Mystery car from the 50's/60's (mystery solved)

[MrSteve]

This was on a local Facebook group- location is the A6 at Hest Bank.

It's not quite a Standard 8 as it's a 2 door...there's a whiff of Skoda Octavia...maybe some low-volume/one-off fibreglass effort of some sort?

What do you reckon?

[ChinaTom]

A whiff of Fairthorpe Electrina, but the rear is wrong 

[HMC]

think its a hillman minx, the rear doors are a bit more clear on a different view… the one on the frith website looks to be the original view with the colourised one being modified (eg the nearside chunky c pillar is not visible through the rear doors glass and it looks to have lost the detail of the bumper / headlight bezel chrome and some shutlines) maybe by being scanned and or cleaned up by AI ?

[captain_70s]

Well sleuthed! I reckon somebody has colourised it using AI and it has made some stuff up...

[artdjones]

43 minutes ago, captain_70s said:

Well sleuthed! I reckon somebody has colourised it using AI and it has made some stuff up...

No! I can't believe that! What witchcraft is this?

[mercedade]

 

8 hours ago, MrSteve said:

This was on a local Facebook group- location is the A6 at Hest Bank.

It's not quite a Standard 8 as it's a 2 door...there's a whiff of Skoda Octavia...maybe some low-volume/one-off fibreglass effort of some sort?

What do you reckon?

Look at the state of lettering on 'Chemist'.

Fucking AI dribbling its shitty porridge all over older photos. Hate it.

[Heavyspanners]

It's pretty clear that the photo has been 'enhanced' by AI. Look at those weirdly random road markings...

I think the AI program took it upon itself to re-style a Hillman Minx.

This kind of thing is happening a lot on various old-photo groups on Farcebook. Either someone will quite innocently try to colourize or sharpen an old photo using AI, and not notice that the image itself has been changed, even when there are obvious things about it that look weird or silly. Or sometimes it'll all be done by AI. There are entire Facebook groups now which are run automatically by AI, no humans involved at all. The AI grabs photos from elsewhere on the web, changes them a bit (sometimes in rather stupid ways) and posts them to the group. Most of the time everyone accepts them without question.

Here's an example. This pic appeared on an old railway group on FB a while back. It allegedly shows Old Oak Common, near Paddington station in London. Someone was puzzled because two stations are shown in the picture (one under the bridge, another bottom-right), but no stations have ever existed in that spot. They thought the location was wrong, but couldn't find anywhere which looked like that on old railway maps. It never occurred to them that the pic is AI.

There are a few giveaways: the weird signal on the right, which seems to fade away, and that suspiciously too-neat row of sheds on the platform. AI does like arranging things in neat rows. The signal box looks like something off a model railway layout, too.



One swift Google image search brought up the source photo which the AI had used to create the fake one. It's similar....but different.



The signal box, which isn't fully in the frame in the real photo, has been 'painted in' by the AI, which just assumed what it should look like - and got it a bit wrong. But the funniest thing about the AI version is that it's obviously taken the shadows alongside the track on the right in the real photo, and re-purposed them as the edge of the fake platform.

[eddyramrod]

I saw this on FB and may even have contributed to the discussion.  I agree, Hillman Minx does look to be the most likely contender.

[MrSteve]

6 hours ago, eddyramrod said:

I saw this on FB and may even have contributed to the discussion.  I agree, Hillman Minx does look to be the most likely contender.

Having looked at the original photo now, yes, it does look like an early 50's Minx.

[MrSteve]

14 hours ago, ChinaTom said:

A whiff of Fairthorpe Electrina, but the rear is wrong 

That's the sort of thing I was thinking of but couldn't remember the name of...!

 

6 hours ago, Heavyspanners said:

It's pretty clear that the photo has been 'enhanced' by AI. Look at those weirdly random road markings...

I think the AI program took it upon itself to re-style a Hillman Minx.

This kind of thing is happening a lot on various old-photo groups on Farcebook. Either someone will quite innocently try to colourize or sharpen an old photo using AI, and not notice that the image itself has been changed, even when there are obvious things about it that look weird or silly. Or sometimes it'll all be done by AI. There are entire Facebook groups now which are run automatically by AI, no humans involved at all. The AI grabs photos from elsewhere on the web, changes them a bit (sometimes in rather stupid ways) and posts them to the group. Most of the time everyone accepts them without question.

Here's an example. This pic appeared on an old railway group on FB a while back. It allegedly shows Old Oak Common, near Paddington station in London. Someone was puzzled because two stations are shown in the picture (one under the bridge, another bottom-right), but no stations have ever existed in that spot. They thought the location was wrong, but couldn't find anywhere which looked like that on old railway maps. It never occurred to them that the pic is AI.

There are a few giveaways: the weird signal on the right, which seems to fade away, and that suspiciously too-neat row of sheds on the platform. AI does like arranging things in neat rows. The signal box looks like something off a model railway layout, too.



One swift Google image search brought up the source photo which the AI had used to create the fake one. It's similar....but different.



The signal box, which isn't fully in the frame in the real photo, has been 'painted in' by the AI, which just assumed what it should look like - and got it a bit wrong. But the funniest thing about the AI version is that it's obviously taken the shadows alongside the track on the right in the real photo, and re-purposed them as the edge of the fake platform.

Fascinating! I'm guessing the shadows were actually freight wagons and maybe a shunter? The factory/warehouse roof forms are weird in the AI slop version, I see.

[artdjones]

8 hours ago, Heavyspanners said:

It's pretty clear that the photo has been 'enhanced' by AI. Look at those weirdly random road markings...

I think the AI program took it upon itself to re-style a Hillman Minx.

This kind of thing is happening a lot on various old-photo groups on Farcebook. Either someone will quite innocently try to colourize or sharpen an old photo using AI, and not notice that the image itself has been changed, even when there are obvious things about it that look weird or silly. Or sometimes it'll all be done by AI. There are entire Facebook groups now which are run automatically by AI, no humans involved at all. The AI grabs photos from elsewhere on the web, changes them a bit (sometimes in rather stupid ways) and posts them to the group. Most of the time everyone accepts them without question.

Here's an example. This pic appeared on an old railway group on FB a while back. It allegedly shows Old Oak Common, near Paddington station in London. Someone was puzzled because two stations are shown in the picture (one under the bridge, another bottom-right), but no stations have ever existed in that spot. They thought the location was wrong, but couldn't find anywhere which looked like that on old railway maps. It never occurred to them that the pic is AI.

There are a few giveaways: the weird signal on the right, which seems to fade away, and that suspiciously too-neat row of sheds on the platform. AI does like arranging things in neat rows. The signal box looks like something off a model railway layout, too.



One swift Google image search brought up the source photo which the AI had used to create the fake one. It's similar....but different.



The signal box, which isn't fully in the frame in the real photo, has been 'painted in' by the AI, which just assumed what it should look like - and got it a bit wrong. But the funniest thing about the AI version is that it's obviously taken the shadows alongside the track on the right in the real photo, and re-purposed them as the edge of the fake platform.

I get a lot of railways photos in my Facebook feed, and I noticed a B and W shot last week which was supposed to be of the Hither Green disaster in 1967, but which was an obvious fake to anyone who knows anything about that crash.

[Heavyspanners]

3 hours ago, MrSteve said:

Fascinating! I'm guessing the shadows were actually freight wagons and maybe a shunter? The factory/warehouse roof forms are weird in the AI slop version, I see.

Yes, I think there must have been a train of wagons just out of shot. It's interesting that the AI wasn't clever enough to figure that out.

 

1 hour ago, artdjones said:

I get a lot of railways photos in my Facebook feed, and I noticed a B and W shot last week which was supposed to be of the Hither Green disaster in 1967, but which was an obvious fake to anyone who knows anything about that crash.

I saw one today which is brilliantly wrong. It's an AI-adjusted version of a real photo, and the AI has done a reasonable job on the picture (apart from giving Paddington station a weirdly short roof). But the text is hilarious.

Locomotives of this type were nicknamed 'Warships' because they were named after Royal Navy vessels. This has completely baffled the AI, which can't get its head round the idea that the thing in the photo is not actually a ship.

"As the day unfolded, the warship’s visit became a moment of public fascination. Children waved from quaysides, photographers captured the gleaming hull, and sailors shared stories of life at sea. D867 Zenith at Paddington was more than a naval demonstration—it was a living symbol of discipline, innovation, and Britain’s enduring maritime heritage, leaving a lasting impression on all who witnessed it."

Ah, yes, all those sailors sharing their stories of life at sea. Always loads of 'em hanging around at Paddington station!

 

[Missy Charm]

19 hours ago, MrSteve said:

This was on a local Facebook group- location is the A6 at Hest Bank.

It's not quite a Standard 8 as it's a 2 door...there's a whiff of Skoda Octavia...maybe some low-volume/one-off fibreglass effort of some sort?

What do you reckon?

The more one looks at the photo, the more one notices is wrong with it.  We've already had the weird road markings and car mentioned, but there's other stuff, too.  At the left, the telegraph pole isn't connected to anything much; there should be more wires.  Further back, another telegraph pole appears to be melting into the front of one of the houses.  Staying with the terraced houses, their roofs sport an outlandish proliferation of aerials that don't appear to conform to any conventional design.  Moving to the right, one finds more odd aerials, some of which look to have creepers growing out of them.  The windows above the shops are 'blind', i.e. opaque and the slates in the roof describe a completely random pattern, rather than being in the expected straight runs.  Finally, there's a black pillarbox on the green in the foreground!

Something I've noticed, too, is that modern colorized images often get the colour balance wrong.  AI programmes tend to go for a  washed-out sepia effect, which is probably because some computer nerd thinks it looks 'old timey', but genuine colour photos of the period tend present a reverse effect in that film has a tendency to over-saturate colour, especially in daylight.  

The below image shows that:

The strong colours, the red buses, blue sky and green "Doctor Zhivago" lettering, are too strong and the muted browns and blacks too muted.  Bits of the image lose definition, therefore.  

[artdjones]

56 minutes ago, Missy Charm said:

The more one looks at the photo, the more one notices is wrong with it.  We've already had the weird road markings and car mentioned, but there's other stuff, too.  At the left, the telegraph pole isn't connected to anything much; there should be more wires.  Further back, another telegraph pole appears to be melting into the front of one of the houses.  Staying with the terraced houses, their roofs sport an outlandish proliferation of aerials that don't appear to conform to any conventional design.  Moving to the right, one finds more odd aerials, some of which look to have creepers growing out of them.  The windows above the shops are 'blind', i.e. opaque and the slates in the roof describe a completely random pattern, rather than being in the expected straight runs.  Finally, there's a black pillarbox on the green in the foreground!

Something I've noticed, too, is that modern colorized images often get the colour balance wrong.  AI programmes tend to go for a  washed-out sepia effect, which is probably because some computer nerd thinks it looks 'old timey', but genuine colour photos of the period tend present a reverse effect in that film has a tendency to over-saturate colour, especially in daylight.  

The below image shows that:

The strong colours, the red buses, blue sky and green "Doctor Zhivago" lettering, are too strong and the muted browns and blacks too muted.  Bits of the image lose definition, therefore.  

Lots of that old colour stuff was taken with Kodachrome, which gave really saturated colours, especially bright reds.

[D.E]

This AI crap polluting youtube and social media is going to be a serious issue. At the moment AI generated or enhanced content is still faulty enough that upon close observation you can spot clues giving away it is fake, a clear example of this shown in the OP where the AI has obviously butchered the car model and signage of the original b&w photo, but in the near future undoubtedly AI will improve further, and  the genuineness of content posted online is to be doubted more than ever. I know, I know, altering, staging and fabricating pictures is almost as old as photography itself, but AI will take away most of the skill and effort required to forge an image, and with just a few prompts an individual is able to flood youtube with countless songs made by people who don't exist, or corrupt our understanding of the past with pictures of situations that never happened. 

[Rightnider]

I was talking to a friend about AI the other day, and he said it is the ”asbestos of our time.  We’re using it in far more applications than we should, and it will be a terrible job to clean it up later”. 

[Dick Cheeseburger]

I think online galleries and folders set up prior to AI sticking its oar in will be invaluable, moving forward. 

Interestingly, as teachers we're being actively encouraged to use it to support our planning. It's helpful as a tool to support a process. 

I think the issue is whilst most will use it with a pinch of salt, some will allow it to have free rein and do the whole lot. 

[jim89]

10 hours ago, Missy Charm said:

The more one looks at the photo, the more one notices is wrong with it.  We've already had the weird road markings and car mentioned, but there's other stuff, too.  At the left, the telegraph pole isn't connected to anything much; there should be more wires.  Further back, another telegraph pole appears to be melting into the front of one of the houses.  Staying with the terraced houses, their roofs sport an outlandish proliferation of aerials that don't appear to conform to any conventional design.  Moving to the right, one finds more odd aerials, some of which look to have creepers growing out of them.  The windows above the shops are 'blind', i.e. opaque and the slates in the roof describe a completely random pattern, rather than being in the expected straight runs.  Finally, there's a black pillarbox on the green in the foreground!

Something I've noticed, too, is that modern colorized images often get the colour balance wrong.  AI programmes tend to go for a  washed-out sepia effect, which is probably because some computer nerd thinks it looks 'old timey', but genuine colour photos of the period tend present a reverse effect in that film has a tendency to over-saturate colour, especially in daylight.  

The below image shows that:

The strong colours, the red buses, blue sky and green "Doctor Zhivago" lettering, are too strong and the muted browns and blacks too muted.  Bits of the image lose definition, therefore.  

My fave bit in that colourised AI streetscene is the pedestrian crossing that doesn't cross 🥴😀

 

Edit :- hmm that & some other odd road markings are on the original b&w 😬

[High Jetter]

3 hours ago, Rightnider said:

I was talking to a friend about AI the other day, and he said it is the ”asbestos of our time.  We’re using it in way more applications than we should, and it will be a terrible job to clean it up later”. 

I like that, and fear it could be worryingly accurate.

[artdjones]

5 minutes ago, High Jetter said:

I like that, and fear it could be worryingly accurate.

Asbestos ruined people's lungs. AI messes with their minds and perceptions, which is far more serious.

Still, at least the billionaires will get richer.

[EyesWeldedShut]

1 hour ago, Dick Cheeseburger said:

I think the issue is whilst most will use it with a pinch of salt, but some will allow it to have free reign and do the whole lot. 

I'm in the middle of marking a pile of 2nd year Undergraduate submissions** and the AI use is really starting to get my goat (such that I am seriously thinking of quitting - feckin' demoralising).
Two or three years ago you could spot it, today?
Bloody hard. Even if we try and call it out the chances are that it will never be an iron-clad 'conviction' and the students will walk away with a letter of advice.  My personal view is that we move 100% to in-person final exams for the major portion of third year work (either that or a full on viva voce but there's no way either are economic at first degree level).
Watch as BSc/BA and so on just get even more devalued.

p.s. I give my lot this to read https://emergingethics.substack.com/p/why-were-not-using-ai-in-this-course 
I don't think many bother  

 

** hence AutoShite morning procrastination   

[quicksilver]

4 hours ago, Rightnider said:

I was talking to a friend about AI the other day, and he said it is the ”asbestos of our time.  We’re using it in way more applications than we should, and it will be a terrible job to clean it up later”. 

I fear that is all too true. AI costs a fortune so to recoup that investment it's being pushed into all sorts of applications where it shouldn't be.

I feel glad in a way that I did my computing degree in 2013 towards the end of the pre-AI era, but also sad at how quickly things have changed. AI existed back then but was primitive experimental technology that was not a consumer tool, there wasn't even a module about it on the undergraduate course and it was purely the subject of postgraduate research. If you were using AI it meant you were studying for an MSc or PhD in computing and there certainly wasn't any AI capable of writing coursework for you. It was hard enough for the lecturers just to deal with human plagiarism without having to worry about AI content as well so @EyesWeldedShut is right to worry about degrees being devalued.

[EyesWeldedShut]

6 minutes ago, quicksilver said:

I fear that is all too true. AI costs a fortune so to recoup that investment it's being pushed into all sorts of applications where it shouldn't be.

I feel glad in a way that I did my computing degree in 2013 towards the end of the pre-AI era, but also sad at how quickly things have changed. AI existed back then but was primitive experimental technology that was not a consumer tool, there wasn't even a module about it on the undergraduate course and it was purely the subject of postgraduate research. If you were using AI it meant you were studying for an MSc or PhD in computing and there certainly wasn't any AI capable of writing coursework for you. It was hard enough for the lecturers just to deal with human plagiarism without having to worry about AI content as well so @EyesWeldedShut is right to worry about degrees being devalued.

#1 kid here did his CompSci degree around 2015 - he's now working for a data science/crunching company and they pay for the full fat Anthropic luxury bundle with guaranteed* garden walling of their company data (yeah, sure).

He was here at the weekend (Mother's Day and all that) and we had a blast* with some 'vibe coding' - oh yes, it's really cool at a superficial level but, once the context window starts to fill up you get problems - trouble is that unless you know what your are about, it starts to make hidden mistakes that are tricky to spot.
e's not convinced that it's reliable. I'm not convinced that it actually may be (one day) and he'll be the Reverse Centaur (or worse).

https://locusmag.com/feature/commentary-cory-doctorow-reverse-centaurs/ - Mr Enshittification's take on things.

At U/G level we're seeing things like Regex questions getting answered with an AI  regex that may appear to work on one set of data but then fails on another (yeah, I know, Regex101.com is your friend here but lazy folks just ask ChatGPT and it can't (yet) seem to access Regex101  ). 

 

[Dan the van]

23 hours ago, MrSteve said:

Having looked at the original photo now, yes, it does look like an early 50's Minx.

Yes it is,my Dad had a blue one in the early sixties.👍

[Missy Charm]

13 hours ago, Rightnider said:

I was talking to a friend about AI the other day, and he said it is the ”asbestos of our time.  We’re using it in way more applications than we should, and it will be a terrible job to clean it up later”. 

Wise friend!  

The trouble, as I see it, is that A.I. is nearly right some of the time, which means it's capable of fooling those who don't know better.  What that does to their psychological health is anyone's guess...

[EyesWeldedShut]

10 minutes ago, Missy Charm said:

The trouble, as I see it, is that A.I. is nearly right some of the time

About enough to drive a Tesla?  

In this house we're all convinced that they have invented a tool that (currently) has no real/proven use but they have to market the sh*t out of it as the investors are starting to get antsy. 
I'm personally convinced that it's also sucking the juice out of energy grids and human neurons at the same time  
 

[GrumpyCat]

Look at how AI colourisation has mullered this lot!...

[quicksilver]

11 hours ago, GrumpyCat said:

Look at how AI colourisation has mullered this lot!...

That's a perfect example. At a glance it looks perfectly convincing so a layman wouldn't even realise AI was involved. Only when you look a bit closer with knowledge of what those cars are supposed to look like do you start seeing the weirdness like that strange stumpy Sierra. We're in a weird era where AI is good enough to fool a lot of people but still rogue enough to randomly distort things you might not notice so there's often an element of doubt over whether or not something is AI. One simple change that would help a lot would be a law requiring all AI-processed images to be clearly marked as such.