I think this author like many people is overestimating the value we'll see in AI. I think we're just going to burn a decade of energy to end up with slighty better models in a select few use cases and nothing else of interest. AGI is the only chance to do anything dusrupive and we won't get there this hype cycle.
what are you smoking, even recent DeekSeek R1 being 10x cheaper is a breakthrough improvement that would upend an industry in normal times.