AI

[DualityNSNO]

I've been tracking AI in the workplace for around 15 years, and that period has been consistent in that tech companies have overhyped their wares, and the impact has been negligible (certainly in terms of layoffs). So I default to looking at what the evidence shows rather than what the tech companies say. It's almost inevitable that, as with every other technology, people will initially try to transpose AI onto how we currently work (and we see this with straightforward automation of tasks). This will have minimal impact. Then we'll start reorganising how we do things around what technology enables us to do. This typically takes 10-20 years, though.

Yeah I agree, and to me it feels kinda pushy currently - mostly by the big boys like MS, Amazon, who keep pushing "AI everything", while the product is still essentially in development.

It's always been a buzzword to sell your product too, our company's trying the same thing too and it's really not that applicable to us to begin with, but here we are.

 

[Bruce Wayne]

Yeah I agree, and to me it feels kinda pushy currently - mostly by the big boys like MS, Amazon, who keep pushing "AI everything", while the product is still essentially in development.

It's always been a buzzword to sell your product too, our company's trying the same thing too and it's really not that applicable to us to begin with, but here we are.

If we look at the internet era, it's predominantly been a winner-takes-all environment. I'm not sure how that'll work in AI. Google and MS can absorb the costs to an extent as they make so much from other areas of their business. OpenAI and Claude, on the other hand, will need to start recouping their costs before long.

 

[DualityNSNO]

If we look at the internet era, it's predominantly been a winner-takes-all environment. I'm not sure how that'll work in AI. Google and MS can absorb the costs to an extent as they make so much from other areas of their business. OpenAI and Claude, on the other hand, will need to start recouping their costs before long.

I think they already have begun that, albeit "silently", exactly by using MS and the big boys (plus government) - Microsoft 365 Copilot (the main one, but also all the other Copilots, as Microsoft are idiotically naming everything *productname* Copilot FFS!) is straight up ChatGPT:

While their GitHub (coding specific) Copilot is using a mix of them, currently leaning on Claude's models, and they are rumoured to have plans to include Claude in other parts of it as well, which will no doubt be reality and not just a rumour.

And AWS is using Claude entirely for their Amazon Q chat or however it was named as well for coding, with more coming soon (TM).

Then there's the government(s) box of frogs approach where models are also getting involved heavily as time goes on.

As you say, it's winner takes all, and there are other models that work quite well, but they're not famous or well funded exactly because of that, not because of something extremely different, technology wise. Mistral is also a high value company, but everyone talks of ChatGPT and Claude, so it's most likely a matter of time before they are absorbed or irrelevant.

 

[Bruce Wayne]

I think they already have begun that, albeit "silently", exactly by using MS and the big boys (plus government) - Microsoft 365 Copilot (the main one, but also all the other Copilots, as Microsoft are idiotically naming everything *productname* Copilot FFS!) is straight up ChatGPT:
View attachment 350695
While their GitHub (coding specific) Copilot is using a mix of them, currently leaning on Claude's models, and they are rumoured to have plans to include Claude in other parts of it as well, which will no doubt be reality and not just a rumour.

And AWS is using Claude entirely for their Amazon Q chat or however it was named as well for coding, with more coming soon (TM).

Then there's the government(s) box of frogs approach where models are also getting involved heavily as time goes on.

As you say, it's winner takes all, and there are other models that work quite well, but they're not famous or well funded exactly because of that, not because of something extremely different, technology wise. Mistral is also a high value company, but everyone talks of ChatGPT and Claude, so it's most likely a matter of time before they are absorbed or irrelevant.

I wouldn't be that surprised if we don't see some sovereign models/systems created in Europe. For all the talk of "digital sovereignty", when you use American platforms, you're tethered to the whims of the US government, which has obvious risks at the moment, especially if AI is used for security purposes. France is likely to be the first to break ranks, but if AI really is as transformative as the tech bros argue, then having public service versions that are not only accessible to everyone, but also developed on more ethical foundations, would seem a no brainer, especially for more left-leaning states. Sadly, the UK seems to be falling over itself to appeal/appease the tech bros, so it's unlikely to happen here, even under a Labour government.

 

[DualityNSNO]

I wouldn't be that surprised if we don't see some sovereign models/systems created in Europe. For all the talk of "digital sovereignty", when you use American platforms, you're tethered to the whims of the US government, which has obvious risks at the moment, especially if AI is used for security purposes. France is likely to be the first to break ranks, but if AI really is as transformative as the tech bros argue, then having public service versions that are not only accessible to everyone, but also developed on more ethical foundations, would seem a no brainer, especially for more left-leaning states. Sadly, the UK seems to be falling over itself to appeal/appease the tech bros, so it's unlikely to happen here, even under a Labour government.

So are many other governments, so at least the UK is not alone, I guess? The US government is basically bowing down to techbros now anyway, just look at who was around Trump during his inauguration(s), the shilling of NFTs/coins/nonsense and the money given to AI, among other things.

As I've said here, I use it and am happy with it, but obviously I can also see where and how it could be potentially horrid for security on national and personal levels. Hope the EU/European companies kick on or at least provide an actual good service that covers some of these things at the very least - some governments have rumoured/tried to implement local and locally hosted models (randomly, we tried to do something like that but it was obviously half cooked - right idea but bad knowhow, expectedly).

I recently had to explain to our "AI Team Lead" why it's horrible to just let some random AI model access all company data/repos - I'm barely okay with Copilot doing it but it's in the MS Infrastructure at least, it's already there and listening in a way. Feels like people intentionally or not skip over so many steps that should be common sense when implementing these things, no doubt the situation is the same for the model providers honestly.

 

[Cels]

​

 

[Moomin]

It's a power hungry beast.

Lot of money to be spent on Data Centres these next few years. Some interesting projects in the North East.

 

[Cels]

​

 

[Harryflashman]

Appreciate there are some fairly heavy discussions in here but this amused me

 

[Bruce Wayne]

It's a power hungry beast.

Lot of money to be spent on Data Centres these next few years. Some interesting projects in the North East.

McKinsey suggest that $7 trillion will be spent on data centres by 2030. That's a lot of money to recoup.

 

[Rob17896]

For me it’s the smart arses in work now who whatever anyone does they just pipe up ‘there’s probably an AI that can do that’

For me it is the people who think they are smart presenting AI info and it is plain as day where it came from.

 

[Bruce Wayne]

It's also interesting that Meta is currently laying off a bunch of people, not because AI is replacing them, but because they've spent so much on AI that they can't afford to keep them on.

 

[davids]

Some of the YouTube AI generated commentaries are comedy gold

 

[Milk]

It's a power hungry beast.

Lot of money to be spent on Data Centres these next few years. Some interesting projects in the North East.

McKinsey suggest that $7 trillion will be spent on data centres by 2030. That's a lot of money to recoup.

It's not the power or the cost ... it's the water usage for cooling.
If warming is the thing, where's the water coming from? Especially if government insist on putting these data centres in places like Lincolnshire which are already water depleted.

Plus, AI porn is rubbish

 

[Rob17896]

It's not the power or the cost ... it's the water usage for cooling.
If warming is the thing, where's the water coming from? Especially if government insist on putting these data centres in places like Lincolnshire which are already water depleted.

Plus, AI porn is rubbish

McKinsey are management consultants. Whenever did they get anything right?