A gun might not win when the gearbox and hub carriers grenade. Looks like his window regulators have gone, too.
1 Like
Gales of 35mph forecasted then.
Of course its no laughing matter, given the lives ended and ruined. Even more recently, hurricanes struck North Carolina; that in itself is not unusual, but the extent Helene it reached inland was. Baxter had a huge plant at North Cove. A few weeks before the hurricane, the plant had an audit, that included preparation for effects of extreme weather. At over 400 miles from the coast, they considered the usual tornados, ice storms, and felt well prepared. Hurricane Helene slammed into it, spawing 7 tornados. The US lost 70% of its saline supply for hospitals in September 24. Capacity has only just been restored. 108 died that night in North Carolina, but incalculable more probably died prematurely because of subsequent rationing of medical supplies.
AI weather forecasting is just another step in improving our preparedness, mitigating damage and loss of life.
I’m working on aligning hospital procedures for 20+ countries. I’ve got about half a million data points to figure out, to effectively crosswalk different coding systems. Yes, AI is going to help me in that task.
2 Likes
Yes, it did last week, as stated. I used the ‘forecast’ for preparedness, all ready for a day’s detailing. The ‘forecast’ wall-to-wall sunshine nice day was in fact light drizzle. Even more funny, was if one ‘rewound’ the day for prior hours, the ‘forecast’ still said it had been nice. So the ‘forecast’ couldn’t predict the weather- after it had actually happened 
So yes, AI got a long way to go 
1 Like
Another Mazda-Toyota mash up, based on a prompt that forced Chat GPT to consider styling trends of prototypes, not production cars.
First, I give ChatGPT a visual reference, a starting point for the MX-5. This is how the designers will work. When they designed the ND, they considered NA, NB, skipped NC. I tell the AI to consider the current 2025 model. There has to be a link between the two. But I don’t want ChatGPT to merely update the ND, because despite what the media says, Toyota won’t be sticking its badge on a warmed over 11 year old ND.
Starting image, me.
ChatGPT’s first crack at it.. Too conservative.
Not special enough. The door handles are naff. I merely asked to to pop on flush door handles and make it more supercar ish.
Interesting. Its sticking with sharp edges, not a soap shape.
Now tried the Toyota, as a Coupe 2+2. Gets a bit unstuck, and tried hard to make a mini Supra GR.
I tried to get it to drop the nose a bit, unsuccessfully. But I don’t dislike the result
Compare to the Supra
Ran out of free credits. Bit more refining, and a change of livery. Interesting that the Mazda has become sharp edgedut the Mazdota became more curvacious. They look like two completely different cars though. The Toyota, if it happens, will be bigger, as its a 2+2. I want to force things like the windscreen line, to see what a platform sharing car would be like.
Mazda Iconic SP went all blobby, and, dare I say it, a bit anonymous. Chat GPT seems to have picked up, a little, on the 2024 Mazda Arata, a concept SUV unveiled in Beijing, which hints at a more angular, crisp design, very different to Iconic.
2 Likes
Horrible trans Atlantic term. Instead of “cleaning the car”.
What’s a “way to go”? Perfection? A basic understand of “weather” will indicate that is impossible, due to the constant new inputs.
When AI was introduced for mammography, they needed to train it with at least 250,000 3D Tomography quality images, showing healthy tissue and diseased tissue. But 250,000 such images don’t yet exist. So the developers used AI to create training images.
The improvement in women’s health is undeniable. One of the problems before was not so much under diagnosis, but over diagnosis; people had difficulty is discerning the difference between normal dense tissue and cancerous dense tissue. The only way to know for sure is referral for a biopesy, tying up surgeons, theatre time, for what might be a false alarm.
“Artificial Intelligence” is supposed to emulate human thinking. Maybe it will never get there. But in many areas, its now doing better than the grey matter, such as deciding if patients have cancer, or what the weather will be tomorrow. Yes, it will get things wrong, but less frequently than humans. The difference is, each time it makes a mistake, even small ones, it learns from it. Humans mostly don’t. It takes a catastrophic systems failure, like failing to anticipate hurricane-like conditions in 1987, to have a rethink.
Look at cars, or specifically, panel gaps. 20 years ago, they were all over the place. 50 years ago, even worse, and often the work of a hammer on the factory floor was needed to get a good fit. Now the cheapest of cars has comparatively excellent gappage, unless subtly, our expectations have moved on.
1 Like
Yes, i was alluding to what seemed to be “best guess” forecasts in those days. Obviously things have moved on, but no matter how good the technology is, nature will always do it’s thing.
1 Like
People pay “loadsamoney” for hand built cars produced back in the day. I believe Morgan still use the “hammer on the factory floor” method. 
What, simply becuse you say it is? lol
If your understanding is that car detailing is the same as ‘cleaning the car’, I would suggest you use your favourite ‘AI’ tool to explain to you what the difference is. Who knows, you might add to this thread, entitled ‘What have your used AI for today’, by saying ‘I found out today what car detailing is’ 
1 Like
Yeah, I don’t like it as a term (it is an Americanism), and I know what it is.
1 Like
They’re not buying those cars because of the fit and finish. Superform panels were a revelation to Morgan owners.
A certain Youtuber rebuilt a crashed recent Ferrari, and found extensive use of washers as spacers on panels, which wasn’t all that impressive.
Still…..
1 Like
Nature is just a set of rules that we, the walking bags of chemicals and water, don’t yet fully understand. Just like when people didn’t really understand the world around them when Moses came off the mountain, and so they used their imagination to fill in the gaps.
1 Like
Does show AI still has some way to go.
A human hand designed the Aztek. A top stylist from the UK, Ken Greasley, was largely responsible for the Ssangyong horrors.
It took me a collective 15 minutes to “design” a NG MX-5. The AI model tried to interpret my half baked ideas. The skill is in the Prompt; the instructions directed at AI. Those who can express themselves clearly will be successful.
Its not the AI that has some way to come, but my ability to interface with it; to Prompt engineer.
I’ve been using the free version of ChatGPT for all sorts of things in the past year or so, mainly with useful results. Today it occurred to me to ask it to guess how old I am, based on our previous interactions. It worked through each one looking for age clues and adjusting its estimate accordingly and eventually came up with the range 50 - 55.
A bit on the young side (I’m 66), but an interesting exercise.
2 Likes
It might cost if they go here lol
lol crazy city. I wonder if that will include 2+ tonnes ‘eco’ EVs 
AI ‘told’ me that average person in the UK walks 15-21 miles per week, 3-4 thousand steps per day lol AI does fake news the best. Probably 15-21 steps a week 
1 Like
And those steps will be to KFC and McDonalds!
1 Like