4 ms·
I’ve yet to see a GenAI system that comes even slightly close. They’re dangerously inaccurate in this space.
by QuinnyPig 1y ago
I’ve yet to see a GenAI system that comes even slightly close. They’re dangerously inaccurate in this space.
- mh- 1y agoThat's been my experience as well. Makes sense, there's little training data available and the (public) docs and (first-party) tooling remain dreadful.
- laurent_du 1y agoYou probably need an ad hoc MCP for this kind of things.
- QuinnyPig 1y agoYou need far more than that. Beyond the basic resource sizing / misconfiguration issues, it’s highly contextually dependent to an environment / institution. A change that’s great for one customer could be ruinous for another.