If you buy it after the robots, it’s very likely you’ll be working for them.
Your call.
quotingThis is a real example of a Human being being paid by an AI to perform a task in the real world for bitcoin.
note153a…p2am
I talked about this recently as Cheatcode Australia (with petermccormack (nprofile…3wlj) ) and I realised I've never shared examples.
The future is closer than you might expect.
I didn't write this task, we let an AI Agent go with a goal "I would like 100% test code coverage".
It uses our Sphinx Code Knowledge Graph to understand which functions don't have test coverage.
It uses a second custom Agent on Stakwork to review the current code implementation, example unit tests and previous bounties.
It figures out a cost and publishes a bounty on our public bounties website for any of hundreds of bounty hunters to complete.
The Bounty Hunters find it on http://community.sphinx.chat/bounties and ask to be assigned.
Currently I still have to click "assign".
In ~ 2 weeks the hunters will be able to assign themselves 👀
An hour later (or less) this will show up on my Kanban board as complete with the PR attached.
I review the PR, I pay the bounty.
A human just got paid for work created, specified and advertised entirely by a machine.
I just helped it out.
Within 2 weeks we expect the final human bottlenecks of self assignment and PR review to be automated to the point we can generate 100-1000s of unit and integration tests a day.
At this point we bring all the apps to 100% test coverage.
Sometimes I wonder if we should just bring every OpenSats (nprofile…azcs) , HRF (nprofile…445u) , and Spiral (nprofile…7m65) Open Source Project up to 100% test code coverage for the lols.
"Bitcoin, best test code coverage of any industry"
Thoguhts: NVK (nprofile…fgca) , gladstein (nprofile…uft5) , ck (nprofile…tey9) , jack (nprofile…cvv9) , moneyball (nprofile…pxq7) , ODELL (nprofile…lldz) , Gigi (nprofile…prpx)![]()