My personal theory is there is a threshold of performance. Below the threshold the experience is bad enough it affects revenue so getting the program up to speed becomes a priority. Above the threshold only features are prioritized to drive more revenue. That's why despite computers getting orders of magnitude faster computer programs seem to run about the same speed.
I think the threshold is more about how much more rent can we seek to collect from users, and making things more performant or ergonomic doesn’t do anything to allow sales to add another 10% to the per-user subscription pricing (I assume this is a product with per-user subscriptions, even though it’s almost certainly unnecessary).
But adding yet another gateway to ChatGPT’s API…that’s a $15/mo/user add-on right there, and not just a monkey, but one of the slower intern monkeys, could write the code for such an incredibly innovative and, there’s no other word for it, powerful new feature that everyone should be thrilled to pay for at a 200-300% (or more!) markup.
So guess who gets that VP of New Bullshit Development position, and guess what kind of choices get reinforced?
