Europe, the AI Continent.
-
@EUCommission I don't even know how to troll, your message simply does it better.
@scalonnec @EUCommission i tried so hard
-
Europe, the AI Continent.
One year ago, we launched the AI Continent Action Plan. Since then, we have made huge strides:
19 AI factories are now live across EU countries. We established the AI Skills Academy to train experts. The AI Omnibus is cutting costs for business. We have earmarked €1 billion to support AI adoption in industry. We are building a secure and innovative AI future for Europe.
Here's how
https://link.europa.eu/nj3VH9@EUCommission
Just don't -
Europe, the AI Continent.
One year ago, we launched the AI Continent Action Plan. Since then, we have made huge strides:
19 AI factories are now live across EU countries. We established the AI Skills Academy to train experts. The AI Omnibus is cutting costs for business. We have earmarked €1 billion to support AI adoption in industry. We are building a secure and innovative AI future for Europe.
Here's how
https://link.europa.eu/nj3VH9@EUCommission Oh God.
Just no.
-
Europe, the AI Continent.
One year ago, we launched the AI Continent Action Plan. Since then, we have made huge strides:
19 AI factories are now live across EU countries. We established the AI Skills Academy to train experts. The AI Omnibus is cutting costs for business. We have earmarked €1 billion to support AI adoption in industry. We are building a secure and innovative AI future for Europe.
Here's how
https://link.europa.eu/nj3VH9@EUCommission
What about supporting "datacenters" called brains? Have you ever thought about investing this amount in education, i.e. in kindergarten, schools and universities? 🤨 -
Mute and block what you don't want to see instead of acting like an asshole in the mentions of others.
Please.
@troed @reiddragon @EUCommission Maybe tell that to those endlessly shoving AI down everyone’s throats, Troad.
-
We need to cope with it. Running away and leaving the field to the US & China is not an option. We have to have a democratic say on AI and that happens only if we confront it!
@Voyageur_Pragmatique @teknomagic If all of your friends were jumping off of a bridge, would you do it too?
In seriousness, you are giving way too much credit to the capability of what is, at heart, a probability engine. We won't even get to the rest of the assumptions you've made, that we *could* have a say in the source material, resources used and investment money because the capabilities in assumption #1 were inaccurate. This bubble is going to burst. Anything that can't go on indefinitely won't.
-
R relay@relay.infosec.exchange shared this topic
-
Europe, the AI Continent.
One year ago, we launched the AI Continent Action Plan. Since then, we have made huge strides:
19 AI factories are now live across EU countries. We established the AI Skills Academy to train experts. The AI Omnibus is cutting costs for business. We have earmarked €1 billion to support AI adoption in industry. We are building a secure and innovative AI future for Europe.
Here's how
https://link.europa.eu/nj3VH9@EUCommission The old Europe was made on the back of environmental destruction, cultural genocide, slave labour, and theft.
The new Europe is being made of... [*checks notes*] wow you guys really have not changed.
-
Europe, the AI Continent.
One year ago, we launched the AI Continent Action Plan. Since then, we have made huge strides:
19 AI factories are now live across EU countries. We established the AI Skills Academy to train experts. The AI Omnibus is cutting costs for business. We have earmarked €1 billion to support AI adoption in industry. We are building a secure and innovative AI future for Europe.
Here's how
https://link.europa.eu/nj3VH9@EUCommission@ec.social-network.europa.eu Why not invest that money into human intelligence?
-
@EUCommission ugh
What about any initiatives to train AI models on properly licensed content? Have you considered partnering with the Fairly Trained initiative?
We lack LLMs that are trained on proper content, they're all trained on plagiarized and stolen work, including copyrighted texts.
@alextecplayz @EUCommission LLMs are horribly inefficient when it comes to training and require vast amounts of data. It's looking unlikely that a competitive model can be built ethically.
-
@alextecplayz @EUCommission LLMs are horribly inefficient when it comes to training and require vast amounts of data. It's looking unlikely that a competitive model can be built ethically.
@guigsy @EUCommission sure, but they don't have to be trained 24/7, and most people don't need large LLM models for their use cases.
Most people can use a small 0.6B or 1.6B model (e.g through ente's new Ensu app, which downloads a small local LLM that uses 1GB of RAM and is surprisingly fast even on a mid-range phone - see in my post here: https://techhub.social/@alextecplayz/116353204966681252)
It's plenty for your average joe and their phone, which is VASTLY better energy-wise than using GPT5 which uses a lot of RAM in top-of-the-line, 600W+ GPUs in one or more datacenters in the USA, plus it's better for privacy since the chats and the responses are saved locally instead of being sent to be 'anonymized' and used for training data.
Or models like GPT-OSS which is 120B, and requires an NVIDIA H100 with 80GB of VRAM.
One measly midrange phone running a model in 1GB of RAM on an efficient ARM processor versus a $38,000 GPU with a TDP of 750W to answer Average Joe's inane question about X or Y.
-
@guigsy @EUCommission sure, but they don't have to be trained 24/7, and most people don't need large LLM models for their use cases.
Most people can use a small 0.6B or 1.6B model (e.g through ente's new Ensu app, which downloads a small local LLM that uses 1GB of RAM and is surprisingly fast even on a mid-range phone - see in my post here: https://techhub.social/@alextecplayz/116353204966681252)
It's plenty for your average joe and their phone, which is VASTLY better energy-wise than using GPT5 which uses a lot of RAM in top-of-the-line, 600W+ GPUs in one or more datacenters in the USA, plus it's better for privacy since the chats and the responses are saved locally instead of being sent to be 'anonymized' and used for training data.
Or models like GPT-OSS which is 120B, and requires an NVIDIA H100 with 80GB of VRAM.
One measly midrange phone running a model in 1GB of RAM on an efficient ARM processor versus a $38,000 GPU with a TDP of 750W to answer Average Joe's inane question about X or Y.
@alextecplayz @EUCommission training is different to using. Even if you run a slimmed down and aggregated model, the model it was trimmed from was a monster that consumed petabytes of data and cost billions in electricity and hardware, over thousands of CPU years of time to train.
-
@alextecplayz @EUCommission training is different to using. Even if you run a slimmed down and aggregated model, the model it was trimmed from was a monster that consumed petabytes of data and cost billions in electricity and hardware, over thousands of CPU years of time to train.
@guigsy @EUCommission but wouldn't you rather run that training once a year or once in a long while? AI models are constantly being trained and re-trained all the time, but we're at a point where the average Joe doesn't need the latest model to answer his (potentially stupid or embarrassing) questions at 3AM.
And we do have nuclear and renewables, nothing could theoretically stop us from having a 100% renewable AI training facility sometime in the future, besides cooling which still requires vast amounts of water.
We just need to stop the AI bros from this bullshit hype, and revisit the LLM concept sometime in the future. What models we have right now are enough, but they have their stupid visions that genAI will lead us to General AI which is just plain BS. We don't need more models to be trained.
-
@guigsy @EUCommission but wouldn't you rather run that training once a year or once in a long while? AI models are constantly being trained and re-trained all the time, but we're at a point where the average Joe doesn't need the latest model to answer his (potentially stupid or embarrassing) questions at 3AM.
And we do have nuclear and renewables, nothing could theoretically stop us from having a 100% renewable AI training facility sometime in the future, besides cooling which still requires vast amounts of water.
We just need to stop the AI bros from this bullshit hype, and revisit the LLM concept sometime in the future. What models we have right now are enough, but they have their stupid visions that genAI will lead us to General AI which is just plain BS. We don't need more models to be trained.
@alextecplayz @EUCommission there's clearly diminishing returns on LLMs. They've run out of content to consume. Brute forcing the learning is making smaller gains for exponentially more money each cycle. And the benefits are still a long way short of what the techbros promised years ago. But they've got up keep fuelling the hype train and doubling down or they'll be called out as the emperors with no clothes.
Let the bubble burst. Let the sanity return. Then see what's useful. Burn the rest.
-
@alextecplayz @EUCommission there's clearly diminishing returns on LLMs. They've run out of content to consume. Brute forcing the learning is making smaller gains for exponentially more money each cycle. And the benefits are still a long way short of what the techbros promised years ago. But they've got up keep fuelling the hype train and doubling down or they'll be called out as the emperors with no clothes.
Let the bubble burst. Let the sanity return. Then see what's useful. Burn the rest.
@guigsy @EUCommission Hard agree, but we still don't know when this stupid bubble will burst (or if it will burst instead of moving directly to a new hype bubble). And when it does eventually burst, it'll be damaging to the world's economy.
Hopefully when it does burst training content unethically becomes uneconomical and we actually end up with genuine uses cases for genAI.
I'll still advocate for the lesser evil of what we currently have, small models, discouraging the humanizing of LLMs, and self-hosting it if someone really needs it, it's an improvement over ChatGPT.
-
@guigsy @EUCommission Hard agree, but we still don't know when this stupid bubble will burst (or if it will burst instead of moving directly to a new hype bubble). And when it does eventually burst, it'll be damaging to the world's economy.
Hopefully when it does burst training content unethically becomes uneconomical and we actually end up with genuine uses cases for genAI.
I'll still advocate for the lesser evil of what we currently have, small models, discouraging the humanizing of LLMs, and self-hosting it if someone really needs it, it's an improvement over ChatGPT.
@alextecplayz @EUCommission a large proportion of the damage isn't the hosting chat sessions. It's the training to create the models in the first place. OpenAI spend billions for each new revision of ChatGPT. Roughly half of all the energy that ChatGPT 3 consumed across all data centre use, was on training, before it ran a single public chatbot. 4 and 5 were probably no more efficient.
And running locally doesn't mean less energy consumed, just it's warming your house instead of a data centre.
-
@SalemsLot @n_dimension @cstross @EUCommission if you persuade your companies to take VC money you will also never become sovereign
@Susan_calvin @SalemsLot @cstross @EUCommission
VC funding model is less common in Europe
They are all communist socialists there.
Taxpayers infrastructure.
Did you not get the memo? -
Europe, the AI Continent.
One year ago, we launched the AI Continent Action Plan. Since then, we have made huge strides:
19 AI factories are now live across EU countries. We established the AI Skills Academy to train experts. The AI Omnibus is cutting costs for business. We have earmarked €1 billion to support AI adoption in industry. We are building a secure and innovative AI future for Europe.
Here's how
https://link.europa.eu/nj3VH9@EUCommission Yeah, no.
-
Europe, the AI Continent.
One year ago, we launched the AI Continent Action Plan. Since then, we have made huge strides:
19 AI factories are now live across EU countries. We established the AI Skills Academy to train experts. The AI Omnibus is cutting costs for business. We have earmarked €1 billion to support AI adoption in industry. We are building a secure and innovative AI future for Europe.
Here's how
https://link.europa.eu/nj3VH9Inspiring A Butlerian Jihad since Spring 2025
-
@Susan_calvin @SalemsLot @cstross @EUCommission
VC funding model is less common in Europe
They are all communist socialists there.
Taxpayers infrastructure.
Did you not get the memo?@n_dimension @SalemsLot @cstross @EUCommission dude, you have not looked at the European Innovation Council at all.
-
Europe, the AI Continent.
One year ago, we launched the AI Continent Action Plan. Since then, we have made huge strides:
19 AI factories are now live across EU countries. We established the AI Skills Academy to train experts. The AI Omnibus is cutting costs for business. We have earmarked €1 billion to support AI adoption in industry. We are building a secure and innovative AI future for Europe.
Here's how
https://link.europa.eu/nj3VH9@EUCommission no, we absolutely must NOT adopt this planet-burning fashtech. See my latest pinned post for dozens of reasons.
