> Every growth function looks exponential until it's not.
What is this about?
Discovering the shape of a curve near the origin is one of the oldest and most mature estimation techniques people have. It has been done for centuries, and it's pretty trivial to differentiate exponential curves from other common shapes.
Anyway, IBM and Google have been delivering exponential growth in quantum chip size with faster (but really doesn't look exponential) decoherence each time. Any promises are irrelevant.
Are you sure you aren't looking at the bottom of an S curve? Extrapolation is a dangerous thing to engage in, particularly from the nascent forms of a technology. Many technologies hit walls early on in their lives and are discovered to be infeasible or not useful.
Also, differentiating an exponential from even a low-order polynomial is not trivial at all given only a few points (we have <10) and a lot of noise. However, the trajectory of each is markedly different when you extrapolate.
Of course it's the bottom of an S curve. Anybody that talks about exponential growth on the real world is talking about the bottom of an S curve.
What is your point? That the curve will inflect soon?
Also, about the "differentiating an exponential from even a low-order polynomial is not trivial at all given only a few points" part. Well, if you want to propose that the curve is a low-order superlinear polynomial, the onus is on you to propose a mechanism or show some analysis. That's not a very ordinary way for things to behave on practice.
What is this about?
Discovering the shape of a curve near the origin is one of the oldest and most mature estimation techniques people have. It has been done for centuries, and it's pretty trivial to differentiate exponential curves from other common shapes.
Anyway, IBM and Google have been delivering exponential growth in quantum chip size with faster (but really doesn't look exponential) decoherence each time. Any promises are irrelevant.