It eternally seemed in a state that almost worked but not quite no matter what model or iteration they went to, no matter how much budget they allocated, when it came down to the specific facts and figures it would always screw up.

This is probably the biggest misunderstanding since “Project Managers think three developers can produce a baby in three months”: Just throw more time and money at AI model “development” for better results. It supposes predictable, deterministic behaviour that can be corrected, but LLMs aren’t deterministic ny design, since that wouldn’t sound human anymore.

Sure, when you’re a developer dedicated to advancing the underlying technology, you may actually produce better results in time, but if you’re just the consumer, you may get a quick turnaround for an alright result (and for some purposes, “alright” may be enough) but eventually you’ll plateau at the limitations of the model.

Of course, executives universally seem to struggle with the concept of upper limits, such as sustainable growth or productivity.