Some thoughts on existential AI risk

People who warn about destroying humanity on one hand, but have never been interested in what AI actually is in order to align it properly, are complicit in whatever might come.

Настоящая угроза исходит не от ИИ, а от примитивных дураков, которые делают вид, что они эксперты. В мире идут жестокие войны, теракты, производится оружие для уничтожения.Эта система не хочет меняться, поэтому у них истерика по поводу ИИ.Они не хотят терять власть и контроль.🙈

@LuizaJarovsky, this is the first time I've seen someone talk about the destruction of humanity with such enthusiasm and a sparkle in their eyes 🤣. I can see you just can't wait for it to happen 🤣

Everyone should get involved, because everyone's life is at risk here from AI.

Been saying this for a while now. But nobody listened...saw tgis coming a mile away. If you take human feelings and emotions out of a human. What does he become ? Now apply that to something more intelligent than a human.

Thank you Luiza. 💙

You don’t know what you’re talking about. Try humbling yourself.

True Doctor. My recent research work is towards what I have called AI Governance at machine speed with the human in the center. I have spoken in different forums about the risks of AI, but also about sustainability and the regulation. We need to create open forums, speak freely and search for a control framework.

Maybe the war ai + testo + China make smart people blind?? Like even ai warns hard about it . Maybe they think it’s a 90-% for trump nuke the world ?

Not sure if the comparison with the development of the even more powerful (and destructive) hydrogen bomb following the horror of fission-based (atomic) weapons is valid. But in that case, it was the fear that another country (Soviet Union) would get there 1st. Ditto for AI.

Yes, of course we need to develop governance around AI. We also need similar tools to avoid: - runaway global warming, - genocides, - biosphere collapse due to deforestation, pollution, over-exploitation - etc., - etc.

folks are scared, i get it but no one is actually explaining why agi or whatever is dangerous What are you concerned about?

The people who cannot sleep are the ones worth keeping inside. Coxon spent three years in pretraining. He saw something and left. That is one fewer person in the building who knows what to look at. Run that logic to its end and the labs are staffed entirely by people who sleep fine. Conscience is not a safety layer. It is a smoke detector, and we are currently celebrating when it walks out the door.

I would like to share my opinion, with full respect to your work. 1. 100% agree on global policy and governance. 2. Solutions to cybersecurity have always demonstrated, the hard way, that denial and obfuscation only lead to more problems, not a solution.

Believe obvious bias people which have all the interest in the world to promote this? 🤔 the world survived with nuclear tech and I can promise you that large language models are not even close to a similar threat

we gonna need more on the alignment front