The category has a new name. Federal agencies now say super intelligence, SI, where they used to say AI, and the industry vocabulary is drifting the same way — the people selling data centres have started calling them SI factories.
A name is not a capability. It is worth being exact about what would actually make a machine super at reading a market, because the answer is unglamorous and it has not changed.
Fluency is not the hard part
Producing a confident paragraph about dealer positioning has been easy for a while. A machine with no data connection will give you a gamma flip to two decimal places if you ask for one. It will be well-formed, plausible and invented, and it will arrive in exactly the same tone as a correct answer.
Whatever the category is called, that failure is what separates the useful from the impressive.
Three things worth demanding
Name the source. Every figure should trace to a tool result or a feed you can open yourself. Something that cannot say where a number came from is telling you what kind of number it is.
Carry the denominator. A hit rate without its sample size is decoration. Sixty per cent of what, over how many sessions, measured against which baseline.
Keep the record where it cannot be edited. A read that was wrong should still be findable after it was wrong. Anything that only shows you its good days is not measuring itself.
What the name means here
Dr. NoVo is a Super Intelligence for the financial markets. In practice that means he answers off the live dealer map rather than from memory, names what he read, and states the window a percentile was measured over. When the desks do not support a call, the honest answer is that the data is thin and which cell is empty — never a confident number standing in for a missing one.
The name changed this week. The test did not: can you check the number?