Cycling UK Forum

"The Onion buys rightwing conspiracy theory site Infowars with plans to make it ‘very funny, very stupid’":
https://www.theguardian.com/media/2024/ ... alex-jones
https://theonion.com/heres-why-i-decide ... -infowars/

But think what they could have achieved if they hadn't told us. ; - )

Jonathan

Thanks for asking.

Discussion of the harm caused by flying is very welcome, and we have active threads about transport policy and climate change. They can be bumped with reminders whenever it's thought appropriate.

But not in threads where someone is asking for help. It's very likely to make others want to respond and consequently to derail the thread from its purpose. And we usually won't know what other steps the OP takes to avoid harm to the planet or how important this trip is to them.

Jonathan

Good advice x3.

Jonathan

In 16 minutes after being told which law you've read it, analysed it, and found its shortcomings?

As explained immediately above the car won't "commit driving crimes".

And the law came into force on 20 May 2024.

Jonathan

mjr wrote: ↑14 Nov 2024, 1:19pm

Jdsk wrote: ↑14 Nov 2024, 1:13pm

mjr wrote: ↑14 Nov 2024, 12:19pm ...
I'm with Not Just Bikes: there's loads of videos showing AVs breaking laws already, not all of which are accepted as bugs, and expecting AVs not to break laws is "incredibly naive".

The UK has made a decision on allocation of responsibilities. It's now law. It seems to have been welcomed, but there must be a lot of uncertainty about how it will work in practice.

What law's that, then?
...

The one cited on the first page of this thread:

Nearholmer wrote: ↑11 Nov 2024, 12:03pm Bedtime reading https://www.legislation.gov.uk/ukpga/2024/10/contents

Jonathan

mjr wrote: ↑14 Nov 2024, 12:19pm ...
I'm with Not Just Bikes: there's loads of videos showing AVs breaking laws already, not all of which are accepted as bugs, and expecting AVs not to break laws is "incredibly naive".

The UK has made a decision on allocation of responsibilities. It's now law. It seems to have been welcomed, but there must be a lot of uncertainty about how it will work in practice.

Jonathan

mjr wrote: ↑14 Nov 2024, 12:19pm

Nearholmer wrote: ↑13 Nov 2024, 10:58pm

I can't imagine anyone approaching development of these systems to only mimic human behaviour. Fortunately.

To do so would be possibly the most bizarre way to design a control system that I’ve ever heard of.

Then welcome to the bizarre state of current AI development: it's basically using lots (and I mean really lots) of human behaviour to train a model and then adding only some rules to stop the most-unacceptable output once it's discovered. https://www.technologyreview.com/2023/0 ... lt-openai/
...

That's about the development of ChatGPT. It's a Large Language Model approach. The two comments that you've quoted are about control systems in situations where immediate physical risk is obvious. That's very different.

Jonathan

"Assisted Dying: What Role for the Judge?"
https://transparencyproject.org.uk/assi ... the-judge/

Updated thoughts from Rozenberg:
https://rozenberg.substack.com/p/assist ... dium=email

Jonathan

tim-b wrote: ↑14 Nov 2024, 9:21am More league tables for the NHS.
...

Thanks for drawing attention to this. Apart from any relation to privatisation it's a terrible idea. Making as much information as possible available to the public is an excellent principle. But selecting a small amount of it, displaying it as if it's a competition, and suggesting that it will improve management or motivate staff shows serious misunderstanding.

Here's Rachel Clarke:
https://www.theguardian.com/commentisfr ... -nhs-staff

And there's many others saying the same.

Jonathan

Thanks for the update.

Jonathan

mjr wrote: ↑13 Nov 2024, 9:10pm

Jdsk wrote: ↑13 Nov 2024, 9:05pm

mjr wrote: ↑13 Nov 2024, 9:01pm ...
But if we just replace that with large reaction models trained by watching human control, then it's going to make similar mistakes, plus some new ones it didn't see in the training data, without the human ability to react to unexpected new situations, then it's not going to be an improvement.
...

I can't imagine anyone approaching development of these systems to only mimic human behaviour. Fortunately.

Not only, but most of the current systems are trained by watching human behavior, aren't they?

+ physics models, + risk models, + outcomes analysis, + differences because the sensors are different, and probably many more.

Jonathan

Pete Owens wrote: ↑13 Nov 2024, 8:35pm If our current technology of ape based vehicle control systems was a model of perfection, with road crashes being a rarity rather than the leading cause of death of young people then some of this paranoia might make sense. As it is you don't have to look far on this forum to see severe criticism of those ape based control systems. And often anger directed at them because of their inability to perform with the reliability of a computer.

Yes, it's fascinating.

Pete Owens wrote: ↑13 Nov 2024, 8:35pm 90% of crashes are caused by driver error, so the sooner we phase out human control of motor vehicles the better.

That might be the first time that anyone in this forum has gone that far. And it's a very interesting ethical point.

Jonathan

mjr wrote: ↑13 Nov 2024, 9:01pm ...
But if we just replace that with large reaction models trained by watching human control, then it's going to make similar mistakes, plus some new ones it didn't see in the training data, without the human ability to react to unexpected new situations, then it's not going to be an improvement.
...

I can't imagine anyone approaching development of these systems to only mimic human behaviour. Fortunately.

Jonathan

National Audit Office: "Department for Transport 2023-24":
https://www.nao.org.uk/overviews/depart ... t-2023-24/

Jonathan

Commons Select Committee: "Unlocking community energy at scale"
https://committees.parliament.uk/work/8 ... dium=email

The committee is now inviting evidence from individuals and organisations on any or all of the following questions:
How could the Local Power Plan to be produced by Great British Energy build upon existing community energy support schemes, such as the Community Energy Fund?
How should the energy market and licensing regulations be reformed to enable community energy projects to sell the electricity that they generate to local customers, without the current barriers, and be properly remunerated for doing so? What lessons can be learnt from other jurisdictions?
How could existing government support mechanisms, such as the Smart Export Guarantee, provide community energy projects with more financial certainty?
What are the regulatory solutions needed to minimise the high costs and long delays incurred in securing a grid connection for community energy projects?
Should the local benefits of community energy projects be formally recognised as a material consideration in planning decisions?
What should be the role of Neighbourhood Plans and Local Area Energy Plans in building local support for community energy projects?
What is the potential for community energy to incentivise consumer demand flexibility at the scale needed to achieve the UK’s net zero targets?

Jonathan