Retail Auto Poised For Strong Finish, F-150 Lightning Production Pause, Is AI out of Training Data?

Paul J Daly  0:00  
Good morning, everybody. It's the Friday before Christmas, December 20. Everyone's going out to shopping. But we're here talking about auto retail being poised for a strong finish. F1 50 production, stopping an AI, running out of training, data out. Really want to know who acts that sound when you get to the bottom of a milkshake, they're all done. Every funny video

Unknown Speaker  0:22  
ASMR channel. That's a gripping milkshakes. Yeah.

Paul J Daly  0:25  
I mean, that's AI running out of data. Apparently, this week feel has felt long and short, long and

Kyle Mountsier  0:34  
short. I feel like next week is crashing into us with a brisk, you know, crash to the face type feel, I don't, you

Paul J Daly  0:41  
know next week's gonna feel like. So my wife and I went shot Christmas shopping last night. Thursdays kind of, we try to keep that as a date night. But typically, the one last one before Christmas, we go out and we just, you know, hit all the stores we can without any kids in tow, and we just zip it. We went to a lot of places we went to, like Party City, Barnes and Noble, Sephora. Cols,

Kyle Mountsier  1:02  
okay, I You said Sephora. I saw the LinkedIn post. I wasn't sure if it was Altair Sephora. Either way, that

Paul J Daly  1:08  
black and white motif, more yellow, yeah. Well, we'll talk more about that later, but what a mess. But what I realized was that it was dead, really I expected, like it's gonna be crazy, and it was dead, and I don't know this weekend is gonna be nuts. I think the feeling that people I have a weekend and then another day before Christmas, oh yeah, people are dragging on it. So I felt like I got away with something. Oh no, was parking up front. Because last weekend

Speaker 1  1:38  
you're like, Yeah, you like, robbed, robbed the week. It's

Paul J Daly  1:41  
amazing. I felt like and we're done, we're finished, done. Not another thing to buy. I'm so happy about that.

Kyle Mountsier  1:46  
Man. Hey, people yesterday, if you did not catch the year in extravaganza, you can go to our YouTube or go to our LinkedIn and go re watch the year end extravaganza, the stream yesterday. Watch it for the first time. It was unbelievable. Actually, I stepped out to watch it with my wife and kids yesterday, because we're homeschooled. I'm at home, right? My kids were like busted laughing on the floor. It was amazing. It was so good. So you got to check it out. There are some graded it's obviously some recap, some looking forward type stuff. Aired it

Paul J Daly  2:19  
out a little bit. If it's easier for you to get there, you can just go to asotu.com it's right up there at the top of the page, or at the top of the fold anyway. So yep, you can just go in and click through and watch it on LinkedIn. Lots of fun comments. And yeah, we pulled out a few things we've haven't pulled out. It was amazing. Little little little more humor, a little back and forth, some great segues. One of the funnest lower thirds I've seen, little surprise present lower third. It was nice, Yep, yeah, it was good. It was a great time. You should go check it out. We hope you do what else we talking about? Anybody else? Let's go Auto Collabs episode, is that out today?

Kyle Mountsier  2:51  
That's right. Matt Murray wide, well, no, that's been it's been out a couple days. You should check it out. That's a great that's a great podcast episode. It really gets into like, what's going on in the industry from, like, where how consumers are feeling about it based on a whole bunch of reviews. So, like, real customer words, really cool insights. And Matt Bernie is just a G. He's a good dude. He's a G. I've

Paul J Daly  3:11  
never heard you say that before. There you go. I'm wearing my ugly sweater. By the way.

Kyle Mountsier  3:16  
I'm proud of you. I don't even have one of those anymore. I wish it's not so ugly. I like that sweater, actually. I think it's dope, you know, just if you can't see it, it's got a whole bunch of like, you know, frilly things, and then it's got rain reopen motors,

Paul J Daly  3:29  
like, Yeah, I mean, it's not that ugly. You're right. I should have just wore a Dallas Cowboys. I feel like a Grinch when it comes to Christmas parties. I'm like, this is just a lot of work. Can we just get, like, really good food? It's a lot of work. It's a lot of work around here in the congruent

Speaker 1  3:50  
hallway. But proud of you for sustaining it. I'm

Paul J Daly  3:55  
festive though I got my ugly sweater. I was like, where am I gonna find this thing? Pain in your voice right now is significant. So once I'm into it, I'm fine. Yeah, I'm not big on dressing up. I

Speaker 1  4:07  
know you, I like you know you already had to dress up once. This

Paul J Daly  4:12  
way better. I'd rather do that again. I'd rather do that again. People really came for this. This morning. They really came for this. I'm sure they left for this is what they did. Let's talk about some news. So the auto industry is gearing up for a solid 2024 finish. As inventory stabilize aggressive year end deals are luring buyers back to showrooms. Analysts are remaining optimistic, optimistic about the continued momentum into 2025 despite all the policy uncertainties that we have, total 2024 sales projected to hit 15.8 to 15.9 vehicles, a solid increase from 15.6 last year, December, seasonally adjusted rate the SAR is expected to reach 17 point 2 million, the highest since 2021 GM is likely to hold this position as the top US auto maker with Honda having. The most gains, overtaking stellantis in market share due to well, we all know couple issues. Retail leasing is up 19% year over year. Even sales forecasted to grow 12% in q4 despite the falling interest rates below 7% for the first time in a year, high. Vehicle prices are pushing monthly payments up seven to $740

Unknown Speaker  5:23  
Wow.

Paul J Daly  5:24  
Wow. That's $150 higher than 2019 that's Jonathan smoke, our guy from Cox automotive says, quote, we've gone from wait to buy into a new marching cry of BUY NOW before prices and rates climb again. Wow. What a little seesaw on,

Kyle Mountsier  5:41  
kind of, yeah. I mean, this is still, there's so much to be determined in in 2025, and I, I'm hearing the backbeat of, like, just the side conversations at church and, you know, around, like people that I know, like people are wondering, should I buy right now? There's actually a lot of incentive for people to think about buying here at the end of the year. So I think it is this like great coming together, of could be an epic end of the year, which we always look forward to that. But you think about just the growth in in in the fourth quarter, I'm hearing a lot of people leads are a little bit up. In the first half of the month, the traffic is a little bit up. So this, this is, this is going to bode well. And, you know, obviously inventory is there. So getting this, that 17,000,017 point 2 million mark is that that would be really good for just the overall volume, even though dealers are feeling the pinch on profitability, abyss.

Paul J Daly  6:41  
Speaking of feeling the pinch,

Unknown Speaker  6:42  
oh. Speaking

Paul J Daly  6:45  
of new inventory, speaking of EVs, Ford has announced a temporary production pause in its f1 50 lightning supply, or production until January 6, 2025 the decisions reflecting a strategic adjustment to align supply with the actual market demand. What a concept, as the EV industry navigates a period of slower than expected growth. Basically, they sold 22,807 f1 50 lightnings this year. Mark again, 86% increase over last year. Well done. This performance places it well ahead of competitors like Chevy Silverado EV, which only sold 5200 vehicles and even rivians, you know, kind of cult favorite, R 1t only sold about half as Ford did on the lightning. The pause is going to be at Ford's Dearborn Michigan plant, and they're just trying to prevent overproduction, to ensure that they're not pushing more inventory, which is great, yeah, everyone should take a note. Ford strategy acknowledges that EV adoption is not growing as fast as earlier projections did suggest. So they're trying to balance the innovation with the market need. As Ford spokesperson explained, quote, we continue to adjust production for an optimal mix of sales growth and profitability. I mean, this feels like it feels like they're maturing now, yeah, definitely maturing. You get a little holiday break up there in Detroit, you know, or near Detroit, and,

Unknown Speaker  8:05  
you know,

Paul J Daly  8:08  
look, 86% growth.

Kyle Mountsier  8:09  
Any market anywhere, any business would get pumped on that, right? So it's still like, let's not take this as like, oh my goodness, the sky is falling, there's still a lot of progress when it comes to EVs. Obviously the f1 50 is still taking the truck market with the lightning. But I love to see these manufacturers. You know, there's been a couple manufacturers, especially American made cars, that are saying, hey, pull it back, just for a little while, reset. Let's see what going into the new year looks like. Yeah, and I think, like, this is, this is the ebb and flow of learning how to create the right supply for the demand in the market.

Paul J Daly  8:50  
Speaking of learning stock segue time, speaking of large learning models,

Kyle Mountsier  8:56  
all right, so in a talk at the neuro p s 2024, AI conference, former open AI chief scientist, Ilya Sumer, good job. If I nailed that, kudos, I'm sure reflected on the foundational 20, 2014 paper that shaped modern AI and looked ahead to what's going next. The era of pre training is ending as AI is running out of Quality internet data, marking the end of what we're what they're calling peak data, saying that data is the fossil fuel of AI suits COVID says that we should expect smarter AI agents, synthetic training data and models that think more like humans, and that future is a eyes could have Things like true reasoning, self awareness and unpredictable decision making. So basically, the raw data is running out. So it's like retraining itself at this point. Play

Paul J Daly  9:49  
the milkshake sound for a minute in your mind, whatever that sound is, and you're saying that all the data on the internet, right? There's only one and. Pretty much they've trained it on everything on the internet. Well,

Kyle Mountsier  10:03  
okay, so let's, let's be clear on that. And I've been talking about this a lot, like inside conversations and and I think I posted on it maybe a couple weeks ago, but the the reality is, is what, what you get is qualitative training data, right? So actually, and I think I talked about this on the podcast last week, but you know, Google has always kind of looked for what is the best data on the internet. How do we surface that to its users in search, right? And there are millions and millions of trailing pages, trailing content on the internet that never get delivered to an end user, and there has to become a point where these models, and these trainers, the companies that are training these models, get to get to a point and say, Ah, that's not really relevant data, or that's not factual data, or that's not data that's going to actually help the algorithm. It's going to that track form everything right it'll exactly and so you're running out of, like the qualitative data, and at a quantity that no longer serves the engine to make it know things better. So now what you have to get into the habit of, and what they're starting to learn, and where thinking and reasoning happens, and Google's Gemini has just come out with a reasoning engine last week in 2.0 but is to say, Okay, you have all the available data now make reason and think based on that data to extrapolate things that maybe we aren't able to feed You and relearn or create net new and so that's where, that's where this, like, you know, kind of infinite game of of data is, is heading toward,

Paul J Daly  11:48  
you know, when you bring that up, the thing that comes top of mind for me is the the big debate on, like, how is the AI trained to basically build a hierarchy of What is most important? Because that is going to, like, steer the reasoning Right. Like, for instance, if you know environmental impact is the most important thing, the most logical thing is to get rid of the humans. No, but I mean, I think that. I mean, that's like a kind of, like a very, high level over the top. Everybody laughs at that. But there are all these other assumptions in business, in industry and in, I mean, not even business, industry, other areas where all of a sudden that level of hierarchy and training is gonna become much more value or important, because all the data is usurped. And now we're saying, like, figure it out. But this is also, you know, we talk about, we've been talking a lot in the back channels about AI agents. And when you deploy an AI agent to have a mind of its own, basically, to go and do a task this, this really is gonna be interesting the next 12 months. On what's gonna happen, you know, Steve Greenfield and I, we were talking in the Euro extrava and extravaganza. He was like, man, we're gonna look at, you know, on a very base level like BDC activity in in 10 years from now, and what we have going on at this moment in BDCs is going to feel like a rotary phone. Yeah, for sure. I agree. So I don't know. We just report the news. We just talk about it. That's awesome context. I don't know what you're going to do with that today, except be aware. Then being aware is one of the best ways that you can get ahead, which is why we have an AI newsletter auto industry.ai. Got them. We'll drip you a little fee without overwhelming you. Once a week. You should sign up for it. But listen, it's Friday before a holiday. I know there's a lot of Christmas parties. There's ugly sweaters going on everywhere. And the reason I'll do the ugly sweater thing is because it's connected to the thing we should all be paying attention to you know what it is? It's the people you.

Transcribed by https://otter.ai

‍