Tremendous thanks to @marcfaddoul.com, @chandrn.bsky.social, and @jonathanstray.bsky.social for coming on this episode to talk about your research on these questions, your vision for 'prosocial media', and your view of the leverage points for change in public policy.

should have a monopoly on deciding what content we see, when none of the content they host is theirs, in the first place?

Could social media 'middleware' take away Elon Musk's megaphone and give news organizations, writers and influencers, startups, and even libraries and museums a new and vital role in curating our feeds? Why do we take for granted that social media companies

...Related to social media, Three Pillars piece links to this blog post exploring just the kind of @nytimes.com middleware service you suggest -- with both algorithmic and editor curation options.
ucm.teleshuttle.com/2024/10/maki...

So well said! A web of community ecosystems, steered in part by communities affiliated with publications they trust, and those pubs’ staff — it sounds like a novel but very natural way to go

I am working on a podcast series on ‘middleware’ and would love to speak with anyone involved with this project or interested in it. Thank you for the incredible work that went into this paper!

Beyond the paper’s scope, but in its spirit, I am interested in exploring:

could bring their own antitrust case against platforms that give monopoly roles to their own ‘editorial’ algorithms, perhaps illegally ‘tying’ these algorithmic products to their platform products. European antitrust authorities could also play a constructive role here.

3. making ‘middleware’ happen through private antitrust action, or through government enforcement outside the U.S. The paper does a good job considering roles DOJ, FTC, and FCC might play, but it stops short of considering (e.g.) how news organizations...

2. ‘middleware’ offerings from OpenAI and Anthropic, which the paper does not consider, but which could be hits. People are coming to trust these companies’ AI more than the social platforms’ AI, simply because you can explicitly tell ChatGPT or Claude what you want more of and less of.

And conversely, if advertisers love what they see in the NYT feed, they might be happy to pay a premium for ad space there over what they are comfortable paying, under the status quo, for ad space on X.

feeds — because our feeds are no longer algorithmically personalized — they can protest (or withhold money) when they don’t like what they see. This would be a natural check against the kind of factionalization and polarization the paper contemplates.

would be safer and better than what anyone’s content algorithm can feed kids.

There is an obvious advantage to non-personalized content feeds, like the kind NYT could offer on X: accountability. If media watchdog groups and ad agencies can easily tell what NYT and Fox News are pushing into our...

or even public library system could give X’s content algorithm a run for its money.

For that matter, a team at PBS could probably hand-tailor an alternative to YouTube Kids that parents would prefer to YouTube’s own offering. Hand-selecting a few hundred hours of wholesome YouTube content...

in place of Elon Musk’s algorithmically-powered one. Seeing all the tweets that people working at NYT find interesting on a given day could be incredibly worthwhile. It would make moot the difficult issues around data processing costs and privacy that the paper explores. Any news desk, think tank...

1. ‘middleware’ offerings that are not software-based. The paper considers technology-powered curation tools a media company like The New York Times could offer, but stops short of exploring what it might look like if NYT were empowered to hand-curate (e.g.) a ‘For You’ tab on X that we could use...

I am working on a podcast series on ‘middleware’ and would love to speak with anyone involved with this project or interested in it. Thank you for the incredible work that went into this paper!

Beyond the paper’s scope, but in its spirit, I am interested in exploring: