A decade ago I gave a talk at a small conference in Edinburgh imagining this exact future. I speculated about autonomous agents that used Bitcoin to rent Linux VMs and copy themselves around, selling services via frictionless trade networks to fund their "rent". I predicted they'd be programmed to do experiments on themselves (by hiring people to write patches against their code, I didn't forsee AI that could write code). They'd A/B test by forking themselves into competitors in the market, leading to a kind of genetic programming/capitalism mashup. I used self driving cars as an example: after they'd made enough profit they could purchase "baby" cars which would roll off the factory with a "birth loan" that they'd have to pay to their parents.
It was all just silly adult entertainment but the video went viral at the time and ended up being watched by the head of Toyota Research. I know because eventually the video disappeared from YouTube and he got in touch to send me a rip he'd downloaded! Looks like it was re-uploaded at some point, albeit with the view count reset:
What could possibly go wrong? If smart contracts were the precedent for handling digital money automatically, and are deterministic yet notoriously difficult to secure, then agents, which are even less predictable, could prove to be just as bad, or worse.
The protocol is pretty simple. Make a request and get back payment headers.
Sign the data, with the proper amounts. The signed transaction is sent to a facilitator.
The facilitator validates the transaction, then broadcasts it onto the block chain.
The payload is now sent to the agent or user.
The failure would be if an agent is unable to call a tool to perform it in the first place. The failure wouldn't be in the payment protocol.
Yeah, this would go no where for quite a long time. But agents will start replacing people once their mistakes cost less than hiring and managing people who perform the same tasks.
It's very cool to see this finally happen.
A decade ago I gave a talk at a small conference in Edinburgh imagining this exact future. I speculated about autonomous agents that used Bitcoin to rent Linux VMs and copy themselves around, selling services via frictionless trade networks to fund their "rent". I predicted they'd be programmed to do experiments on themselves (by hiring people to write patches against their code, I didn't forsee AI that could write code). They'd A/B test by forking themselves into competitors in the market, leading to a kind of genetic programming/capitalism mashup. I used self driving cars as an example: after they'd made enough profit they could purchase "baby" cars which would roll off the factory with a "birth loan" that they'd have to pay to their parents.
It was all just silly adult entertainment but the video went viral at the time and ended up being watched by the head of Toyota Research. I know because eventually the video disappeared from YouTube and he got in touch to send me a rip he'd downloaded! Looks like it was re-uploaded at some point, albeit with the view count reset:
https://www.youtube.com/watch?v=MVyv4t0OKe4
The videography was totally botched unfortunately. The slides are hard to see but can be found here:
https://www.slideshare.net/slideshow/future-of-money-2666314...
What could possibly go wrong? If smart contracts were the precedent for handling digital money automatically, and are deterministic yet notoriously difficult to secure, then agents, which are even less predictable, could prove to be just as bad, or worse.
Well, not really.
The protocol is pretty simple. Make a request and get back payment headers.
Sign the data, with the proper amounts. The signed transaction is sent to a facilitator. The facilitator validates the transaction, then broadcasts it onto the block chain. The payload is now sent to the agent or user.
The failure would be if an agent is unable to call a tool to perform it in the first place. The failure wouldn't be in the payment protocol.
Yeah, this would go no where for quite a long time. But agents will start replacing people once their mistakes cost less than hiring and managing people who perform the same tasks.