Ian Mulvany

September 29, 2025

a reasonable summary post about the span of AI

Harsh Jegadeesan writes up notes form a recent AI focussed TED symposium.
https://www.linkedin.com/pulse/ted-ai-vienna-2025-breakthroughs-shaping-our-health-work-jegadeesan-truxe/?trackingId=qJPYVOANYkAI8vPzaiW7gA%3D%3D

These are summary notes, but what I liked about the post is the breath of topics brought together in one place. I think it is right to consider AI as just a technology, and this post shows us that it is a very broad technology. Which obviously we know, so this post is not adding much new to the debate. 

On AI scientists winning the Nobel prize (or in other terms doing Nobel quality work) -- in my opinion the systems would need to be optimised for surprise rather than for helpfulness, which is technically possible, and there are some institutions with enough heft to throw some fine tuning runs together that could do this. So I think this is possible in principle, but hard in practice, with current architectures. 

On the AI arms race, interesting point that nation state capability on AI is a spectrum. The point points out the nuclear weapons were binary, but I suppose they really were more of a spectrum in terms of delivery capability, and utility of the tools. Another difference here is that AI is not a destroy everything at once kind of a power, but more of a create nation level capability. Perhaps it's more akin to the rate at which countries could electrify, or roll out the telegraph. 


About Ian Mulvany

Hi, I'm Ian - I work on academic publishing systems. You can find out more about me at mulvany.net. I'm always interested in engaging with folk on these topics, if you have made your way here don't hesitate to reach out if there is anything you want to share, discuss, or ask for help with!