🏝 TheTechOasis, the only newsletter you need to stay ahead of the curve with AI & Crypto

Today we have the world’s outcry to press ‘pause’ on AI and revolutionizing Crypto for iPhones

🤖 AI insight of the week - World AI thinkers & leaders want to press the ‘pause’ button on GenAI… but can they?

As Italy has banned — or will ban — ChatGPT, a selected group of individuals more or less related to AI have signed an open letter that demands a complete and unwavering halt to the training of “superior to GPT-4” models… now.

Among the signees are people like Elon Musk, Yoshua Bengio, considered one of the top-3 most influential AI scientists in history, and a smattering of other prominent entrepreneurs, researchers, and thinkers.

But why are they asking for this, and more importantly, is it really possible?

Black boxes and a threat to humanity

In an open letter you can read — and sign if you wish — here, AI is described as a turning point for the human race, stating that:

“Advanced AI could represent a profound change in the history of life on Earth, and should be planned for and managed with commensurate care and resources.“

The problem, outlined by the letter, is that LLMs like GPT-4 are black boxes.

In other words, not even OpenAI knows what is really going on inside.

Thus, these models are completely unpredictable in what they will output and, even more worrying, about what new abilities they learn as they increase in size (often referred to as ‘emergent behaviors’).

This, added to the “out-of-control” race between AI labs that the letter argues the AI industry is on, creates the perfect recipe for clumsy, no-guardrail decision-making in the name of profits and margins.

For those reasons, it demands a complete pause on the development of larger models until regulators and researchers alike understand what we’re developing.

The letter goes as far as suggesting that if there’s no coordinated agreement, governments should institute moratoriums (prohibitions).

But is this wish actually feasible or a madman’s dream?

Absolute decisions require totalitarian measures

Such measures would require a globalized agreement between all countries and enforcement agencies on two possible fronts:

Banning the hardware used for AI; banning consumer and private access to GPUs, not only crippling the AI sector but also other relevant industries like gaming or computer hardware

Treat model files as illegal assets (contraband)

However, there’s a big problem.

Every decision in this world depends on incentives.

Would China lose the opportunity to get ahead in the AI version of the long-running China⚔️US confrontation?

Would Russia, immersed in an all-out war with Ukraine and a geopolitical one with all of the West, lose the opportunity to level up its AI game?

Would the US or Europe allow the aforementioned countries and respective companies to raise above them and its companies?

“An impossible agreement” Source: Author with Diffusion model

Let’s be real, it’s simply not feasible.

That being said, and as the open letter points out, we must protect the AI principles defined and agreed upon in…

Enjoyed the introduction? Subscribe or sign in to continue with this edition.