Man I never seen someone do agreegument better. We basically are in accord on most matters of substance here
I find nitpicking silly, but it is traditional in the agreegument format. I will only nitpick two things:
1. While I did put scarequotes on "innovation" for a reason (it is SV's favorite euphemism of course), it is worth noting that things like funneling money into a ridesharing app to create an entire fiefdom of "gig workers" to misclassify and then try to codify anti-labor policies into law by pretending to go to bat for them, or selling massive troves of data to a british analytics firm in order to steer elections toward "business-friendly" fascist demagogues, or creating a "non-profit" to make "safe and open" AI then partnering up with one of the most consumer-hostile and closed-source companies to ever walk the earth once a good opportunity to make bank popped up are certainly uh... creative
2. While a lot of progress in AI recently can definitely be summarized at a high level as throwing massive datasets and parallel processors at neural networks, if you zoom in there are definitely different algorithms coming through the noise, new kinds of network layers, training protocols, and objectives that we didn't have when ANNs first showed up. Like most actual real world science, it was a world of incremental progress that occasionally spawned a really useful application. I still think some advances in machine learning that have come out of all this nonsense have been meaningful and potentially useful, but it's difficult to separate signal from noise when the industry brain-drains academia to do research behind closed doors on whatever sci-fi shows up in the molochian wet dreams of the capitalist investor class