So Into Chatgpt for GNRI Guniness Bogie Van Drawing in 4mm

[Auto-Train Original]

OK, even I can see that some things are off. But is there anything to be said for using this program for something like this? Did it get anything correct?

Prompt: "based on these photos and stats can you generate a OO Gauge/4mm scale drawing with dimensions in mm"

RESULTS (2)

Edited Monday at 16:12 by Auto-Train Original

[Mol_PMB]

Try comparing what it produced to the real drawing of the real thing! 

Also, look at its unit conversions… 

 

[Flying Snail]

It did spell Guinness correctly, so that’s a point in its favour ....  or should that be a pint? 

AI is improving at a rapid rate, but it’s not there yet. I think, like a good pint of the black stuff, it needs a little longer to settle. ...

[jhb171achill]

1 hour ago, Auto-Train Original said:

OK, even I can see that some things are off. But is there anything to be said for using this program for something like this? Did it get anything correct?

Prompt: "based on these photos and stats can you generate a OO Gauge/4mm scale drawing with dimensions in mm"

RESULTS (2)

There’s virtually nothing correct about that van beyond the number of wheels! 

I'm afraid AI is absolute light years away from even the most basic accuracy in model design. 

Verdict: avoid, unfortunately! 

[Mol_PMB]

Lots of people at work sing the praises of AI. When it can do my job, I’ll retire. Looks like I’ll be working until I drop then. It can produce pretty but false pictures but it’s absolutely useless for anything requiring an accurate answer. Just do the sums on the dimensions it’s put on those sketches… 

[Broithe]

35 minutes ago, Mol_PMB said:

Lots of people at work sing the praises of AI. When it can do my job, I’ll retire. Looks like I’ll be working until I drop then. It can produce pretty but false pictures but it’s absolutely useless for anything requiring an accurate answer. Just do the sums on the dimensions it’s put on those sketches… 

Technology usually gets better, and the early stages can be a bit poor. Like automatic chokes forty years ago, but the systems generally work OK after a while and we forget how it used to be without it.

I know an 'edgy journalist' who uses AI to write his articles. Having discussed some of the subjects with him over time, it's fairly clear that he knows so little that he wouldn't be able to tell if his articles were nonsense or not. But, I suspect a lot of his audience are much the same - they just want to hear the opinion they want reinforced.

Facebook pops up a lot of stuff on aircraft for me and I noticed a nice picture of a Phantom the other day. I was about to click away, having seen enough Phantoms to last me into the grave, when I spotted something odd - the presumed occupant of the rear seat, from the way the two crew were standing, was a woman. Reading the caption, the picture was alleged to be from 1974, so this seemed a little unlikely. Looking at the picture more closely, it was full of errors, although they 'looked right', but were either very incorrect or even physically impossible.

There are likely to be problems with people falling for incorrect AI output - a lot of people seem to not fully understand what is and is not possible - I have had somebody tell me that an obviously faked 'Victorian photograph' had to be real, because it was taken before Photoshop was invented...

[Horsetan]

I've seen annual accounts from the Scottish National Party that are more trustworthy...

[Mayner]

AI can be a useful tool but has its risks and limitations

I wonder did Chatgpt gathers (scrapes) information thats published on the internet. Presumabably it experienced an "AI Hallucination" (made things up) unable to find accurate photos or drawings when asked to generate a OO/4mm drawing of the van.

Back in the day the "Transport Research Associates"  a group of IRRS members published a seemingly accurate detailed 4mm drawing of the GN 20 Ton Bogie Covered Wagon (Guinness) complete with dimensional data, construction (material & sections), brakes, running numbers and wagons converted to carry bulk grain 1938. The drawing (possibly CAD) prepared by Herbert Richards 1968.

The GNR Wagon Diagram book(now possibly in the care of the IRRS) also contains diagrams of both the bogie Guinness wagons and bulk grain conversions.

There is also the thorny question of potential copyright breach if Chatgpt lifted photos or drawings that are subject to copyright from the internet.

Perhaps the IRRS may be able to come to an agreement with Herbert or his family for his extensive photo collections and excellent wagon drawings to be made available for future generations through the Archive.  

Meanwhile in this part of the World.

An acquaintence of mine a sports photographer is experimenting with AI as a photo editing tool as opposed to using AI to creat images. 

Meanwhile the NZ Government is planning to axe 8700 public sector jobs by 2029 to save $2.9B and adapt AI tools as a replacement https://www.rnz.co.nz/news/politics/595847/replacing-public-servants-with-ai-could-come-with-hidden-costs-critics-warn

Personally I try to avoid the AI Overview on my browser like the plague as it simply regurgites original material from other sites on the web.

[Auto-Train Original]

Quote

Back in the day the "Transport Research Associates"  a group of IRRS members published a seemingly accurate detailed 4mm drawing of the GN 20 Ton Bogie Covered Wagon (Guinness) complete with dimensional data, construction (material & sections), brakes, running numbers and wagons converted to carry bulk grain 1938. The drawing (possibly CAD) prepared by Herbert Richards 1968.

Is this available? 

 

[Broithe]

To some extent, AI is just automating a style that has been going on for decades - selecting things that are 'near enough' to be used in a presentation.

One of my favourites was in a TV programme around 25 years ago, about a bombing raid on Romanian oil refineries. The bombers flew from Libya and there was an issue with one which had to divert and "landed safely in Malta".

A clip of film was used to illustrate this. The clip did at least have the correct type of aircraft in, a Liberator, but it was completely unrelated to the subject of the documentary.

What gave it away initially was the large range of mountains in the background as the plane descended onto the runway. These mountains must have taken a lot of work to remove before I lived on Malta in the late 50s/early 60s, as they certainly weren't there then.

The film clip was only a few seconds long, showing the main gear engaging the runway, enough to illustrate what "landing safely" meant, I suppose, but it was ended before the rest of the landing was shown. This was presumably because the clip was long enough, or, as the plane had no nosewheel deployed, it would have ceased to illustrate "landing safely" quite abruptly.
 

There's a huge amount of this kind of stuff around and it gets cited as evidence over and over again, as it is endlessly regurgitated - to the point where it eventually becomes the truth....

A lot of things are intended as 'infotainment', rather than real education, and it might not matter in small doses, but there's a lot of small doses around.

If you watch, listen to or read anything 'popular' about a subject that you actually know about, it's usually full of simplistic misinterpretations, often for presentational effect.

It's a large part of the whole conspiracy theory culture.

[Mayner]

I've had examples of designers and planners making assumptions and getting things seriously wrong and no none noticing in the days before A1.

Housing development Rathfarnham 1980

I worked for a company that obtained Planning Permission/building consent for a development of 11 (expensive) detatched homes on a site in Rathfarnham during the late 70s. 

Only to discover on my first day on site complete with bulldozer, that the architect had made an error in the site plan with one of the homes seriously encroaching on a neighbours land, no one noticed the error during the planning and consent stage.

I basically ended up marking up the architects drawing as an 'as built" drawing with the corrected boundary lines and a number of houses re-positioned/juggled around to fit all drawn in pencil with the incorrect lines tippexed out.

I don't know if the architect produced a revised drawing from my 'as-built' for the title documents or they simply produced dye-line prints from my as-built if so could have caused some head scratching with clients and building society/bank solicitors at loan approval stage.

Fun n games Harlequin Centre Watford 89-92

Real cases of designers, planners and contractors apparrently experiencing hallucinations on Phase 2 of the Project where the planners tried (had to use) a number of unconventional construction techniques due to the constricted nature of the site built to link the new centre with the existing Charter Place Shopping Centre and the existing M&S, NHS & C&A Department stores without disrupting their operation.

1. First stage of the job was to build semi-circular oval loop linking Beechen Grove (Watford Ring Road) with the top level of the Charter Place  car park via a temporary bailey bridge so that we could demolish the existing (straight) ramp from Beechen Grove to the car park & demolish an existing 5 storey stair and lift tower where the Harlequin Centre was intended to link up with Charter Place.

Funnily enough this phase turned out quite straightforward with few problems, even managed to drop in the Bailey Bridge as planned on a Sunday with the new ramp opening for business on Monday.

2. Second stage intended to form a basement service area with temporary service lifts for M&S turned out a bt more tricky. First job was to form a contigious piled wall with capping beam around the new basement perimiter to support the surrounding area before digging out the basement https://www.bacsol.co.uk/solution/contiguous-retaining-wall-piling/ almost forgot temporary ground anchors after digging down about 3 meters before completing the basement excavation works. Second job was a section of 'top-down" construction alongside the BHS store where we basically formed a short tunnel to link the existing Charter Place basement service area with the new centre, by casting a section of ground floor on top of two sections of reccently cast basement wall before excavating the tunnel. Third stage was to cast the basement floor and install temporary lifts to serve M&S in what was once the BHS service area. BHS agreed to take their deliveries from the High Street or a temporary covered walkway across the construction site. Again everything went reasonably to plan except the engine blowing up on the first night of a 7day 24hr operation & the Council shutting down nightwork setting the programme back by four weeks.

3. Fun b games set in in earnest at the beginning of Stage 3 planners assumed that we would serve the piling operation for the permanent ramps that would replace the Bailey Bridge from Beechen Grove but the City Council (also client) refused to allow concrete and muck-away trucks to cross the Bus Lane on Beechen Grove.

I came up with the suggestion of a second bailey bridge this time heavy duty and 12m span, client not exactly pleased with the additional cost, but it did the job and no real alternative.

Further blupers emerged as we began to build upwards forming a section of first floor slab to support a temporary ventilation system for the basement and enclosing the basement area by forming a ground floor over the basement. At this stage Phase 1 and 2 of the project were two construction sites separated by the area between Queen Road and Loates Lane occupied by the Trewin (John Lewis) department store separating the two sites. It was intended to decant Trewin's to a new store on Phase 1 once complete. It was then planned to demolish the existing store, remove Loates Lane then extend Phase 2 basement and structure to Queens Road. 

Snag was Loates Lane was supported by a temoporary contigious wall that would be removed once the basement structure was completed on the Queens Road side of the temporary wall. First bluper to emerge on the design architectural and engineering drawings was an emergency penetrated by the tempoary basement wall with a lack of clarity on phasing. It was enentually agreed we would construct the stair core (walls, starcase (all reinforced concrete) after the ground floor was constructed and temporary piled wall removed.

The more comical was that the designers and planners did not realise the HVAC contractor intended to install the extract plant for the temporary ventilation system on a section of the first floor was not scheduled to be constructed for a further 6 months, resulting in extensive 'on site' modification. A similar problem arose with structural steelwork for an extension to the BHS store arriving which litteraly did not fit, resulting in an expensive return visit to the fabricators for modification, no one from the fabricators apparently had taken site measurments or from the design team checked the shop drawings or asked the question.

Final doozie most comical of all was myself and bricklaying contractor asking resident architect what was the purpose of a series of rooms drawn without doorways in the Car Park. Eventually came back that the Client imposed a 'design freeze" on the design team alarmed at rapidly escalating costs. The Architect team ceased work on a number of Plant Rooms omitting doors and ventilation grills.

Ended up with receiving a back handed compliment from the Client's project manager at a company function, thanking me for identifying the problem and costing the Client an estimated £100k in increased costs.

I guess one thing I learned in my working life was to take nothing at face value whether on pen and paper or digital