2 Comments

Not sure I agree that the last graph supports the point you think it’s making.

From an individual firms perspective, wouldnt spending on (externally developed) AI appear as service procured in these accounts?

In which case this type of firm might indeed not require more (in fact less possibly, if cost saving) IT-outlay, but some big providers would have more of it…on aggregate a type of pattern that could look exactly like what we see in the graph?

Expand full comment

I don't know how exactly the expenditures were categorized, but you could be right that increased spending is masked as services.

I do think that even if more spending than is indicated HAS been done, the main point still remains, because during that time we don't see a big rate of increase change in productivity. This would be expected if companies had seen that AI alone were more effective than labor alone or labor with some AI.

Expand full comment