Yes, AI is a classic labour saving invention...*when* it does actually save labour.
Sometimes it's use might cost more labour in which case it is not labour saving. I think on the whole and all else being equal, AI *is* labour saving.
Why is this a good thing in the final analysis?
Because like all LSI's in the past it will boost the production of wealth - more output of wealth will be produced per unit hour of labour and capital input. A great thing.
What is greater still is given this will increase the demand for the locations in which this new wealth can be produced, real estate will also boom to that extent, perhaps a bit more due to the expectation of even higher future land prices(speculation).
We've been forecasting a Great Recession for 20 years. One is due within the next 12 months. Yield curve is signalling. Rates are signalling. Mortgage repos versus approvals are signalling. Timing is about right.
One thing the states of the world will be desperate for as they send rates toward zero and adopt huge QE to stimulate, is a means of exchange which works at much lower cost to before. And can be deployed almost immediately. (recovery always is a factor of increased productivity again versus the price of annual rental value)
My view is the AI datacentres will suddenly be converted to hash power. That is, for mining Bitcoin as ecash at very high transaction scale (maybe even without the HW acceleration).
Teranode is ready...I think.