At TED AI in Vienna, Europe fights for its place at the table; filmmakers are worried about AI; AI agents are getting more capable; Meta criticized for calling models "open source"

Can AI help Africa close the development gap? Anthropic CEO publishes lengthy essay on AI's future; US tech companies investing in UK data centers; OpenAI researches stereotyping

Oct 18, 2024

∙ Paid

I’m spending the week at TED AI Vienna, the first AI-focused TED conference held in Europe. As I listen to people debate issues related to innovation and regulation, I can’t help but think of a clip that went viral on X where French president Macron finally said the quiet part out loud: Europe has spent the last decade over-regulating and under-investing, which has put the region on a path of technological irrelevance.

But, despite these missteps, I believe there is still a path forward for Europe—one that draws on its long-standing industrial strengths. The real opportunity for Europe to lead in AI doesn’t lie in trying to outdo Silicon Valley in consumer software or Beijing in massive data projects. Instead, Europe should double down on what has always made it formidable: its ability to deliver efficient and practical engineering solutions for precision-driven industries.

Automotive, energy or healthcare, built on decades of precision and quality, represent the perfect platforms for AI innovation. Sepp Hochreiter, inventor of the LSTM and founder of AI startup NXAI, illustrated this point very clearly in a conversation at TED AI with Liquid AI CEO Ramin Hasani. While admitting that Europe lacks the AI infrastructure to turn academic research into products that generate revenue, Mr Hochreiter showed what’s possible today in terms of industrial innovation: his company is achieving speedup of 10,000x for simulations of particles in industrial systems such as chemical processes in steelmaking or fluid dynamics for automotive design that would not be possible with traditional computation. More importantly, these simulations can also be scaled up to replicate the entire properties of a real-world system, as opposed to smaller or isolated segments of a whole. Mr Hochreiter also had an interesting observation: European AI companies should not build mechanical servants that just do what they’re told (e.g. chatbots or agents of the ChatGPT variety), but specialize in developing machines that are able to extend human intelligence and help us analyze and identify patterns that are limited by our processing abilities, from the physical (computer vision-enabled robots) to the cognitive (generative AI systems with a world model).

By the end of the week, it became clear to me that we need a major mindset shift. The EU’s regulatory-first approach has clearly now become its Achilles’ heel, producing a cottage industry of well-meaning people dressed in businesswear that have spent the last few years seeking philosophical answers to AI regulation, while the rest of the world has moved at speed to develop and deploy this new technology.

The EU AI Act and GDPR, though well-intentioned, have created bureaucratic friction for EU Inc., forcing European companies to divest resources away from innovation and focus too much on compliance, therefore opening the space for North American and Chinese companies that thrive on agility and iteration to dominate the European tech landscape.

We urgently need to wake up and start building the future, not just imagine it.

And now, here are this week’s news:

❤️Computer loves

Our top news picks for the week - your essential reading from the world of AI

⚙️Computer does

AI in the wild: how artificial intelligence is used across industry, from the internet, social media, and retail to transportation, healthcare, banking, and more

🧑‍🎓Computer learns

Interesting trends and developments from various AI fields, companies and people

Keep reading with a 7-day free trial

Subscribe to Computerspeak by Alexandru Voica to keep reading this post and get 7 days of free access to the full post archives.