Latest / BREAK/FIX the Motorsports & Vehicle Enthusiast Variety Show / AI and Cars: State of the Nation 2025

Transcript

AI and Cars: State of the Nation 2025 · BREAK/FIX the Motorsports & Vehicle Enthusiast Variety Show

• 0:00
  John Summers is the motoring historian.
• 0:02
  He was a company car thrashing technology sales rep that turned into a fairly inept
• 0:07
  sports bike rider hailing from California.
• 0:10
  He collects cars and bikes built with plenty of cheap and
• 0:12
  fast and not much reliable.
• 0:15
  On his show, he gets together with various co-hosts to talk about new
• 0:18
  and old cars driving motorbikes, motor racing, and motoring travel.
• 0:35
  Good day.
• 0:35
  Good morning, good afternoon.
• 0:37
  It is John Summers, the motoring historian.
• 0:40
  I went to another one of those Society of Automotive Engineer
• 0:43
  events in the, uh, in the valley.
• 0:47
  Sen. Once again, thank you for inviting me to, to the event and listener if you're
• 0:51
  interested in, in this kind of technology futures niche that I'm falling down,
• 0:57
  which as a car lover is a bit depressing, but if, if you are interested in it,
• 1:01
  there are other episodes in my archive.
• 1:04
  Cover Lidar and Vallejo, where the people that talked about that,
• 1:08
  there was a German software house that talked about the software
• 1:11
  defined car they were involved in.
• 1:13
  How can you make money off all the data streaming off the car That's lurking in
• 1:17
  my archive, the software defined car.
• 1:20
  There was also the safety people, and that one's auto live with
• 1:24
  a company that spoke to there.
• 1:26
  And that's, uh, it's in that heading in my, in my archive here.
• 1:30
  I was also the keynote speaker at the Automotive User Interface Conference
• 1:36
  that was at Stanford in 2024.
• 1:38
  And look, I'm, I'm mentioning all of this because I've been looking
• 1:41
  at this future stuff thanks to phe.
• 1:43
  I've sort of seen how it's happening, PHE and the Society of Automotive Engineers
• 1:48
  and, and this is another one in this series of me trying to understand.
• 1:55
  How Automobility is is changing, I think.
• 1:58
  And, and if you are listening to the pod, I hope I'm gonna be
• 2:01
  able to try and convey that now.
• 2:02
  In the past, it wasn't clear to me that the title that span, who's
• 2:07
  the SAE rep, gives the events in order to encourage people to come.
• 2:11
  It's slightly different from.
• 2:13
  Who the people actually are and what they may have actually crafted
• 2:17
  the presentation originally to do.
• 2:19
  So what do I mean by that?
• 2:21
  I, I mean that when these people come and present, usually they're
• 2:25
  doing a presentation about their company, what they do.
• 2:28
  You know, I work for Vallejo and this is the leader products that
• 2:33
  we do, but what's so skillful about the way that spend packages it up?
• 2:38
  Is the, well, let me give you the title of last night's event.
• 2:43
  It was productive adoption of AI in engineering design.
• 2:49
  And the guy who was presenting to us was, uh, I think the CEO of
• 2:54
  a company called Neural Concept.
• 2:56
  His name was Thomas v Sharma or v Sharma or something.
• 3:01
  He was, it was a very German name, but he was.
• 3:05
  Had a French accent, very excellent, excellent English, but had a French accent
• 3:09
  and had studied some college in Luanne, which is like well known as being like
• 3:15
  basically Luanne Polytechnic, but it's like, you know, world leading Swiss, like
• 3:20
  clever engineering kind of a university.
• 3:24
  So look, let me try and sum up what I learned first of all.
• 3:27
  So if you've got a limited attention span, you can learn the important
• 3:30
  bits and I can fill in all the details and share the notes later on.
• 3:34
  So the important bits here is the, this is saying, okay, so how is AI actually being
• 3:42
  used to develop the cars and the roads that we're actually going to experience
• 3:48
  and have to encounter even if we ourselves want nothing to do with ai and.
• 3:53
  AI powered cars and all that.
• 3:55
  It's, it's, you know, so on there to understand for us, lay people to
• 3:59
  understand how the world is changing.
• 4:01
  Let's understand the mechanics of, of what the future of driving and
• 4:05
  motoring is, is gonna look like here.
• 4:07
  He had.
• 4:09
  Three use cases, and I'll look at my notes in, in a moment.
• 4:13
  But really, I think this is the sort of takeaway a around it.
• 4:17
  I mean, he had more examples, but one example and the, and and an example he
• 4:22
  kept on coming back to was, uh, formula One aerodynamics and how when you use ai.
• 4:30
  Instead of exploring one idea, you might be able to explore five ideas and
• 4:37
  therefore, you know, in, in the week that you've got or in the hour that you've got.
• 4:43
  And he had a lot of metrics around just how quickly you can
• 4:47
  use AI for design iteration.
• 4:52
  The other thing that I thought was really interesting was he talked about
• 4:55
  at the moment, engineers need intuition.
• 5:00
  Their research is guided by intuition with ai.
• 5:05
  Our AI tool will help them probe in the right direction.
• 5:09
  So in other words, if you're a Formula One team, the data
• 5:12
  streaming off the car, the AI.
• 5:15
  Can help point you in the right direction rather than you needing
• 5:20
  somebody like Adrian Newey to be able to say, aha, this is my vision.
• 5:25
  This is where I think we should be going.
• 5:27
  And then the data can be used to craft and help you move in that kind of direction.
• 5:32
  So I think there's a subtle nuance there, and I hope I've made that clear.
• 5:37
  The third.
• 5:39
  Element was ingest workflows.
• 5:44
  This is, in other words, you know, you have a bunch of different engineering
• 5:48
  processes that lead to the final process, and what you can do is you can
• 5:52
  chuck it all into the AI and say, dude, how can you make this more efficient?
• 5:56
  And then it will, one of the examples that he had, it may not have been
• 6:01
  this specifically, but it illustrates.
• 6:03
  Kind of complicated problems that he's working on in the automotive space.
• 6:09
  All these EVs, their batteries need cooling.
• 6:12
  There's these cooling pads and the cooling pads basically feature water moving
• 6:17
  around like some kind of coolant moving around under the cooling pad, right?
• 6:21
  Well, the way in which you can make the water move.
• 6:25
  You know how it swigs and squawks, how it zigs and zags, what you do
• 6:30
  to make it move in a particular way.
• 6:32
  This has a very meaningful impact on how effective the thing is at cooling.
• 6:37
  If you think about the way that a traditional car radiator works and
• 6:41
  how, if you know an old one that's got a few stone chips in it and.
• 6:45
  Bashed and you know, it, it noticeably doesn't work as well.
• 6:49
  Right?
• 6:49
  That might be why the car overheats on that hot summer day
• 6:53
  climbing a along hill, right?
• 6:55
  So it's those kind of, uh, of ideas.
• 6:58
  So once upon a time, what an engineer would've done is be like, well, I know.
• 7:03
  From my experience designing other cold pads that if I make
• 7:07
  it do a figure of eight motion, say that worked well in the past.
• 7:11
  So that will be the beginning of my research.
• 7:14
  So in the first instance, right, what AI is gonna be able to do is help
• 7:17
  me iterate on that five 10 x faster by doing lots of different models.
• 7:23
  But what he's saying also is that I needed to be know that
• 7:29
  figure of eight thing what?
• 7:31
  We can do now is have the AI suggest, you know what, if you do zigzags
• 7:37
  all the way up and down zigzag, crazy zigzag, that might be better.
• 7:41
  Or, you know, how about Pacman a Pacman system?
• 7:44
  You know, where he goes around in a, in a Pacman kind of pattern.
• 7:47
  That might be better.
• 7:48
  Well now it's not it, it's not like this is.
• 7:52
  Better than the figure of eight.
• 7:54
  It's the beforehand you only had the figure of eight method, now you've
• 7:57
  got the figure of eight and the Pacman and the zigzag and whatever
• 8:01
  other stuff the AI's gonna pull out of its computer, silicon ass.
• 8:06
  Um, so that's kind of fascinating, isn't it?
• 8:08
  So that was the, the important bit.
• 8:12
  So let me give you some, some bullets around that.
• 8:14
  So how did he call it?
• 8:16
  He called the first one.
• 8:17
  We can capture the physics.
• 8:18
  For the Formula One team and somebody in the audience was saying, you work
• 8:22
  with, you know, Alfred Cent or whatever, Toro Rosso's called Racing Bulls,
• 8:26
  whatever Toro Ross is called nowadays.
• 8:28
  Anyway, he was saying they've done a lot better recently.
• 8:30
  I don't know if they have an unfollow Formula one, but did
• 8:33
  you have something to do with it?
• 8:34
  And the guy was like quite smug, but obviously felt as if he did.
• 8:37
  So that was really cool to experience.
• 8:40
  So that's case one.
• 8:41
  Capture the physics, right?
• 8:43
  Case two, is this a, a design becomes more efficient, right?
• 8:49
  And the example here was Marla, the piston people who were now
• 8:53
  developing electric motors.
• 8:54
  Of course, the motor that they developed using AI was 15% more efficient and
• 9:00
  three decibels quieter thanks to to ai.
• 9:04
  You also talked a lot about how the engineers then went back to see what
• 9:10
  the AI had done that made it 15% more efficient and three, so in other
• 9:15
  words, there, oh, an engineering intuition was being informed by the ai.
• 9:21
  This symbiosis is really weird, isn't it?
• 9:24
  We've seen it before, right?
• 9:25
  Initially, the car was just a way to get around.
• 9:28
  Then suddenly it gave us the suburbs and suddenly it gave us mass consumer credit.
• 9:33
  Didn't it?
• 9:33
  Suddenly all that stuff happened in the twenties.
• 9:35
  Right?
• 9:36
  But it's just fascinating stuff, isn't it?
• 9:38
  The example for the ingestion of workflows, and I thought this was amusing
• 9:42
  given her the lead, our presenter, um, was Vallejo and the idea of, of
• 9:49
  their product is to unite these sort of three efficiencies, uh, together.
• 9:55
  So that was the sort of cliff notes, I guess a lot of people.
• 9:58
  But I mean, yeah, you can turn off now.
• 10:00
  I mean, I does say a lot of people would've turned off 'cause of my
• 10:03
  useless prattling before then, but that's my, uh, my attempt to
• 10:06
  do, uh, to, to do a sum up there.
• 10:09
  Let me move through my notes now.
• 10:12
  Flashing out what we just talked about, let me also mention that.
• 10:16
  Span whilst we were there, had a couple of giveaways sa, international Edge
• 10:23
  Artificial Intelligence in Connected, cooperative, and Automated Mobility.
• 10:29
  I haven't even had a look at that yet, but it's like this
• 10:32
  is like a pamphlet kind of guy.
• 10:34
  I guess it's more of that software defined car stuff that I've talked
• 10:38
  about in other presentations.
• 10:40
  This is definitely this material edges right towards my know
• 10:45
  the enemy kind of thing.
• 10:46
  Understand how they're deve inventing.
• 10:49
  Cars and driving.
• 10:50
  Now that's interesting, right?
• 10:52
  I that I funny that I thought of that because they laid on nice food.
• 10:56
  Thanks for that, Stan.
• 10:58
  And, uh, after I scarfed the nice food, I was, uh, hovering about to, uh, to
• 11:04
  sit down and, uh, somebody struck up with me and I thought I recognized
• 11:08
  him and he was like, I, I know you.
• 11:09
  And it was one of the people that I presented to when I went and
• 11:13
  presented five great BMWs to BMWs.
• 11:17
  Group of designers in, um, Silicon Valley there, he was one of the people that had
• 11:22
  had attended, and in fact, I was thinking of him just recently because he had been
• 11:28
  involved in the development of the BMWS 1000 RR Uber Sports Bike, and we talked
• 11:36
  about how his insight for me was how.
• 11:41
  MW Faithful were all about the tele leave of suspension.
• 11:45
  And of course the S 1000 RR was away from that.
• 11:48
  It was, we're just gonna do things the way, basically the Japanese do
• 11:52
  it with the upside down conventional forks and you know, BMW will.
• 11:57
  Engineers felt that their tele lever solution was better.
• 12:00
  So there was resistance around that internally.
• 12:03
  So, uh, the S 1000 RR had to be a knockout because just internally
• 12:10
  it would've crippled the careers the people involved had it.
• 12:14
  So he's passionate about the S 1000.
• 12:16
  RR was really pleased that I'd included it in my like list of like five great BMWs.
• 12:23
  He showed me he'd recently been to Munich.
• 12:25
  There was a BMW Design classic studio, and he was scrolling through the photos
• 12:30
  that he had until he landed on the.
• 12:33
  PK Era Bra and BMW Formula One Car, which had been another one of my
• 12:39
  great cars that I just felt showed like BMW's immense design superiority
• 12:45
  twice in Formula One with that bra and BMW of the early eighties.
• 12:49
  This was a. Basically A BMW 2002, you know, 1,602 block with a special
• 12:56
  head on it that could take the turbo.
• 12:58
  And those crazy statistics, like 1500 horsepower statistic that you
• 13:02
  always hear being bandied about the Turbo Era Formula One cast.
• 13:07
  That was the BMW in line four.
• 13:09
  That produced that statistic.
• 13:11
  My understanding is as a flash reading, so you know, certainly
• 13:16
  centers Lotus Reno of the mid eighties.
• 13:18
  They reckoned 1300.
• 13:20
  1350 in qualified people may know more now.
• 13:23
  I've not been on the forums.
• 13:25
  That's a, a different kind of rat hole.
• 13:26
  But yeah, so I was chatting with this BMW fellow, he was showing me these
• 13:32
  pictures of these awesome classic BMWs and saying that he'd thought
• 13:35
  of me and I was musing on my, uh.
• 13:38
  Role as, as influencer and I, as we were stood there span was like,
• 13:43
  you can take all these books away.
• 13:45
  Well, people rushed up and took away all the books.
• 13:47
  So the Edge Artificial Intelligence pamphlet was the only thing that I
• 13:52
  got from the pamphlets, and the only book that I got was this Biblical
• 13:56
  looking tome, which is entitled Transport Transitions, advancing,
• 14:03
  sustainable and Inclusive Mobility.
• 14:06
  Proceedings of the 10th TRA conference, 2024 Dublin Island volume two.
• 14:13
  Sustainable transport development, and now this is full on no thine enemy stuff.
• 14:19
  This is, if you're familiar with the way that modern cities are creating like
• 14:25
  exclusion zones by putting bollards and.
• 14:28
  Corralling the car corralling cars into, uh, lanes that
• 14:32
  are wide enough for the cars.
• 14:34
  Just, but you know, that extra space that they used to be for you to
• 14:39
  like cut the corner a little bit.
• 14:40
  Now there's room for pedestrians to stand in the road or in the case of San
• 14:44
  Francisco, FAL attics to fall in the road in that space, and they put these annoying
• 14:49
  bollards around and, and all of that.
• 14:52
  That's, these are the people.
• 14:53
  They were creating the theory.
• 14:55
  But behind that, another one of these pods that I did when I was asked to do
• 15:00
  a keynote at the Auto User Interface Conference, uh, in 2024 last year,
• 15:08
  I, a number of the sessions that I went to focused on, on this kind of
• 15:13
  sustainability and, and, and so on.
• 15:16
  It really feels to me a lot like, well, think about the way
• 15:21
  that cars of the 1960s were.
• 15:25
  And then compare them with the way the cars of the 1970s had those big ugly
• 15:30
  bumpers and big pillars, and were heavy and were unresponsive and de-tuned
• 15:38
  and, and really the whole freedom of design and freedom of expression and
• 15:44
  the sheer joy of motoring it left.
• 15:48
  Didn't it at that time?
• 15:49
  Well, that I feel like is what's happening with these people who are doing this.
• 15:56
  It's interesting.
• 15:57
  That should sit at the nexus of both of 'em.
• 15:59
  'cause it's all traffic technology.
• 16:01
  It's interesting that the, uh, AI systems that have been developed on
• 16:06
  Formula One or that, one of the other applications that this guy talked about
• 16:10
  last night was this SP 80, which is the.
• 16:15
  Waterland speed record for unpowered vehicles, so for sale vehicles.
• 16:20
  So this is like a clay foil thing.
• 16:22
  I'll put a link in for it.
• 16:23
  This was learning about, this is one of the most awesome things that I've learned
• 16:28
  and, and seen about for, for some time.
• 16:30
  So it's interesting.
• 16:31
  There's this exciting sort of motoring adventure that sits at the
• 16:36
  same time as this really dystopian suffocating, the freedom of motoring.
• 16:44
  I might even review this properly.
• 16:46
  I might have a proper look through it.
• 16:47
  I've not done more than flick through it, so there, I've,
• 16:49
  I've prejudged it, haven't I?
• 16:51
  What kind of an asshole YouTube, but am I to prejudge the, uh,
• 16:57
  transport transitions, advancing sustainability and inclusive mobility?
• 17:02
  My god,
• 17:05
  enough for the preamble and the setting up and the setting the
• 17:10
  stage and, uh, and, and all of that.
• 17:13
  Productive adoption of AI and engineering design.
• 17:16
  So it's AI workflow processes.
• 17:19
  So his, so the speaker's background is that he developed some AI work
• 17:23
  processes when he was at that low sand college that I talked about,
• 17:27
  which became the industry standard as neuro concepts about a hundred people.
• 17:30
  Now, automotive is their strongest sector.
• 17:34
  I've got a photo.
• 17:36
  I mean, it's basically like an open box and all the flaps of the box
• 17:40
  are like all of you know the ways that the product can connect and
• 17:44
  the partnerships that they have.
• 17:46
  And then the inside of the box is their AI shit stuff out of the flaps of the box
• 17:52
  with stuff like, you know, the Nvidia.
• 17:55
  Net core network infrastructure, which I assume is something like,
• 17:59
  you know, the Microsoft vendor network that I sold through when,
• 18:03
  back in the days when I was a, a tech salesperson 25 years ago, I could be
• 18:07
  wrong, could be talking outta my ass.
• 18:09
  I've included the slide.
• 18:11
  If you can be asked, go to the notes and click on it.
• 18:14
  11 OEMs in automotive, 18 tier one suppliers.
• 18:19
  They're also managing a growing expert community, whatever the fuck that May.
• 18:24
  The next was where we, we talked about these three main pillars, and that was
• 18:28
  the language that he used AI to capture.
• 18:31
  The physics IE, what arrow package should I use In Formula one is the defined.
• 18:38
  Requirement and you know, it can really help with that.
• 18:41
  And the example that I gave earlier about the figure of eight
• 18:44
  flow or the PackMan flow of water around a battery cooling pack.
• 18:51
  You know, the same if you think of it, aero swirl.
• 18:54
  It's a similar kind of super complex modeling.
• 18:57
  In fact, another use case they had was for an architecture firm
• 19:02
  who used to build these models to create the wind tunnel effect.
• 19:06
  So that they could forecast a wind tunnel effect on pedestrians 'cause otherwise
• 19:11
  in San Francisco's a bit like this.
• 19:13
  And you know, a lot of American cities are, the streets become, when they
• 19:16
  have the high rises, it makes like a wind tunnel kind of bacteria effect.
• 19:20
  So it's really horrible walking as a pedestrian, especially if it's raining.
• 19:24
  'cause it blows the bad weather right up in your face kind of thing.
• 19:28
  Well.
• 19:29
  Nowadays, rather than having to like make the models and take two
• 19:32
  weeks to like get results, nowadays the AI can do it from in 10 minutes
• 19:36
  flat, so that's another application.
• 19:38
  It seemed to me that was pretty much the same as how do you shape the
• 19:41
  error foil on a formula of one card.
• 19:43
  Not that different from, I'll come onto the water speed record one in a minute.
• 19:49
  'cause that is more complicated.
• 19:50
  Well, I'll talk about it now.
• 19:52
  It is more complicated, right?
• 19:53
  Because it's, it's not just the shape of the foil, it's
• 19:56
  the flexibility of the foil.
• 19:58
  It's not just the shape and physics of the water.
• 20:01
  It's also the pull of the kite on the foil.
• 20:07
  Dragging it along.
• 20:08
  And then there's the wind as well.
• 20:10
  So that's a really super complex piece of modeling.
• 20:13
  It's also the single coolest application and four oh one's pretty
• 20:17
  cool, but this foil, SP 80 waterland speed record thing is so cool.
• 20:22
  So just to be clear on that, Waterland speed record at the moment is 65 knots.
• 20:27
  They're trying to do.
• 20:28
  80. He was saying last night they've got up to 55.
• 20:31
  And the little bit of footage of it, my God.
• 20:33
  I mean, 'cause you all know 55 knots on water that's gonna feel like a
• 20:37
  hundred miles an hour on land and it's rough, you know, anyway, whatever.
• 20:42
  So I was pretty taken with, uh, with that.
• 20:44
  Wind and foil and all of that kind of stuff.
• 20:47
  That to me was cooler than the formula one.
• 20:49
  Anyway, that's pillar one, capturing the physics pillar two,
• 20:53
  you know, capturing the physics, what should I do, kind of thing.
• 20:56
  Pure use case.
• 20:57
  Pillar two is the MLA electric motor example where it's, you know, where they
• 21:03
  built a more efficient motor more quickly.
• 21:06
  And the measure of it was 15% more efficient and three decibels quieter.
• 21:11
  And then we talked also about, because the way their software works,
• 21:15
  you can drill into what the AI did.
• 21:18
  Your engineers know what it did, so they can learn better next time and
• 21:22
  have better ideas next time so they can inform the AI better next time.
• 21:27
  So, and, and you know, that's kind of obvious, right?
• 21:29
  If you've used any ai, as you well know, if you put shit in, you get shit out.
• 21:34
  If you asked it, who won the Battle of Gettysburg?
• 21:36
  And why?
• 21:37
  If you a good example, if you ask it, who won the Battle of Berrys burg?
• 21:42
  It gives you some total bullshit 'cause it doesn't know what you're talking about.
• 21:45
  It takes a guess.
• 21:47
  It might be right, it might be wrong.
• 21:49
  You have to be informed to understand that I was to illustrate this and
• 21:53
  this is where AI is at the moment.
• 21:55
  To illustrate this, I was researching earlier today a particularly
• 22:00
  gruesome shooting incident in London's history where three
• 22:04
  policemen were shot dead by gang.
• 22:08
  Members, basically what we call now gang members.
• 22:12
  1966, it was Harry Roberts was the main villain, but one of the other villains
• 22:18
  was, uh, was this guy John Duddy.
• 22:22
  Well, I was googling up John Duddy, but AI kept on wanting to give me answers
• 22:27
  about John Duddy, who's some Irish boxer.
• 22:30
  So that illustrates how, if I hadn't have been on the case to say, no,
• 22:34
  not John Duddy, the Irish boxer.
• 22:37
  I mean, John Duddy and Harry Roberts, who were involved in this rather,
• 22:43
  uh, unpleasant shooting incident in, uh, in London in the 1960s.
• 22:49
  The bad guys.
• 22:50
  I don't.
• 22:50
  They they, it was a standard vanguard involved in the shooting.
• 22:54
  If you picture up Google up a standard vanguard.
• 22:57
  I mean, the cars as ugly as the, uh, as the story and the people involved in it.
• 23:03
  And it gives you a real feeling of this sort of guy Richie
• 23:08
  Craze kind of underworld London.
• 23:10
  But it was still shocking because it was a shooting.
• 23:13
  I guess at that time, villains used like knives or you know, one of the three,
• 23:18
  the three that went to jail that didn't actually shoot any of the cops, but
• 23:21
  went to jail anyway, when he came out of jail, he was released in the 1980s
• 23:27
  and uh, he was like living in a flat in Bristol and his flatmate, who was a heroin
• 23:32
  addict, beat him to death with a hammer.
• 23:34
  I was like, God damnit, there's some retribution there.
• 23:38
  But there isn't, right?
• 23:39
  Because the main villain, it was a really nasty piece of work.
• 23:42
  This Harry Roberts, he's still alive.
• 23:46
  They released him in 2014 as far as I can make out.
• 23:49
  The bastard is still alive.
• 23:51
  All of the accounts talk about how he gloried in killing.
• 23:55
  Apparently he learned to do it when he was in the British Army on national service.
• 24:00
  There were some like rebellion and he records.
• 24:02
  He killed four people brags in the Wikipedia entry about
• 24:07
  killing four people there.
• 24:08
  That's when he really.
• 24:09
  Decided he enjoyed killing people.
• 24:12
  My God, how did people become so twisted?
• 24:15
  But anyway, back to where we were going.
• 24:18
  And the third pillar, so one is capture the physics two is
• 24:23
  more efficient, more effective.
• 24:25
  The Marlo electric motor.
• 24:28
  Three is the Valeo example of ingesting workflows.
• 24:33
  So the overall process of developing whatever widget it is from cradle
• 24:38
  to grave becomes faster and and AI informed, and then that offers
• 24:42
  the same kind of circular flow.
• 24:45
  Of learning that we talked about, in example, number two, Mar electric
• 24:50
  motor, and this is where I've really gotta take my hat off to
• 24:53
  spend for the formatting spend.
• 24:56
  Then asked a question and this led to a backwards and forwards with the
• 25:00
  audience, which was really cool because what I'm trying to understand and
• 25:05
  what I'm trying to share with you, the listener, is where are we with this shit?
• 25:10
  Like, is this bloke talking about vaporware or is this something which
• 25:15
  is actually gonna be on the next BMW?
• 25:17
  Like, 'cause there's three guards from BMW sitting right over there and the
• 25:22
  campus were on like opposite across the road from the building we were
• 25:26
  in was Ford's base in Silicon Valley.
• 25:29
  So there were probably four people there as well.
• 25:31
  I mean, I don't know, nobody was wearing a Ford badge.
• 25:34
  I mean, everyone's just wearing like.
• 25:36
  Engineer Silicon Valley neutrals, aren't they?
• 25:39
  You know, the Rohan pullover and the uh, jeans kind of thing.
• 25:44
  But anyway, yeah, fan had a question about.
• 25:47
  AI versus traditional simulations and the IE. Why would I use AI
• 25:55
  instead of a traditional simulation?
• 25:58
  And the answer is that AI can be predictive.
• 26:01
  So it turns days of work into the work of minutes or seconds.
• 26:05
  It's an evolution of current processes and makes them faster.
• 26:11
  So it's not like you're chucking out everything you did before.
• 26:14
  Fuck on that bollocks.
• 26:15
  We're using the AI now.
• 26:17
  No, it's not that.
• 26:18
  Rather it is looking at your current products.
• 26:21
  Allowing the AI to ingest the workflow and then looking at what
• 26:27
  efficiencies it might suggest, and I find myself thinking as I'm saying
• 26:32
  this, the evolution of sports bikes.
• 26:34
  And there's a jigsaw video, which I'll add a link to, which shows
• 26:38
  the evolution of the Suki S xr.
• 26:41
  When you were living here, it didn't seem like it had changed very much.
• 26:45
  You know, 'cause each year they only looked a little bit different and
• 26:48
  yeah, they were evolving generations.
• 26:50
  But even when the generations evolved, you know, when they went to like water
• 26:54
  cooling and fuel injection, you know, that seemed like an enormous change.
• 26:59
  And I guess it was, but it's still an evolution, right?
• 27:02
  It was still basically a G six R if you sat on the previous generation and you
• 27:08
  get on, you know, the all new water cool fuel injected, when it still felt like
• 27:12
  a big, heavy, scary G six R. A few years later, they'd evolved to kind of feel
• 27:18
  like 600 had a few years ago, right?
• 27:20
  So, so what I'm saying is that from my like nineties, eighties, nineties,
• 27:26
  geo six R through the early water cold, you know, it's away from the oil
• 27:31
  cold ones to the early fuel injected.
• 27:34
  Water.
• 27:35
  Cool ones through to the K five, you know, and the last one that I have,
• 27:40
  which is arguably, you know, which a lot of people feel, is peak analog
• 27:43
  sports bike design, which does feel like a 600 easy to ride, like a 600.
• 27:49
  Not as intimidating as the earlier bikes, although it's just as powerful, right?
• 27:54
  That is a long explanation about how a particular design thing evolved.
• 28:01
  Right.
• 28:02
  What we're saying is that that evolutionary process can be
• 28:08
  incredibly powerful, right?
• 28:10
  It seems like it's just small changes incrementally, but when they
• 28:14
  compound, even if it's just from a 1993 jigsaw to a 2001 jigsaw, from
• 28:21
  a 2001 jigsaw to a 2005 jigsaw.
• 28:24
  The changes are enormous taken all together.
• 28:27
  They're enormous.
• 28:28
  The thing in the 2005 jigsaw is nothing like the 1993.
• 28:33
  I mean, it's still a sports bike.
• 28:34
  It still has like the GSXR genes, but oh my word, the thing has evolved an an awful
• 28:40
  lot, so this is my way of underlining.
• 28:43
  I guess the other example of continuing evolution is the Porsche nine 11.
• 28:48
  And it's had its revolutionary changes, right?
• 28:51
  But even then it's been like, oh, it's water cold now.
• 28:54
  Like, oh, it's the 9 9 1.
• 28:55
  Now I can barely look at it.
• 28:56
  It's so big and fat.
• 28:58
  Yet now you're looking at it and being like, actually it's all right.
• 29:01
  And when you look at like an LG 50 car, you're like, wow, it's the same size.
• 29:05
  There's a Volkswagen Beetle.
• 29:06
  It's so titchy.
• 29:07
  Right?
• 29:07
  You know, you, you get used to the change and the thing evolves whilst
• 29:12
  the DNA re remains the, the same.
• 29:16
  That is the evolutionary process can happen much, much more quickly with ai.
• 29:21
  Let's be really real about that, and that I think is something that I hadn't fully
• 29:28
  wrapped my head around until last night.
• 29:30
  But to answer Spence's question, AI is not a replacement for simulation.
• 29:35
  Instead, it's a way to make the simulation more effective and happen faster.
• 29:40
  Oh, this is interesting.
• 29:41
  So like all AI clarity around understanding your goals in using
• 29:48
  the AI delivers much better results.
• 29:51
  So just like we know, the prompt is key.
• 29:54
  If you ask who won the la, the Battle of Lake Berry Sea, the AI's not got a clue.
• 29:59
  If you ask it who won Gettysburg, it can give you a really good answer.
• 30:04
  Now, the example that the people sitting next to me were asking about was.
• 30:10
  If I like, forget to tell it, tell the AI that the car's got a crumple zone.
• 30:17
  Can it figure out after the crash has happened by measuring the
• 30:22
  data, can it figure out that it had a crumple zone, or do I need to
• 30:26
  tell it about the whole concept of crumple zones in the first place?
• 30:30
  And, and the answer to that was.
• 30:32
  AI cannot invent a phenomenon it hasn't been trained on.
• 30:37
  So in other words, if you don't tell it about the concept of crumple zones,
• 30:43
  it can't pull that shit out of its ass and just make up the concept of that.
• 30:47
  You have to say, well come back to what I said before is you can't ask it.
• 30:52
  Who won the Battle of Lake Berry Sea?
• 30:54
  You have to know like Berry Seed's, a good place to do power boating
• 31:00
  and Gettysburg was where there was a big important battle that determined
• 31:05
  the outcome of the Americans of War.
• 31:07
  The purity of F1 means that it is the perfect example and illustration of the
• 31:13
  application of this ar, and it really was.
• 31:16
  He kept, I would say he came back to it three or four times
• 31:19
  just because it's such a pure.
• 31:22
  Illustration of ways.
• 31:23
  Think about that.
• 31:24
  There's a race in two weeks time.
• 31:26
  You can't wait three weeks for the data to come back.
• 31:28
  You need to be modeling something now, and it literally, if you've got
• 31:33
  three ideas for what a wing might look like once upon a time, you only
• 31:37
  add time to work on one wing design.
• 31:39
  You have to like engineering intuition.
• 31:41
  Take a guess what you are gonna work on now.
• 31:43
  You can work on all five ideas that you've got.
• 31:46
  It's really, the impact is exponential.
• 31:49
  That's what you end up feeling about.
• 31:53
  It's really, really fascinating.
• 31:56
  So on that theme of aiding engineers, intuition, the AI is capable of
• 32:05
  looking at the data and then making suggestions about where to go.
• 32:11
  And what it can't do is come up with a phenomenon like a crumple zone.
• 32:15
  That he didn't know.
• 32:16
  But what it can say to you is, oh, if you wanna avoid crashes, maybe
• 32:20
  work on Thai technology rather than working on crumple zones.
• 32:25
  That's not a very good example.
• 32:26
  But you understand what I'm saying is that the notes that I made last night was,
• 32:31
  um, evolving from intuition led design evolution to data led design evolution.
• 32:38
  AI can show the best direction to go.
• 32:41
  It could prevent dead ends.
• 32:43
  And I was laughing because one of the other cars that the chap from
• 32:48
  BMW had been looking at at BMW Classic was something that he was
• 32:52
  like, I've ever seen this before.
• 32:54
  And I was like, good lord.
• 32:56
  I remember it.
• 32:57
  It was the Twin Keel Williams BMW Formula one car.
• 33:02
  I might be wrong on this.
• 33:04
  Google it, find out more.
• 33:05
  But I remember they fundamentally an idea to design a chassis in a different way.
• 33:10
  It was a huge gamble, and if it had worked, it would've given them
• 33:13
  massive competitive advantage, but it didn't, and it just put them behind.
• 33:18
  You might even be able to make a case.
• 33:20
  It was.
• 33:21
  One of the things, if not the thing that was like Williams Spiral downward from
• 33:27
  a front of grid team to a midfield team.
• 33:30
  But I dunno, my contemporary Formula one history enough to be
• 33:33
  able to, uh, to comment on that.
• 33:35
  But at least the Twin Keel car is a case in point.
• 33:39
  AI can do a bunch of work.
• 33:41
  AI could have potentially told the Williams team, guys, this is a dead end.
• 33:45
  You don't wanna do this.
• 33:46
  This is interesting.
• 33:47
  Right.
• 33:48
  He wanted to emphasize what his product does is enables your engineers to
• 33:55
  utilize their domain experience better.
• 33:58
  So he's saying, look, you guys know better than I do what prompts to put in to my ai.
• 34:05
  I'm just saying my AI is gonna help you guys get better if you use it properly.
• 34:11
  So I, but I thought that was interesting.
• 34:13
  Uh, enable your engineers to.
• 34:17
  Utilize their domain expertise better.
• 34:20
  There were a number of mentions of sort of scope of performance
• 34:25
  games and, and 10 X seemed to be the number that kept cropping up.
• 34:30
  Now, I dunno how you can kind of measure that.
• 34:33
  Maybe it's like, you know, Malcolm Gladwell's 10,000 hours before you
• 34:38
  achieve like symbiosis with the machine and what you're doing or whatever is.
• 34:43
  10,000 hour theory is, I never read the book sadly.
• 34:47
  The point is that that's, they're not suggesting that it's gonna
• 34:51
  like, make you 50% more efficient.
• 34:53
  They're suggesting really, uh, orders of magnitude and, you know,
• 34:58
  I think it's a recognition of that potential that might explain why.
• 35:02
  The tech companies are falling over themselves to have the lead there
• 35:06
  because they see the productivity gains that are are gonna be had.
• 35:10
  I mean, the rest of us just see mass unemployment, don't we?
• 35:13
  But there you go.
• 35:14
  At least when people are unemployed, they always do productive things like phishing.
• 35:18
  They don't do things like get involved in stupid revolutionary
• 35:23
  politics, but we won't go there.
• 35:25
  Uh, we talked a lot about this concept of the black box.
• 35:29
  What he said was, customers think they want a black box.
• 35:32
  You know, they just push it and their shit's more efficient.
• 35:34
  That's what they think they want.
• 35:35
  But actually when we work with them, they realize that what they want
• 35:41
  is why I described when we were talking about the M Electric motor
• 35:44
  example right at the beginning, was that what they're looking for is,
• 35:49
  yes, great, it's 15% more efficient.
• 35:52
  But I want to dig in and see what processes it did to get to be.
• 35:58
  15% more efficient.
• 36:00
  I wanna understand how it got there because I'm gonna be able
• 36:02
  to apply that to other projects.
• 36:04
  And if I, as an engineer and a team of engineers can understand that,
• 36:08
  that's gonna improve my engineering intuition next time around.
• 36:12
  So my prompts for the AI are gonna be that much better.
• 36:16
  So it's gonna be this fly.
• 36:18
  Effect, isn't it that, I mean, when you think about what that means sort of down
• 36:24
  the line for the speed and efficiency of any and every engineering project,
• 36:30
  it's kind of mind boggling, isn't it?
• 36:31
  And there were a bunch of questions that were kicking around the topic
• 36:36
  of, do you tell the model of physics?
• 36:39
  Or do you let the AI look at the data and work out the physics itself?
• 36:45
  And I think if you come back to the rumple zone example, well, I think the answer to
• 36:51
  that would be that you have to tell it.
• 36:53
  If you built a rumple zone on the car, you have to tell it about that.
• 36:56
  If you are trying to set the water speed record, you have to.
• 37:01
  Tell it that.
• 37:02
  Yeah.
• 37:03
  The craft itself can bend a little bit.
• 37:05
  It's not just about the wind and the water.
• 37:07
  The craft itself might bend a little bit.
• 37:09
  You have to tell it all the parameters, don't you?
• 37:12
  In fact, right?
• 37:14
  If you think about my example of John Duddy, when I first surfed on John
• 37:19
  Duddy, I was using Google's Gemini.
• 37:22
  What it gave me was the Irish boxer I, I then put it with
• 37:25
  Harry Roberts in 1966, and then.
• 37:30
  It was able to deliver a meaningful result.
• 37:33
  Uh, it was a question also about how to prepare and clean
• 37:37
  the data before feeding the ai.
• 37:39
  And of course, this was a general, there was the general feeling in the
• 37:43
  audience that, that was sort of the bread and butter of any AI project was
• 37:48
  how you feed and, and clean the data.
• 37:52
  Properly.
• 37:52
  Then, uh, he wrapped up with two examples.
• 37:55
  The first one we've talked a little bit about already was the, we'll talk about
• 37:58
  both of 'em a little bit already, but the first one, these Canadian architects that
• 38:02
  do highrises and they use AI modelings to prove that the new building won't create
• 38:09
  horrible wind tunnels for pedestrians.
• 38:11
  So I talked about that as well.
• 38:13
  The additional wrinkle that I forgot to mention earlier was that originally they
• 38:17
  would just model for the wind tunnel.
• 38:18
  But now they're able to add sun and shadow modeling and then they were able
• 38:23
  to blend that and some other measure they had that I didn't remember to
• 38:28
  create a total pedestrian comfort.
• 38:31
  Kind of measure.
• 38:33
  It's quite interesting, isn't it?
• 38:34
  You see the way the AI is able to not just improve the wind tunnel modeling, but
• 38:41
  also have all these additional dimensions that overall develop and deliver a much
• 38:46
  more holistic, well for out solution, but none of it's possible without an
• 38:54
  engineer giving it the right prompt.
• 38:57
  Pretty soon it's gonna be, isn't it?
• 38:58
  Let's, let's be clear.
• 38:59
  I mean, we're all engineering our way out of a job here, apart from these SP
• 39:04
  80 CH water speed record fellows, right?
• 39:08
  That's gonna be the future, isn't it?
• 39:09
  Everyone's gonna be busy doing water speed records and fixing
• 39:13
  vintage motorcycles, aren't they?
• 39:14
  They're not all gonna be involved in pointless politics.
• 39:18
  I just wanna say as well, just with this thought, right, the way the Romans
• 39:20
  dealt with this was the corn doll.
• 39:22
  And I thought of that when Andrew Yang came along and
• 39:25
  proposed his like minimum wage.
• 39:27
  On the face of it, you are like, Andrew Yang passed me the bomb, man.
• 39:31
  What kind of hippie bullshit is that?
• 39:33
  And then when you stop and think about it, it's like it is the way the Romans
• 39:37
  controlled their unruly populous.
• 39:41
  So it's the Unpowered water speed record that these SP 80 CH guys are doing.
• 39:47
  As I said, the, the record now stands at 65 knots.
• 39:50
  Their goal is 80 knots.
• 39:53
  It's a hydrofoil kind of design.
• 39:56
  If you've not had a little look at the clip or the thumbnail
• 39:58
  image here, the physics of the field is extremely complicated.
• 40:04
  Because of the multiple factors, the kites pull, the resistance of
• 40:08
  the water, the shape of the water, wind blowing across the foil and
• 40:13
  the surface of the water itself.
• 40:15
  The shape, stiffness and friction, coefficients of the foil itself.
• 40:21
  So an interesting project.
• 40:23
  And then to wrap up.
• 40:26
  A really interesting thought that it just leaves your jaw like dragging on the floor
• 40:31
  a little bit because I just hadn't thought of it in these kind of terms before.
• 40:37
  But this is a good way actually to just wrap up my whole little session here.
• 40:43
  It's that the product works mostly off graphics.
• 40:47
  Two dimensional or three dimensional.
• 40:50
  It doesn't use words and numbers so much.
• 40:53
  And in fact, when somebody asked about that, he made the point that there
• 40:59
  were many other applications that could do words and numbers better than his.
• 41:04
  Like it was a bit of a, like, wasn't it obvious that was was what we do.
• 41:08
  Like we can do these two and three dimensional objects instead
• 41:12
  of just words and numbers.
• 41:14
  The future, my word, the future is a mind boggling place, isn't it?
• 41:20
  Thank you.
• 41:20
  Drive through.
• 41:32
  This episode has been brought to you by Grand Touring Motorsports as part
• 41:35
  of our Motoring Podcast network.
• 41:37
  For more episodes like this, tune in each week for more exciting and educational
• 41:41
  content from organizations like The Exotic Car Marketplace, the Motoring
• 41:45
  Historian, break Fix, and many others.
• 41:47
  If you'd like to support Grand Touring Motor Sports and the Motoring
• 41:50
  Podcast Network, sign up for one of our many sponsorship tiers at
• 41:54
  www.patreon.com/gt Motorsports.
• 41:58
  Please note that the content, opinions and materials presented and
• 42:01
  expressed in this episode are those of its creator, and this episode has
• 42:05
  been published with their consent.
• 42:06
  If you have any inquiries about this program, please contact the creators
• 42:10
  of this episode via email or social media as mentioned in the episode.