Cor, wouldn't it be useful to *predict* where pot-holes were going to be rather than waiting for them to occur?

Sort of thing a bit of machine-learning would be perfect for, wouldn't you say?

@Edent I've taken more damage from potholes personally 😅

@Edent using locally powered machine learning - a great use of that technology.

You're right about the LLM hype stuff though. The industry has proven its own worst enemy there.

@Edent I think there's a slight disconnect though, isn't there? ML can be used to *identify* potholes, but identifying isn't the same as tackling. At least part of the negative response the other day was pointing out that even when councils know about potholes they don't always get fixed, because they're generally lacking in money and road maintenance is lower down the list of priorities.

I take the point that hypothetically investing in something like this can save money in the longer term. But when there are multiple councils that have been on the verge of bankruptcy in the last couple of years, I don't know if a long-term investment is the best approach.

@Edent This has become a failure of marketing. If people using machine learning (non-generative AI, which I have little problem with in many cases) haven't learned by now that the phrase "AI" has now become irrevocably contaminated with the multiple horrors of LLMs, then they're going to have to put up with the bad reactions.

@Edent Reminds me of when I used to work in railway safety. Once upon a time we used to inspect track by a bunch of people in high-viz walking along sections eyeballing (and hitting it). Now it's done by trains that constantly travel the network, without closing it, constantly measureing all sorts of factors.

I wouldn't be surprised if there was a bunch of neural nets at the heart of some of that tech.

I do find it funny that potholes are one of the problems that people claim will be solved by whatever the current hype is. But that's because they are one of those problems that are way more complicated and difficult to tackle than the vast majority of us realise. People (I'm sure including me) have a terrible tendency to assume that things we don't have day-to-day experience of must be as simple as they appear. Most people assume that other people's jobs are easier then theirs.

@Edent I always figured the problem with potholes was simply the cost/manpower to fix them. They remain on roads long after being reported — it doesn’t seem like “knowing that they’re there” is the blocker…

@Edent But if we get rid of the pot holes (or "incidental traffic calming measures") what will councillors point at for photos in local election adverts?

@benjamineskola did you read the piece?
They *are* using it to tackle them by doing preventative maintenance.
And, as the article says, multiple councils *are* using this now.

@Edent @benjamineskola I remain skeptical of the cost/benefit of this - the savings generated by firing "highway safety inspectors" seem unlikely to pay for lots of "preventative maintenance" (ignoring the millions of miles of already decrepit roads that need fixing anyway), especially once you factor in the contract costs for the outsourced software needed to run the video analysis. IMO, it's a nice idea for a rich country, but a bad place to start for a poor country like the UK