Paul Rowlands

Technology evangelist for the Rail Industry. Automated design systems, data integration and visualisations. #BIM, #electrifyallthings, #augmentedreality

[Paul Rowlands @RowlyPaul]

@SteveWhiteRail @GarethDennis What data does the industry have as to the root cause of late handback?

Is this information used when planning future works with similar tasks?

[Ross Hendricks @Ross__Hendricks]

This has all the familiar hallmarks of the peak of an investment bubble

Stage 1: An influx of investor capital (i.e. dumb VCs with money burning a hole in their pocket) gives rise to fundamentally unprofitable business models (AI Labs)

Stage 2: This sends false price signals throughout the economy (i.e. selling compute at negative margins), causing massive misallocation of capital (businesses and employee workflows built around artificially cheap compute) .

Stage 3 (you are here): The false price signals inevitably clash with economic reality. In a race to show less horrific cash burn ahead of their IPOs, OpenAI and Anthropic have switched to consumption-based pricing models, and now we're suddenly finding out that companies like Uber and WalMart are cutting back on consumption

Stage 4 (coming to a stonk near you): Lots of paper wealth gets vaporized

[Paul Rowlands @RowlyPaul]

Back to MicroStation VBA today updating a Pantograph Modelling tool.

It’s genuinely more rewarding working through and making code work with your own problem solving skills rather than asking AI.

[Paul Rowlands @RowlyPaul]

@neilwirving First we must admit we have a problem.

‘We’ve always done it this way and I don’t see the problem’

[MERICA MEMED @Mericamemed]

“Outdated business built a bot, A-I-A-I-O(bsolete).” 🐄🐖🐓🚜🤖

[Paul Rowlands @RowlyPaul]

@brownliberite Contractors will lose money if they accelerate. Zero incentive to speed up. The game is to find as many errors in design, challenge the client and get paid more/extend the programme due to unforeseen works.

[Paul Rowlands @RowlyPaul]

@MichaelGannotti @andrewlee07 @Microsoft @Microsoft365 For many smaller business we need stability and not be lured into a product only for the pricing structure to totally change in 12 months. It’s a risk businesses should assess before fully committing down the AI path.

[Paul Rowlands @RowlyPaul]

@andrewlee07 @MichaelGannotti @Microsoft @Microsoft365 Be cheap/minimal at first then they will change to token based pricing. The impact this is having on GitHub CoPilot is immense and driving many away. It’s just unaffordable. Also, I doubt I will get token reimbursed when the returned ‘solution’ is utter rubbish.

[Paul Rowlands @RowlyPaul]

@neilwirving As designers are informed to use ‘published’ models there was an assumption it was suitable. ProjectWise is full of models and with a lack of suitability, imports risk.

I made the suggestion of ‘engineering suitability codes’ to the project…..don’t think it went anywhere.

[Paul Rowlands @RowlyPaul]

This was the bridge where the 3D model rather than raw point cloud was used for clearance assessment.

Huge variance between the two. Immense detail doesn’t guarantee accuracy.

A suitability code for application would ensure models are using for intended purpose (looking pretty)

[The Transpennine Route Upgrade @theTRUpgrade]

It’s day 4️⃣ of TRU improvement works between #Huddersfield & #Leeds ⚒️

Team TRU have been hard at work at Dewsbury station where we've been removing the old tracks, ready for new ones to be installed, completing improvements to the station footbridge, and continued with platform extensions 🏗️

All of this work will provide customers with better facilities, while allowing longer trains with more seats to stop at the station🚆

If you’re travelling, please plan ahead via: https://t.co/2u14XDrZCL

[Paul Rowlands @RowlyPaul]

Rail eggs in the AI basket please take note.

[Walnut Ave Value @walnutavevalue]

GitHub copilot moving to usage based pricing & shows the real token costs

[Paul Rowlands @RowlyPaul]

True costs of AI are now emerging.

[Stefán Jökull Sigurðarson @stebets]

Current status after less then 24 hours of GitHub Copilot usage (Pro subscription) with the new usage-based billing. I haven't even done any really complex work yet.... I have a feeling I'll be going somewhere else pretty soon. Been using GPT 5.4 today.

[Paul Rowlands @RowlyPaul]

@hoffman_noa Kier cancelled electrification to Sheffield……House or train…it’s the same!

[Paul Rowlands @RowlyPaul]

@NoelDolphin We can keep prodding!!

[Paul Rowlands @RowlyPaul]

We didn’t even implement cost challenge V1 so why move to V2?

When assessing V1 against the data challenge knowledge graph so much is not covered/considered.

[Paul Rowlands @RowlyPaul]

Possession limits are not identified as anyone would expect but refer to a page in sectional appendix (thanks Mr Ward). Said page details ELR, feature and miles + chain.

So now having to create code to extract sectional appendix diagram data. Opened a right digital tin of worms

[Paul Rowlands @RowlyPaul]

To supplement my works planning tool I am creating a cloud database of all access/possessions from Network Rail’s EAS documents.

If all goes to plan this will allow me to geospatially import possession details and limits.

Digital twins aren’t limited to 3D models

[Paul Rowlands @RowlyPaul]

@DWPgovuk Why don’t you start with ensuring all UK funded works are undertaken within the UK?

Too much design work is sent to ‘global design centres’. These are normally design entry level jobs, perfect for this age group.

Simple line in all contracts and opportunities will appear.

[Paul Rowlands @RowlyPaul]

@NoelDolphin I wasn’t expecting such a low %…..wow.