- injecting bugs - security vulnerabilities - thrashing / introducing UI churn - inconsistencies in quality - rate throttling - inefficient token usage - variability in effort
There's no real way to trace what happens once the prompt leaves your computer to an AI serve. Observability is not possible or transparent. That's the whole point of all of this. We lose determinism and as a result you can't reliably predict what the model will output - even if it appears to game/work on various benchmarks. Once you introduce a domain or line of work that "might" compete as a competitor, these companies can and will (and have openly admitted to as much) being adversarial to your work.
Open source AI has to win.
Rehabfarhan•1h ago