Farcaster

big question: does this get better or worse with permissionless protocols and open data?

Having access to the entire network data helps training AI to chase bots and fake news, and leveraging decentralisation can help with reinforcement learning for these.
I think it eventually gets as bad, then better.

building /anser @anser-social.eth @troopcast 
https://anser.social https://troopcast.xyz | prev. Ledger

They don’t do it because they most probably estimated that the cost is higher than the benefits for them

If access to this data is all it takes to detect/prevent bots, then what is stopping big companies from doing it? They already have access to waay more data, and even have entire teams working only on improving feed.

With user-centric distance scoring, end-user clients will be able to get rid of a lot of bots in feeds (that just defeats the interest of bots).
I also value a lot the verified addresses as a way to identify and get rid of bots.
Hopefully our friends at @bubblemaps as going to integrate Farcaster data soon.

The commoditisation of data creates an interesting opportunity for a feed market to exist

It’s just a matter of whether the opportunity is big enough for multiple serious teams to participate

If it’s not, it’s hard to see how the network doesn’t get overrun by bots without totally sacrificing discoverability