Ford deleted all of its social content across some of its accounts yesterday, and many were wondering what that meant. The answer is a little boring on the surface. It’s Ford’s first new global ad campaign in more than a decade, replacing the anonymous and meaningless “Go Further” with a tagline that, at the very least, has the name of the company inside of it.
But what’s interesting about the “Ready, Set, Ford” global campaign to me isn’t the aesthetic. I’m fascinated by the strategy behind it, which seems to be rooted in fear and uncertainty. Specifically, it’s at least partially inspired by the uncertainty around artificial intelligence.
This is important! Volkswagen is spending a billion dollars, for instance, in hopes of saving months of development time for new vehicles. If it’s successful, that could save the company many billions. For Software-Defined Vehicles (SDVs), the prospect of AI could make your car more like “connected devices” than static machines.
According to the new-old-new CEO of Volvo, all of this is not just important, it’s existential. I’d say everything feels existential now. It sucks! I’m going to do The Morning Dump in reverse this morning, with the lead story at the bottom.
Ready, Seat, TMD!
‘AI Is Our Key To Greater Speed, Quality, And Competitiveness’ Says VW Group Board Member
One of my favorite films is the Billy Wilder-directed dark comedy The Apartment from 1960. It’s also probably one of the most overlooked films of the 21st century, lacking the awareness of Wilder’s other projects like Sunset Boulevard or Some Like It Hot.
Go watch it tonight.
I mention The Apartment because there’s a scene early on in the film that I think about all the time. It shows Jack Lemmon as an insurance clerk for the fictional Consolidated Life Insurance Company of New York. What strikes me here is that none of the people on this massive office floor (pictured above) represent a job that exists anymore. Take this huge scene of people, and leave two, and that’s what the present is.
Most of these jobs were replaced by computers, and not even particularly smart ones. Microsoft Office knocked out at least 75% of them, and most of that was probably Excel. That’s to say nothing of the people in the mailroom killed by Outlook.
This made Microsoft an extremely valuable company until Google made all of that stuff essentially free. This is the price of progress, and in economic terms, these represent huge gains in productivity. Filling in a bunch of spreadsheets by hand and typing them up is inefficient and not a particularly rewarding job.
Volkswagen started its slow march towards engineering prowess and capability around the same time that The Apartment was in theaters, but it was an approach built on hiring a ton of engineers. It wasn’t fast, and it was labor-intensive. As a company, VW is on its back foot. It needs to do something to catch up.
The plan, according to Volkswagen, is to leverage AI to make developing cars faster:
“With artificial intelligence, we are igniting the next stage on our path to becoming the global automotive tech driver”, says Hauke Stars, Member of the Board of Management for IT at the Volkswagen Group. “AI is our key to greater speed, quality, and competitiveness – across the entire value chain, from vehicle development to production. Our ambition is to accelerate our development of attractive, innovative vehicles and bring them to our customers faster than ever before. To achieve this, we deploy AI with purpose: scalable, responsible, and with clear industrial benefits. Our ambition: No process without AI.”
The company claims that it’s using 1,200 AI applications already, and that the more than $1 billion it’s spending now will save it about $4-5 billion by 2035. In particular, the company thinks it’ll speed up the process of designing and building cars:
In vehicle development, for example, the Volkswagen Group is building an AI-powered engineering environment together with its partner Dassault Systèmes – for all Group brands and across all regions. It is designed to support engineers through virtual testing and component simulations, significantly accelerating development processes. Alongside other initiatives, this collaboration aims to helping to shorten the product development cycle for Group brands to 36 months – or less – making it at least 25 percent (around 12 months) faster compared to today.
AI integration is also advancing in production: Leveraging the Volkswagen Group’s proprietary Digital Production Platform (DPP) – a “factory cloud” now connecting more than 40 sites – Volkswagen is continuously introducing new AI applications into its manufacturing processes. These help optimize the interaction of complex processes in vehicle assembly, contribute to more efficient use of energy and materials, reduce costs, and lower CO₂ emissions.
I have to be that guy and point out that Volkswagen tried to do this with software, creating a company initially called Car.Software, which became Cariad, which became such a failure that Volkswagen had to promise Rivian about $5 billion to help solve the problem it created for itself.
There’s the money you save by making the building of cars faster, but AI also has a role to play in the cars once built, at least according to the people who are supposed to know better.
Do People Think Of Cars As ‘Static Machines’?
There’s an S&P Global Mobility report out about the future of SDVs, and a lot of that has to do with developing a common software platform (as Rivian and Tesla have done) in order to run a vehicle, as opposed to trying to cobble together a bunch of systems from various suppliers.
AI, according to this report, has a role to play on the consumer level and not just the production level:
Christoph Grote explained how BMW is embedding intelligent lighting systems that respond to conditions in real time and digital key technology that lets drivers use smartphones instead of traditional fobs. These features show how AI makes vehicles more context-aware, enhancing both safety and personalization.
Furthermore, AI can analyze vast amounts of data from vehicle sensors to improve predictive maintenance, ensuring that vehicles are serviced before issues arise.
By weaving AI into SDV architecture, vehicles can stay updated and responsive for years after purchase, helping them feel less like static machines and more like connected devices that learn and adapt to the owner.
That static machine line is interesting. Obviously, cars move, and are not static in that sense. But they have historically been stuck in place technologically. Your car stops becoming more advanced the day it rolls out of the factory. With OTA updates and a more SDV approach, that may not be as true anymore.
Volvo CEO: ‘Some Companies Will Adapt To New Circumstances And Survive. Others Will Not’
I don’t know if I’ll be entirely retired when I’m 74, but I plan to at least be semi-retired. That’s not the case for Hakan Samuelsson. He’d left Volvo as CEO of the company, but an uncertain future brought Samuellson back to the company he used to run.
He did a long interview with Bloomberg and he touches on the moment we are in:
Q: How do you see the car industry evolving from here?
A: The industry will be electric — there’s no turning back. It may take a bit longer in some regions, but the direction is clear. In (about) 10 years, cars will all be electric and they will be lower cost.There will be new dominant players, exactly as Ford, GM, Toyota and Volkswagen were in the old world. In the new world, there will be two or three very strong Chinese brands. That makes the room for the old ones tougher. So this will trigger a (wave of) restructuring. Some companies will adapt to new circumstances and survive. Others will not.
I think that is largely true. Who will the old, tougher brands be? Ford is making a play for the future.
‘Ready Set Ford’ Has Roots In Everyone’s Insecurity
Everything is changing. I can feel it. I know it. To some degree, I want the change to come.
As a website, The Autopian is disadvantaged in the short term by its membership model. There are competitors who are able to fill their pages with AI slop or, somehow worse, human-written articles that are designed to appeal to specific Google products. While we don’t ignore the traffic that can be gained from various platforms, we aren’t going to do it at the cost of posting articles from paid consulting firms that make the site worse.
The uncertainty of when and how and how much weighs on me, as it should given that that’s my job.
Ford killed all of its social media content yesterday, and people noticed. What was happening? Was there new product coming?
Nope. The company is still there. There’s no immediate new product. There’s just a new global branding strategy that replaces “Go Further.”
I think this is an improvement and, also, I don’t think I’m going to think about a catchphrase all that much. There is an explanation of it from Ford’s marketing chief Lisa Materazzo in The Detroit News that I am thinking about:
“Ford is a very optimistic and resilient brand, and that’s reflected in this idea of Ready Set Ford,” she said during a briefing. “We really tapped into some universal global trends that we were seeing. And one of those trends that is probably 24 months or more in the making has been this idea that consumers can really feel overwhelmed. The term that kept popping up is called polycrisis. … People are a bit overwhelmed, probably starting right around the time of COVID: So, health concerns, financial concerns, technology concerns — AI, what does AI mean for me?
“There are so many things that people feel angst about,” she continued. “However, the really interesting insight in all of that is people are actually very optimistic and very resilient when you give them the tools and the capability to feel empowered. And that was one of our big insights, because we think that we do that better than anybody else through our products, services and experiences.”
What the uncertainty around AI means for me is that I still want a Maverick Hybrid.
What I’m Listening To While Writing TMD
As noted by Styx in their song “Mr. Roboto,” we all need control. I want control.
The Big Question
How do you feel about AI?
Top photo: Honda
AI sucks for a lot of different reasons.
But I expect the charade to continue to grow, and continue until something happens to convince the majority of us that it’s all bullshit.
As a matter of fact AI seems to be a good description of the bullshit coming out of Washington DC these days. On another note, be careful about who you send birthday cards to, and what you write on them. Even if it is fake news..Fake news, LOL.
YMMV as always.
How do you feel about AI?
It doesn’t matter how I feel. The machine has been switched on, the flywheel has been spun up, and the inertia will carry us along the inevitable path that an AI arms race leads to. I’m not an AI doomer, I’m actually excited to see how this plays out…I just think the odds are that it’s going to end badly…either because it didn’t work, or because it worked too well. In the meantime, we’ll consume a MASSIVE amount of energy to find out and create a whole new set of problems in the process.
When I read the VW presser concluding with “Our ambition: No process without AI” my first thought was ‘Whatever. Harmless PR chum cast out to lure investors into their net.’ All it’s saying is “Seeking alpha in the manufacturing sector? Choose #VW! We will AI all the things! Add #VWAGY to your portfolio TODAY!”
Then this shook me from my complacency:
“I have to be that guy and point out that Volkswagen tried to do this with software, creating a company initially called Car.Software, which became Cariad, which became such a failure that Volkswagen had to promise Rivian about $5 billion to help solve the problem it created for itself.”
Oh yeah. That happened. At the same company. Shining example of how corporations forget that while pandering to shareholders can add a juicy sumthin-sumthin to boost the short-term numbers, they ultimately exist because they sell actual products to individual people, and it hurts badly if those products are scheisse.
It’s as if it’s 1954 and they’re saying “Plastic is inarguably the future. Our ambition: No component without plastic!” Making a blanket statement with little support from practical reasoning is one thing, implementing it across a global organization it is a far loftier plane of collective foolishness.
No real AI use at my engineering job thankfully. Being on an old govt contractor weapons program probably helps.
The problem is that the term AI means too many things, but everyone focuses on LLMs like chatgpt. The underlying tech is really good at ingesting large quantities of data for analysis and synthesis, and I believe it is very helpful in the research community. From space exploration to weather prediction, the tech has the potential to be extremely useful.
The public-facing use of AI, however, is fucking shit. It’s social media all over again, where despite the numerous examples of the danger it poses to the general health of society, some assholes are making a lot of money off of it. So they continue to push it, consequences be damned. Oh sure a couple kids killed themselves, but we put up half-ass guard rails that get overridden the day we release them! We’re not stealing everyone’s material, we’re simply training!
For use in cars? Man, I don’t even like auto-sensing wipers; I will turn them on when I feel it necessary. I don’t want my car thinking it knows what lights/music/whatever is best. F off and let me decide.
I was in a Target last night, and the shelves were definitely NOT full to abundance as was normal. There was product, but not as much – particularly seasonal items. Is the economy near a breaking point? MASA – Make America Soviet Again.
Have noticed the same thing here.
Someone WILL NOT be receiving 30 dolls for Christmas this year…
How do you feel about AI?
Mixed.
Used it quite a bit and found it useful for parts of my job (summarizing papers, creating the framework of documents and stuff like that) and was a total failure at much of what I do (managing databases, writing T-SQL/PowerShell/KSH/PLSQL).
Most of my experience is with ChatGPT and I found that you need to be an expert in the technology you are asking about in order to catch the poor code, wrong answers (that AI pulled from some random forum) or the hallucinations that are all too common.
AI is particularly poor at writing performant SQL (any variety).
I am lucky that my management is not really pushing AI for what we do.
I work for a govt contractor on a weapons program, a very not exciting and old one, but I think that has helped prevent AI from becoming too intrusive at work. The most I’ve seen is that we have an AI writing assistant tool for our goals and accomplishments for the year. Which I find very funny, because to me it is the people running things saying “yeah all this performance metric crap is stupid, just let this tool tell your manager how well you did and how ethical you’ve been.”
Ai is an ugly tool for lazy people.
I work in tech and like every tech company we have a big push to use more AI. In my experience, it’s kind of crap. It’s very rare that it just gets things right and doesn’t need a bunch of human intervention to fix its mistakes. Most people I’ve talked to feel like they spend as much time babysitting the chat bot when they try to use one as they would just writing the code or whatever themselves.
That said, there are cases where it’s Good Enough(tm). I actually love the Google Meet AI meeting notes. Nobody likes taking notes and it frees up whoever would have done that to actually participate in the meeting. The AI notes aren’t anywhere near accurate, but they’re close enough to get the gist of what was said and for something like that it’s fine.
However, we looked at adding a chat bot to our Slack to answer questions and it was trainwreck. It got some things right, it got some things extremely wrong, and sometimes the errors were buried in a block of otherwise accurate text, which made it difficult to parse out what was fact and what was fiction. Which is a huge problem with the whole technology – it’s not designed to be correct with its answers, it’s designed to look correct with its answers.
And for the limited places that AI is actually useful, is it really worth the environmental cost we’re paying? I’m extremely unconvinced, especially because I don’t think AI models are going to get appreciably better. It’s emphatically not learning at an exponential rate the way some ignorant proponents will tell you it is.
At this point I hate AI because I think in the long run it will have more negative impact on society than positive, much like social media has. And social fallout is an externality to all of these companies, so they’re not going to give it a second thought.
I used AI to make my job faster. To such a degree that I can see needing fewer people in my position. It is a little scary.
Can I ask what you do? I’ve had the same corporate job since college graduation so I don’t have exposure to what roles AI actually helps in.
Broadly – risk and compliance consulting for IT systems.
Huh, I wouldn’t think AI would help a ton there but I have no knowledge of what’s involved. I read ArsTechnica regularly, and the amount of stuff I see about IT issues makes it seem exceedingly complex to get right. The vast majority of it goes over my head though, so I’ve probably lost the plot on it.
I’ve read that AI is going to impact a lot of knowledge-based jobs. I think mine fits into that. We’ll see, though.
Not great. Not because there isn’t use in LLMs and the like, but because the tech industry and many other industries have decided that AI is their hammer and every problem is a nail. Want to clean up your database? Deletion is cleaning, I guess. Want to write procedures with AI? You’ll spend as much time fixing factual errors and scrubbing references to procedures that don’t exist as you might have spent writing the procedure yourself.
It’s great for pattern recognition, since that’s what LLMs are–they recognize the patterns of words and phrases used when people write about topics and piece them together into something that sounds like what people would say. Can you feed it the data you want presented and have it clean it up into a few good paragraphs? Probably, but you better have someone familiar with the data check those paragraphs for accuracy. Can it find patterns you may have missed in the data? Absolutely, and you can then have an expert consider those patterns and determine relevance (this is an important step that some technophiles want to skip).
The problem I think we’re going to run into is with AIs training on AI data. After some time, all the “hallucinations” and quirks build on each other and you have something completely useless. Garbage in, garbage out, as they say.
We very much agree. 🙂
Agree! AI is terrible at anything database related. Seen too many instances were AI will write T-SQL code and add PostgreSQL or Oracle syntax. Not to mention how poor the performance of the code is.
Absolutely. The only people who get excited about AI writing code don’t understand AI or coding. In theory, if you trained one entirely on one programming language, it might do alright, but it’s still not likely to be good code. AI is already inefficient and wasteful; we don’t need it writing inefficient and wasteful code, too.
Looking forward to the massive wave of VW recalls in a few years
Like computers, AI is just another tool in the kit. All of these companies are falling into the same trap that Musky has. Computers made certain tasks vastly more efficient, thereby reducing the number of workers per function, but absolutely, categorically did not replace them. People still had to do the thinking and then program a computer to replicate specific processes the humans were performing, likely with a much lower error rate. That is reduction, not replacement. AI is exactly the same. Exactly, I repeat. It is a distinction without a difference. These machines do not think, they process. Merely using a different flow chart does not constitute anything remotely like thinking. It is not creating, it is processing. No matter how well done, that’s all it is.
This matters because, as Jason likes to point out, life is full of edge cases. AI is grossly, violently misnamed. It is not artificial intelligence, it is simulated intelligence. There is no actual thinking ever going on, just a simulation thereof which can indeed appear as thinking, but is just plain not. Has a semi-novel way of processing information been revealed by LLMs? Yes, it has, but it is not anything like human intelligence. Complicating matters is that the decision trees of the LLMs can be opaque to even the persons who wrote the code simply because back-tracing output involves too many decisions based on probabilistic equations, not true comprehension. It is not and never will be thinking, just fancier processing. That difference will always limit “AI” in the exact same ways computers have always had limits. It will never, ever be anything but simulations of thinking. And every edge case will stump AI in the identical way every computer in history get’s stumped by any such cases. AI will always be trapped by it’s code and the impossibility of true comprehension. It will never truly create, it will only ever regurgitate variations of the data available and never more.
/end screed.
If it is a tool it is a terrible one. Why would you want a tool that gives you wrong answers?
I totally agree it is terrible if misused, which will indeed happen. However, like any tool,
AISI will have uses that it will excel at, just like any computer. My point is thatAISI is not the panacea lot’s of people think it will be. You just cannot program out life’s edge cases no matter how much you want to. Those are the times it will fail, butAISI should do just fine when used appropriately.I’m glad I’m not the only one who thinks simulated intelligence would be a better name for what the technology is today.
Wonder if all the Dieselgate shenanigans are part of the training data.
AISI will certainly be used for nefarious purposes like cheating, so yeah, probably.AI sucks and I can’t wait for this bubble to burst
QOTD: I am fine with it. I use it to complement my work and to deliver faster. I’ll survive.
I am still just learning about the cost for the data centers, their energy requirements, and the external cost to groundwater. I don’t know enough yet to have an informed opinion, but I know it makes the hair on neck stand up a little bit. I am and always have been concerned with fresh water, AI or not.
Why do products need to be updated? This is BS that we’ve been sold by marketing parasites. Was the machine matching expectations upon purchase? Does it still work perfectly fine? If so, what is the need to change? At this point, very few things are lacking in features or capability that can/needs to be added and “improvements” are largely tacked on nonsense or excessive performance that serves no real purpose other than trying to convince fools that they need a new one. What they won’t give is the reliability and resiliency that they’re capable of so that we need to replace more often, but that schedule wasn’t short enough so then they went to changing styling a little and shuffling the board around every year or several and hope people are convinced the new pattern is so much better or the obviousness of it being new makes the consumer think the neighbors are impressed (they’re not, they don’t care, and if they did, those aren’t the people worth trying to impress) and that’s what we had with analog/independent devices. Even that wasn’t enough for them, so they came up with forced obsolescence through tying it to constantly connected software (almost exclusively written by hacks so that they need OTA updates and introducing larger security holes that now need OTA patches), where a perfectly functional item they did their damndest to build to fail hardware-wise is finally rendered unusable due to “necessary” software connectivity that they claim they have to sunset for the sake of maximizing the performance of dubiously improved new product. Almost nothing is real anymore and nearly everything is a scam.
I’m trying to force myself to start using the AI options our company has more – it’s very clearly becoming one of the more useful tools in the kit. Yesterday was the first time I actually tried out Copilot. Unfortunately, it only confirmed my plan of attack for a CAD-import issue and didn’t really help. That said, it did give a pretty exacting run-down of what I was already planning to do.
Where AI has really shined within the place I work for is with Python assistance. One of my bosses has used chat GPT to help write custom software that incorporates a bunch of graphics libraries which up until a few weeks ago, none of us knew existed. He’s not much of a coder, but has a basic working knowledge of the language and has managed to do some really impressive stuff in regards to processing security footage in a few hours that would’ve taken weeks to do previously.
As to Ford, their new ad campaign is making me Hungry For Apples for some reason.
“Go Further” was going to get them sued for false advertisement after all the recalls. Had to switch it up.
I would prefer my cars being more “static” with less OTA updates. I work in tech and feel like the more connected something is, the worse it is*.
* Don’t @ me. I know that an unconnected device can still be hacked and regular updates can add functionality. I’m generalizing here.
IoT certainly adds an interesting layer to a lot of things. People seem to love smart homes.
You know what I like? Timers and motion sensors. Cause the only thing I’ve ever felt needed automation in my home is the lights. They’re connected to power and that’s it.
And we all know what the ‘S’ in IoT stands for:
“Security”
Smart lighting is one of my favorite new-ish things. I love that I can be sitting on my back porch and turn on the downstairs spotlights without leaving my comfy chair.
Yeah, agreed. I love lighting control. But I keep it isolated on its own network, because if someone hacks my lighting… um…. I don’t much care. 🙂
Reminds me of one of my favorite tweets of all time.
Can confirm.
Though I wouldn’t leave the gun within reach of it.
Torch is going to LOVE the new AI- engineered cars. Three turn signals on one corner and none on the other side. Want a physical glovebox latch? There’s 3 of those too and they all work different ways. Wheels have a total of 24 spokes, best not to ask how they’re distributed and there are so many power window switches the AI hallucinates them running into the pattern molded into the door card…