woodenhead

I also love that CoPilot loves to reference Stack Overflow in preference to MS Learn, typically it's early example code and referencing is Stack Overflow and then somewhere down the list it refers you to the Microsoft pages.

The company I work for develops applications that use Generative AI, I see what it does and I am impressed. It's very clever and clearly it is a massive leap forwards that will make big changes in how we interact but it still makes mistakes and when a guru advises you to check everything it does carefully because it lies / hallucinates, making up it's own references at times to justify it's answer, then you know it's got some way to go. There are lots of applications for the technology and it will replace jobs in significant numbers when the truly good applications come to full fruition.

I bet the NCB shunter driver at Bold couldn't believe his luck when he came on shift and was presented with all kinds of steam traction to do his HAA marshalling with 😄

I didn't call LLM AI, I made a statement and followed it up with why an LLM is not AI. They are a sophisticated guess the next word model, that is very good at what it does, but it is not sentient.

So the Jacobite is running (or is it) https://westcoastrailways.co.uk/news/jacobite-seats-available-to-book Cash only, pay the guard on the platform if there are seats available. So if you are not in Fort William and the train happens to be in the platform and is allowing passengers to enter then you're not getting a trip. Has it actually run since Monday? Edit: it did run today, it arrived back 18 minutes early from Mallaig and went off to the coach sidings 25 minutes early

Some analysis of the service that failed? Has the train actually run since Monday?

Large Language Model - what AI is apparently. Actually it's not, it just a bloody clever guess the next word in the sequence model. They call it AI so lay people think it is, but it isn't. The LLM cannot think for itself, cannot decide for itself, it's all built on rules. But what it does do is learn, it does guess the next word well and it can draw things, it can also find a needle in a haystack. True AI is a self thinking machine that makes it's own decisions, so far we haven't been presented with any true AI, or at least we don't think we have.

Most companies are fully aware of the dangers of AI LLM using personal or sensitive data for training the models so take steps to obtain locked copies of the LLM that sit within firewalls so the main source LLM cannot train of it. Companies who want to consume AI products want to be sure their data is safe and those selling reputable AI products who know that security of data is paramount and have to invest in locked LLM models that will not be training other LLMs using someone else's data. In terms of redundancy, the scary bit is that most CEO and senior people think they can dispose of developers as the LLM will be able to do that job instead and make no bones about it, which is great for me listening to senior leaders in my company talking about this in front of me. The talk is all about enriching everyone else's work experience not redundancy, but really it comes down to doing more with less people, they just don't talk about 'automation' in the same language anymore but it's still the same. Microsoft is now promoting a product that will actually replace the developer, manna from heaven for any CEO, but the question then comes what happens when the LLM/AI develops it's own language to do stuff because it's more efficient and the few developers left cannot fathom how it does stuff - i.e. who fixes it when it breaks. There are some good uses for LLM/AI, not going to deny it, but to make out it is going to do everything and our lives are going to be so enriched is typical bluster. The one thing I learnt recently and this is from an LLM/AI guru - don't trust the output, you need to check everything, and how do you do that, you have to check that it's references are real because LLM/AI make things up to justify their findings - really it makes stuff up. We are doomed!!!

Might be helpful if your diet has been lacking in fibre and you need some assistance....

The question is, with Bachmann now offering it's 016.5 models, will you be widening the roster on this layout despite them being different parts of the UK, or will you be doing a little slate mine as your next project?

It's not criminalising, I've not mentioned anything about punishing. This to me feels more like protecting someone from danger that they might not be able to perceive.

And childhood doesn't end arbitarily at 17 years old, pushing the boundaries in a fast moving vehicle is not a good way to learn where that brick wall, lamp post or other solid obstacle is.

@kevinlms https://www.brake.org.uk/get-involved/take-action/mybrake/knowledge-centre/young-drivers https://www.rospa.com/resources/hubs/young-drivers/after-the-test/young-drivers-at-risk https://assets.publishing.service.gov.uk/media/5a80782ded915d74e33fa9b6/young-car-drivers-2013-data.pdf

So on that analogy, let young children out in traffic to get a few glancing blows from passing cars so they understand cars are dangerous and to be treated carefully? We all learn what's good and bad when we are young, we get that from our parents, our peers and educators. But the teenage brain is a fast developing vessel that unfortunately thinks doing certain risky activities is a great idea regardless of previous experiences and even more so when it is to impress friends and potential companions.

I think the issue is that males are not aware of the perceived risk because they have not developed that part of the brain, so they don't know they are being reckless. Obviously there is a spectrum of those who are sensible and drive carefully right across to those who are extremely reckless and a massive curve between the two where there will be people who unconscious of their lack of awareness. The black box clearly has an impact as it is an external monitor that they are aware of and if it believes you are being reckless will result in you losing the insurance and ability to drive. Sadly it is not possible for people to discriminate between those who will always be sensible, those who don't give a stuff and those who simply don't know, so all come under the same banner and have to accept the black box. My son recently passed his test and drives his wife's car which used to be ours. As he was mid twenties when he passed, and married he was not asked to install a black box and neither was his wife for her previous car. It might feel like a form of big brother but I believe there is sound reasoning behind the need to control the driving of cars for males under a certain age. I was one of them once, and I can vouch it's easy to stray under peer pressure or the desire to show off. Luckily I never came to grief and typically after doing something stupid I would not do the same thing again. And thank goodness I never purchased the original Suzuki Swift GTi pocket rocket I test drove or the Peugeot 205 1.9 GTi, I would have been lethal. These days I am happy to use the speed limiter on all roads, let my car guide me on efficient foot use and get there when I get there.