Alan Jones

True, but at least some will be funnelled into the University which may allow them to expand their UK operations thus benefitting their home towns. Who knows, maybe they'll get more students this way

Magic Mushroom Theory captures the Greens Zack Polanski discovers the debts held by the Bank of England are not real

Maybe Paul Mason should have a chat with Zack Polanski
open.substack.com/pub/htsf/p/m...

Not terribly surprising that an AI evangelist would want to use AI in his own operation

The Guardian are reporting that this might not be the first time. Why would you publish before the event rather than after? It doesn't make sense.

That's speculation rather than hard information, isn't it?

Interesting. Can you tell me where you got those figures from?

Cheaper in terms of dollars but not necessarily cheaper in terms of energy consumption. At least one study suggests that it is not.

It does seem to be much cheaper (in terms of energy) to train, though.

I find it somewhat annoying that these things a written as if it's all the AI's idea to do these things. It's people who are instructing it! It's a tool which can be used for good or ill.

Would that be more or less factually accurate than the mainstream press?

Building data centres is govt policy, how they will be powered, ditto. It's also govt policy to invest in AI. Who else are you going to go to?

"they've been given permission to embed" is a sinister sounding way of saying that their technology is being used isn't it?

Trump donations are irrelevant.

Yep, about six fridges :)
No, Kenny is absolutely right that bundling your queries into a single question is more efficient.

What's the carbon footprint of using ChatGPT or Gemini? [August 2025 update] A new study from Google suggests its Gemini LLM uses around 0.24 Wh per text query. That's the same energy as using a microwave for one second.

Just to be a bit geeky (`cos I am), please don't get too worried about the energy use when asking AI questions. Google reckons one query is about the same as running a fridge for 6 seconds.
www.sustainabilitybynumbers.com/p/ai-footpri...

Not quite. It's built on the open source Chromium engine as are Chrome, Edge, Opera, etc.

It makes sense to build on something that is already established.

I think we know the answer:)

To be clear, does this have any material consequence? Doesn't he still hold his privileged position but will no longer be expected to perform any duties associated with that position.

Apart from damage to his pride, what is the personal consequence of this for Andrew?

The article has a sensible conclusion but it has to be said that everything you do changes your brain. And it's not surprising that your brain works harder when it's you doing work rather than an LLM. Neither is it surprising that you remember things you wrote yourself better than things you didn't

I cannot believe that he has not read Careless People

Jack Clark of Anthropic tells @lewisgoodall.com that governments need to engage with AI companies in order to shape the future use of AI.

Lewis asks thoughtful questions (without getting too much into the doom and gloom mindset of most commentators) and receives equally thoughtful answers.

He didn't say that. He made a few points points: numbers are coming down, the backlog is being reduced, we need more young people and recently it's becoming more uncomfortable in the UK if you look foreign. I may have missed something.

Agreed. But I worry for the next election when PP/Vox maybe in power

"could" is doing a lot of heavy lifting there. The MOU does not OpenAI give access to that data

This article could have been titled:

"Government rejects spending 1 billion pounds on ChatGPT".

The responses mostly assume the opposite is true and criticise the government for something it won't be doing.

OK, but it's not the government that are bothered about giving us all access the ChatGPT - it's Sam Altman trying (and failing) to flog it to them.

I'm not sure if you're criticising the Guardian or the government. The Guardian has plenty of other stories on it's Web site, including those that you are concerned about, and Peter Kyle is working to his brief (and, it seems, didn't take the OpenAI offer seriously)

The person who admitted hate speech is regarded as a hero by the right...

Noooo, why did you do that?

AI text is weird. You can't quite put your finger on what's wrong, though.

@iandunt.bsky.social is excellent on how, as AI gets more human-like and ubiquitous, humans are in danger or becoming more like machines.

iandunt.substack.com/p/machine-me...

I think they understand. Obviously there are legitimate and/or necessary uses of VPNs. This is not about that. It is trying to persuade people not to get around age verification in the home so that kids can't access unsuitable content.

No, not really. This guy is just asking people not to try and get around age verification and so make it more difficult for kids to access unsuitable content.