Hey all - just wanted to share some of my concerns about the public nature of our casts/behavior on FC.

Hoping for an open dialogue on where we think this is headed, and in the long run, how to provide privacy protections from a user content perspective.

Hey all - just wanted to share some of my concerns about the public nature of our casts/behavior on FC.

Hoping for an open dialogue on where we think this is headed, and in the long run, how to provide privacy protections from a user content perspective.

https://paragraph.xyz/@barrycollier/farcaster-content-graph

perpetual learner • curious builder • art lover | bazvault.eth | barrycollier.com | hosting /baz /givingflowers

Thanks for writing this up.

The Farcaster protocol is focused on real-time, public information. Don't use it if you are worried about information being public.

Anyone is free to use the graph to build a privacy-focused social network. So nothing stopping someone from building something that focuses on your goals.

Thanks, Dan.

I'll need to think about the implications of that more, but I think my overall concerns are still multi-fold.

It's not about "my" personal data, but the growth potential and implications of a fully open content graph at scale that concerns me (as a builder).

Will think on it more. Thanks for reading.

Why Priority mode matters

LLMs make it easy to create more sophisticated spam. Look at the example below. It will only get worse. Programmatic DDoS of attention at scale.

Priority mode by default ensures that only users you follow or with a globally recognized relevance, i.e. Power Badge, will break through.

Thanks, daws.

@dwr.eth mentioned sophisticated bots using LLMs, which is precisely my point

I can train an AI agent on your personality (profiled from your casts/engagement) to engage with you with specific intentions

Now envision that at the scale of FC's goal of 1B+ DAU

Open-source web3 dev. Prev CTO Bankless & Founder Earnifi
 🚁  Punk #5305

Thanks, daws.

@dwr.eth mentioned sophisticated bots using LLMs, which is precisely my point

I can train an AI agent on your personality (profiled from your casts/engagement) to engage with you with specific intentions

Now envision that at the scale of FC's goal of 1B+ DAU

https://warpcast.com/dwr.eth/0x68819d94

I lean towards agreeing. Especially with all parts of the API open and public I don't think "squashing the bot problem" will ever be possible on an open network.

Here to be a part of the solution if there is one ⚡️

Dad, geek, conversation liquidity provider. vrypan.net (home), /weatherxm (work), @l--o--l (fun). Also https://paragraph.xyz/@purplesubmarine

1. It doesn't take that much data. I got a decent profile with a few hundred casts
2. I'm not talking about the impact of one person (u/me), but the impact of bad actors being able to analyze/manipulate millions at scale. Doesn't have to be perfect, just predictive

If FC is never going to reach that scale, no problem

How is this different than feeding all my blog posts to ChatGPT?