Savage Geese Takes a Long Look

[minivanasatruck]

I need to add that I think a key problem people are not identifying, isn’t so much software operations or features, but the integration of the software with vehicle hardware. I bet the system works perfect on their little desks and tests screens and then all hell breaks loose when it’s tasked with communicating with and verifying the status of various vehicle hardware components. We’ve seen this repeatedly. Owners seats lying flat so the car software didn’t recognize what state the hardware was in, same for windows being stuck down, same for keycard not being recognized, same for key fob, same for not being able to put car in reverse, or button push for auto on the HVAC not putting HVAC into auto, or rear climate coming on even though you turned it off. I think the actual software is just fine but it’s failing at multiple reproducible point when it comes to accurately communicating with and controlling vehicle hardware. False alerts galore, like headlight errors, dynamic handling errors, the list goes on. It’s like the car and software aren’t talking the same language to eachother.

When you frame it like that it clearly sounds like hardware driver issues. That's still a software issue, just at a lower level. I wonder if supply chain issues caused them to swap vendors on a few parts, contributing to the glitches we see now.

 

[TahoeMichael]

I just took the time to watch the video again, with a focus on the driving they did at the end. Their commentary about the on-road driving was very interesting, and reinforces my decision to buy a Gravity. I don't think they explicitly said, but I assume the on-road driving is with the same vehicle they had on the track with 22"/23" wheels, and their discussion about small impacts like expansion joints also makes me think that I was right to stay away from the largest wheel size.

Regarding the track portion of their testing: while it was interesting to see their lap time compared with other vehicles, and hear their commentary, I just don't care. No vehicle I own is ever going to spend a moment on a race track. I think the entire auto-journalism field drifts into entertainment instead of information when they conduct high-speed track testing. I don't care what it feels like to brake a 6000 lb Gravity going into a curve at 120 MPH, because I'm just never going to drive it that way. I know that high speed testing behavior does have some correlation with highway speed testing, but so what? 99.9% of purchasers will only experience vehicle dynamics at highway speeds, where even on mountain roads the curve radius never gets small, or on more twisty back roads where you're probably driving 20 MPH slower. If you want to usefully test what happens when you push the limits of the vehicle at higher speeds, design a test regime of emergency maneuvering tests that simulate unexpected obstacles on the highway, abrupt stop of traffic as you come around a curve, being cut off by someone (you assume) who's wearing a clown nose, etc. Try a test of how well it behaves when you exceed the limits of grip and cannot hold your desired line at 40 or 60 MPH. Test how well it behaves if you're panic steering around a deer in the road, and in the middle of that you transition from pavement to a packed dirt shoulder. All those scenarios are way more relevant to me than going into a track's curve at 120 MPH.

 

[Murcie]

He should've added a disclaimer that he's not responsible for stupid financial decisions as a result of watching his review lol, despite his honestly about the software-related issues

 

[PGHGravity]

Contrary to being “useless” I can say that Atmos music has turned my car into almost a second home for me, as I spend quite a bit of time just sitting and listening and not even driving.

The Atmos mix of Bohemian Rhapsody is pretty fun too.

Side note I would pay Tidal an extra $1/month for a switch to filter out everything besides the Atmos versions of things since they are so tedious to find otherwise.

 

[Rogue]

Hopefully without trying to sound critical of the Lucid software team, I'm not sure if we have any automotive software engineers on here, but any feedback on the words from the software team in the video.

For the software engineers and developers on here: Based on what the Lucid software team members said in the video, do you hear or get the sense that the software skillset/knowledge currently exists in-house for Lucid to have software that matches its engineering?

I know that's a hard, maybe impossible question to answer, so maybe a better question is: Do you see the software team as being on the right track for where they need to go?

My specialization and interests are in HPC and distributed systems with a focus on software architecture and performance optimizations. A good chunk of this was UI/UX, but there were a lot of hints and nudges beyond related to software architecture and non-UI/UX features.

As someone who envisioned possibly working at Lucid on the software side, the interview with those two engineers has completely and utterly eviscerated any desire.

Even if you were completely ignorant of the subject field, a bystander can surely understand and appreciate the day and night difference between them and the interview on suspension or any of the tech talks hosted by Lucid themselves. They are engineers who devote themselves to abstraction, smoke and mirrors, and all things other than building a solution most effective in tackling the objective outcomes. My impressions were that of individuals who build massive unsustainable monstrosities and call them beautiful to behold because of the sheer complexity of their creation and the idealized, purely conceptual achievements rather than the objective reality of the execution.

The vagueness of the answers with an utter lack of detail to the nuance of why or what concrete measurable outcome and associated trade-offs were numerous. As a fellow software engineer, the hardware matters and fundamentally alters your solution. To have two car models with fundamentally different hardware magically achieve nirvana of a singular codebase and design is not possible, or at the very least, the sacrifices necessary would give birth to a monstrosity and evil that far outweighs whatever slipshod excuse of benefits they are touting.

I wish I could unsee and unhear that part of the video. Ignorance is bliss. Maybe these two just happened to be the worst possible folks, but the fact Lucid offered them up expresses the sheer circus behind the scenes for the software. May the company have someone bold enough and empowered plus the technical competence to purge such individuals from the organization and achieve something worthwhile than the current folly they appear to be undertaking.

 

[Rogue]

I have given the company significant grace in previous posts up to this point. However, I don't know how you see past this for the software not related to driving dynamics type pieces.

 

[electric]

Contrary to being “useless” I can say that Atmos music has turned my car into almost a second home for me, as I spend quite a bit of time just sitting and listening and not even driving.

I just signed up for Tidal to see if I can appreciate Atmos music over plain old Spotify and it did not disappoint
Noticeable improvement in sound depth and richness on Atmos playlists, I'll be spending more time in a parked car now.

 

[TahoeMichael]

My specialization and interests are in HPC and distributed systems with a focus on software architecture and performance optimizations. A good chunk of this was UI/UX, but there were a lot of hints and nudges beyond related to software architecture and non-UI/UX features.

As someone who envisioned possibly working at Lucid on the software side, the interview with those two engineers has completely and utterly eviscerated any desire.

Even if you were completely ignorant of the subject field, a bystander can surely understand and appreciate the day and night difference between them and the interview on suspension or any of the tech talks hosted by Lucid themselves. They are engineers who devote themselves to abstraction, smoke and mirrors, and all things other than building a solution most effective in tackling the objective outcomes. My impressions were that of individuals who build massive unsustainable monstrosities and call them beautiful to behold because of the sheer complexity of their creation and the idealized, purely conceptual achievements rather than the objective reality of the execution.

The vagueness of the answers with an utter lack of detail to the nuance of why or what concrete measurable outcome and associated trade-offs were numerous. As a fellow software engineer, the hardware matters and fundamentally alters your solution. To have two car models with fundamentally different hardware magically achieve nirvana of a singular codebase and design is not possible, or at the very least, the sacrifices necessary would give birth to a monstrosity and evil that far outweighs whatever slipshod excuse of benefits they are touting.

I wish I could unsee and unhear that part of the video. Ignorance is bliss. Maybe these two just happened to be the worst possible folks, but the fact Lucid offered them up expresses the sheer circus behind the scenes for the software. May the company have someone bold enough and empowered plus the technical competence to purge such individuals from the organization and achieve something worthwhile than the current folly they appear to be undertaking.

I have a lot of expertise in distributed embedded systems. I don't take as harshly a negative view as you. Their discussion sounded to me like a classic case of software people glossing over the details for a non-technical audience. The lack of sensible software questions from the interviewer would seem to justify that approach. When talking between themselves, they alluded to just enough embedded systems architecture to make me think that they're not incompetent. Although one of them is obviously a UX person brought in from outside the automotive world.

The missing interview that would have been informative was talking to the most senior software quality person.

 

[Bunnylebowski]

The Atmos mix of Bohemian Rhapsody is pretty fun too.

Side note I would pay Tidal an extra $1/month for a switch to filter out everything besides the Atmos versions of things since they are so tedious to find otherwise.

Yeah it’s unfortunate they’re sandbagging one of the best things about Tidal. Apple Music just literally defaults to the Atmos version if your system is set up for Atmos, but that’s only on home systems, and I still have yet to see a proven example of a native Apple Music app correctly playing back Atmos in a vehicle. I think the reports of Rivian R1S doing it are incorrect as the speaker identifier tracks combine the heights and surrounds.

 

[PGHGravity]

Yeah it’s unfortunate they’re sandbagging one of the best things about Tidal. Apple Music just literally defaults to the Atmos version if your system is set up for Atmos, but that’s only on home systems, and I still have yet to see a proven example of a native Apple Music app correctly playing back Atmos in a vehicle. I think the reports of Rivian R1S doing it are incorrect as the speaker identifier tracks combine the heights and surrounds.

I don’t think I’m qualified to say if it’s playing correctly but Atmos in a Gen 2 Rivian definitely plays more surround-y than non-Atmos in the native Apple Music app.

 

[Bunnylebowski]

I don’t think I’m qualified to say if it’s playing correctly but Atmos in a Gen 2 Rivian definitely plays more surround-y than non-Atmos in the native Apple Music app.

Yeah it does but not discreet from what I had heard. Can you try this test track and see if it’s actually doing left rear height and right rear height correctly? Maybe in works in yours but I know another Rivian owner it didn’t play discreet for.

 

[gaskiller]

I’m sure they are already keenly aware of the consequences of releasing a fault prone vehicle to reviewers. I don’t think a single reviewed Gravity has been issue free thus far. Things will be great for them if they can wrangle control of these issues and rapidly solve them and I do think they’re very motivated to do so and are making progress at it. You can’t go back in time but being able to course correct is a huge asset. They’re taking steps to rectify the disaster that is Lucid Financial Services as well. I’ll try to remain positive and imagine a check list of issues that are getting crossed off quickly, and think happy thoughts. I don’t like it when Lucid makes me sad.

Well, they shouldn’t have let this happen……software needs to work from day 1, or at least day 365.

And I hope they fired the person who was responsible for these dumb keys

 

[borski]

To have two car models with fundamentally different hardware magically achieve nirvana of a singular codebase and design is not possible, or at the very least, the sacrifices necessary would give birth to a monstrosity and evil that far outweighs whatever slipshod excuse of benefits they are touting.

I think you’re making a lot of assumptions here. For example, it’s entirely possible that there ends up being another CCC upgrade for new Airs, and perhaps that is what’s unified with the Gravity infotainment codebase, or perhaps not.

It is entirely possible they were only referring to infotainment, for example, and that could be a unified codebase without too much craziness. Of course the firmware for the various components would be different, various under the hood systems would be different, etc.

But Android Auto is Android Auto, and they could theoretically unify the infotainment codebase at some point without it being insane.

 

[DeaneG]

I do wish the seats would heat and cool a little bit along with the HVAC

?? my Air does

 

[Smorgasbord]

As someone who envisioned possibly working at Lucid on the software side, the interview with those two engineers has completely and utterly eviscerated any desire.

Even if you were completely ignorant of the subject field, a bystander can surely understand and appreciate the day and night difference between them and the interview on suspension or any of the tech talks hosted by Lucid themselves.

I think it was the UX design lead and the manager of the infotainment software team, not two engineers. TBF, part of the problem was that they were "interviewing" themselves, so it was full of "as you know" instead of addressing someone outside the company, as the suspension engineer got to do. That said, I do not disagree that those two were not impressive.

Also, while suspension is important, infotainment has far less constraints, far less history anywhere (even Tesla), etc. than suspension. And watching the video, he wasn't that happy with how Lucid tuned the one newish thing in suspensions - the air suspension, comparing it unfavorably to Porsche. Although there is opportunity for Lucid to tune the software in the suspension via OTA updates.

The infotainment choice of Qualcomm and QNX with an Android hypervisor is pretty solid, actually. The problem is that they essentially started all over again. At some point, Lucid needs to stick with some basic architectural choices and then they can iterate improvements. Note that the keyfob is almost certainly from a third party company called Pektron (they did the Air as well as numerous other brand vehicles for decades now). So, that it doesn't work is Pektron's fault - but Lucid engineering management should have been all over them months ago to nail that out the gate. With Lucid's low volumes, Tier ones like Pektron don't pay enough attention nor put their best engineers on - they save those for the big deals with high volumes and multiple vehicle contracts.

One good side of using QNX, beside security and performance is that there are many implementations of CarPlay and Android Auto built on top of QNX, so it shouldn't be huge challenge for Lucid - except maybe that Apple tests pretty thoroughly, and if Lucid's rest of infotainment is buggy Apple will likely withhold final approval.

 

[DeaneG]

...Note that the keyfob is almost certainly from a third party company called Pektron (they did the Air as well as numerous other brand vehicles for decades now)...

Lucid switched to a different well established supplier, Marquardt, for Gravity's keys.

 

[msaunders9430]

My specialization and interests are in HPC and distributed systems with a focus on software architecture and performance optimizations. A good chunk of this was UI/UX, but there were a lot of hints and nudges beyond related to software architecture and non-UI/UX features.

As someone who envisioned possibly working at Lucid on the software side, the interview with those two engineers has completely and utterly eviscerated any desire.

Even if you were completely ignorant of the subject field, a bystander can surely understand and appreciate the day and night difference between them and the interview on suspension or any of the tech talks hosted by Lucid themselves. They are engineers who devote themselves to abstraction, smoke and mirrors, and all things other than building a solution most effective in tackling the objective outcomes. My impressions were that of individuals who build massive unsustainable monstrosities and call them beautiful to behold because of the sheer complexity of their creation and the idealized, purely conceptual achievements rather than the objective reality of the execution.

The vagueness of the answers with an utter lack of detail to the nuance of why or what concrete measurable outcome and associated trade-offs were numerous. As a fellow software engineer, the hardware matters and fundamentally alters your solution. To have two car models with fundamentally different hardware magically achieve nirvana of a singular codebase and design is not possible, or at the very least, the sacrifices necessary would give birth to a monstrosity and evil that far outweighs whatever slipshod excuse of benefits they are touting.

I wish I could unsee and unhear that part of the video. Ignorance is bliss. Maybe these two just happened to be the worst possible folks, but the fact Lucid offered them up expresses the sheer circus behind the scenes for the software. May the company have someone bold enough and empowered plus the technical competence to purge such individuals from the organization and achieve something worthwhile than the current folly they appear to be undertaking.

I have a lot of expertise in distributed embedded systems. I don't take as harshly a negative view as you. Their discussion sounded to me like a classic case of software people glossing over the details for a non-technical audience. The lack of sensible software questions from the interviewer would seem to justify that approach. When talking between themselves, they alluded to just enough embedded systems architecture to make me think that they're not incompetent. Although one of them is obviously a UX person brought in from outside the automotive world.

The missing interview that would have been informative was talking to the most senior software quality person.

Two very insightful responses that remind me of a guy I worked with for a few years during my work career.

Guy is hired by government agency to improve the integration of technology in agency operations.
Agency head at the time was not the most technology literate person, so new guy is able to impress agency head with seeming innovative ideas such that he is eventually promoted to role of Chief Technology Officer (CTO).
Agency head eventually retires. New agency head is very technology literate such that he is able to actually understand the innovative ideas of the CTO. New agency head starts holding CTO accountable by setting timelines and milestones. CTO struggles. The "conceptual achievements" that @Rogue mentioned are insufficient. New agency head seeks"execution". Within months of new agency head onboard, the CTO is escorted out of the building.

At some point that "senior software quality person" that @TahoeMichael mentioned, needs to be held accountable for execution from the software team.

I surmise that the future of the current interim CEO and software team leadership is highly dependent on the resolution of the software issues.
Nothing surprising with this supposition.
Happens with companies/organizations everyday; results matter.

 

[jered]

The biggest thing people are going to be critical of is what they see and it’s the software. While many of us die hard enthusiasts can appreciate the remarkable level engineering that has gone into the Gravity, to put it bluntly, the average car owner doesn’t give a shit.

I complain a fair bit about the software, but I do have to qualify that this is coming from an X. Despite all of the bugs and quirks, I think the UX for Gravity is quite good and I have hope for it becoming stable. It's also just so far beyond the painful mess that all of the non-tech-centric companies (Audi, BMW, Ford, Kia, etc) ship. Grading on a curve, they're doing great -- but for top of the market, they need to do better than that.

 

[petergottlieb]

Having lost all those things with our 2021 Tesla, I couldn't agree more. That car was a real pain in the tuchus to operate, especially as the automatic gear selector tried to drive us forward into the garage wall instead of backing out of the door as often as not. We ended up always using the swipe feature on the screen, which was only slightly less annoying. A quick finger flick of a lever without even taking your hand off the steering wheel was so much easier.

Tesla basically lost me with the swipe screen based drive selector. It was hard to use and clearly worse than a physical alternative. It was also for a function that has to work correctly and predictably every time. No thanks.

 

[TahoeMichael]

Two very insightful responses that remind me of a guy I worked with for a few years during my work career.

Guy is hired by government agency to improve the integration of technology in agency operations.
Agency head at the time was not the most technology literate person, so new guy is able to impress agency head with seeming innovative ideas such that he is eventually promoted to role of Chief Technology Officer (CTO).
Agency head eventually retires. New agency head is very technology literate such that he is able to actually understand the innovative ideas of the CTO. New agency head starts holding CTO accountable by setting timelines and milestones. CTO struggles. The "conceptual achievements" that @Rogue mentioned are insufficient. New agency head seeks"execution". Within months of new agency head onboard, the CTO is escorted out of the building.

At some point that "senior software quality person" that @TahoeMichael mentioned, needs to be held accountable for execution from the software team.

I surmise that the future of the current interim CEO and software team leadership is highly dependent on the resolution of the software issues.
Nothing surprising with this supposition.
Happens with companies/organizations everyday; results matter.

I would hope the head of software quality wouldn't also have responsibility for software development execution. One tends to value quality, one tends to value speed of development, and those are frequently at odds with each other. It's often tough for a single person to balance them well. The most functional engineering orgs I've been involved with have had an independent quality lead. That kind of independent voice for quality has gotten way less common in most of the tech industry, as most businesses value feature richness over stability. The embedded systems world is still struggling to adapt to the modern desire for extremely complex software in vehicles that also meets the traditional expectations of a vehicle that always just works. Lucid is far from being the only engineering org that's struggling in this regard.