Stephen Downes

This article describes new study that "provides evidence that the behaviors of these large language models have 'drifted substantially'" This is not necessarily a degradation of capabilities, though many users are interpreting it as such. No matter which interpretation you prefer, as Dickson warns, "the findings are a warning about the risks of building applications on top of black-box AI systems like ChatGPT that could produce inconsistent or unpredictable results over time." It's good advice: don't build your product on someone ese's platform! We could add: make sure you understand what your software is doing.