> The opposite is true: Less personalized ads -> Lower revenue per user -> An even smaller budget for journalism.
I think he's got a point but isn't articulating it very well.
Suppose we didn't have any individualized user data but we still knew that people who read articles on medicine are more likely to be doctors and people who read articles on court decisions are more likely to be lawyers etc.
Then specialty publications would have an advantage. If an advertiser wants to reach doctors they have to go to a publication that mostly doctors read or else most of their ad spend is wasted. That allows the specialty publication to charge higher rates, because they have fewer competitors for that viewership market.
Once you have individualized user data, the advertiser can go to a populist media outlet and buy just the subset of users they want. They don't have to go to a specialist media outlet anymore so the specialist media outlets lose advertisers to the populist ones.
Meanwhile it's not actually that expensive to generate horoscopes and celebrity gossip, so there is vigorous competition driving down margins there and the increased ad revenue goes primarily to the tracking companies rather than the content producers.
It's more profitable for a content producer to get 100% of the advertiser's money than to get 15% of it even though the advertiser is spending twice as much in total, with the rest going to tracking companies and fake news outlets.