43 karma · joined February 20, 2026
A use case may be for example give it access to your side project support email address, a test account on your site and web access.
While we used to use Excel and now use Jira, in my experience even 25 years ago never spoke to users. Spoke to their proxies in house. If lucky those proxies are from the industry and were end users once.
This misses a point that software engineers initmately know especially ones using ai tools:
* Proofs are one QA tool
* Unit tests, integration tests and browser automation are other tools.
* Your code can have bugs because it fails a test above BUT...
* You may have got the requirements wrong!
Working with claude code you can have productive loops getting it to assist you in writing tests, finding bugs you hadn't spotted and generally hardening your code.
It takes taste and dev experience definitely helps (as of Jan 26)
So I think hallucinations and proofs as the fix is a bit barking up the wrong tree
The solution to hallucinations is careful shaping of the agent environment around the project to ensure quality.
Proofs may be part of the qa toolkit for AI coded projects but probably rarely.
You can sandbox anything yourself. Use a VM.
It has a web ui.
We need a change in law but more to do with fining security breaches or requiring certification to run a site above X number of users.
No expert but I assume anything you do that is good faith usage of the site is OK. And take screenshots and report the potential problem. But making a python script to pull down data once you know? That is like getting in that car.
Real life example of fine would be you walk past a bank at midnight when it is unstaffed and the doors open so you have access to lobby (and it isnt just the night atm area). You call police on non emergency no and let them know.