Was just reading @ted post about proof of human efforts underway.

Having built an ai platform that helps creators repurpose content, I can say without a doubt, it’s not enough to just have the verification.

Verification ≠ Value

The new @digg reboot beta will launch with multiple paths to provide proof of human. the number below is troubling and will quickly ramp left unchecked. we're working with partners here and our own zkproof technical solutions to protect privacy 🙏🚫🤖.

a common blind spot i see in references to the Em Dash conspiracy:

people assume that AI-generated posts are the result of bots and, therefore, "proof of human" can help fix it.

this is naive, imo.

there are plenty of real humans using ChatGPT to farm attention.

and they do

ted (on farcaster)

helping build farcaster | surfing @ venice, ca | eating @ gjusta / gjelina | “lucky me, lucky mud”

Was just reading @ted post about proof of human efforts underway.

Having built an ai platform that helps creators repurpose content, I can say without a doubt, it’s not enough to just have the verification.

Verification ≠ Value

https://x.com/tednotlasso/status/1922106036724584465?s=46&t=_BupcSTiBs4Anrfgi_7-dQ

CTO @beelinesai, CPO @snickerdoodle, CEO Zealous. YC/TS Alum & Mentor

Absolutely a tool. The interesting part is the output of those tools existing within or disrupting a former value chain for a new one.

Humans doing human things ✅
Bots doing bot things ✅
Bots doing human things ⁉️
Humans doing bot things ⁉️
Etc

Interesting times indeed

We don’t automate it all so there is still a human curation model left to the process.

More importantly, most of the folks who use our product are creators and have a voice of their own and audience they are protective of which I think is a great rate limiter.

If I had any specific pointers, they would be:

1. Make the ai good but not too good so there’s still room before publishing

2. Minimize the output, talking points over polished output

3. Invert media of you can - show an image to inspire text, text to inspire video, etc.

All of these are just variations on the same theme - leave a space for the human to fill.

You can either build a bot that uses an LLM to create content (and hopefully includes some regulation/error checking). Or you can have humans use AI as a tool to improve their own communication of content. 

The line is blurring. But what never changes is the reality that an LLM is just a tool.