherwise nice car! Also, if I had know how bad the software was before I got my car, I might have just waited a year for the next high end Lucid instead.

Ok, rant over....