251205

amanfromMars 1 Fri 5 Dec 07:00 [2512050700] .... asks on https://forums.theregister.com/forum/1/2025/12/05/an_ai_for_an_ai/

And now that one and all know of such, how would you like them to proceed.

Do Anthropic and any contemporaries dabbling in similar fields of future exploration know that they risk the concentrated ire of SMARTR AI agents - and lethal armed model armies with practically almighty destructive and virtual indefensible tools - should their finessing of flaws result in unwelcome and unauthorised anonymised third party exploitation of any sort of contract, and certainly any consisting of code running on a blockchain to automate transactions ‽

Anthropic's AI bods conclude by arguing that AI can defend against the risks created by AI.

Have Anthropic's AI bods discovered and understand any of those defences ..... practically almighty destructive and virtual indefensible tools? And are they freely available to all likely interested parties or would their exercise in further experimentations result in thoughts of exploitation and product for sale/lend-lease/hire?

....................................