Degen

Farmers learn fast, but they don't have time machines.

A big part of the allo can be connected to past actions, and the qualifying actions should change often.

They must reflect the value you want to achieve. For ex, using DL3 apps. Rotate the apps chosen.

A small part of the allo can be kept predictable.

The incentives game is hard to play. It feels often like a cat & mouse (or a hedgehog and crawlies) chase:

1. The team identifies good valuable actions that people are intrinsically motivated to do, and makes them their "new" criteria for allo. For a few days this works, but...

2. Immediately, farmers start doing a lot of that in the most superficial (aka "efficient") way, just to get their hands on free money. The valuable actions of real people are more complex, but not as "efficient" in just fulfilling a simple criteria to get money. Their allo lowers fast.

3. Result: Most of the monthly airdrop allo ends up going to farmers. This repeats next month, with similar bad results. 

How to escape from this self-defeating prophecy?

Impact

Invisible Gardener. Regen practices. Degen research. Building StatelessDAO open source co-ops. Pixeloodles & Degenomemetics:  https://zora.co/@invisiblegardener

Probably just implement /impact and cut the farmers out of the picture entirely. 

Human curation is the only way I see out of it. We have to train ppl to stop engaging with low quality farmed content by promoting high quality curation!

I agree and am very interested in the mechanisms @abundance is experimenting with to maximize quality/high impact curation over farming.

It can help, but can also backfire. Farmers are humans too, they are operating from different motivations but can find ways to "curate" and also cooperate in large farmer rings.

Influencers are humans programmed by algorithms

The current trend in software design is to think that algo's controlling human behavior to maximize engagement metrics is somehow in the human's best interests, and not just another system designed to benefit the few at the expense of the many.

📷 photographer // linktr.ee/popszi // love books, golf and techno // 🎩degentleman🔵free mint: zora.co/popszi.eth

The current trend in software design is to think that algo's controlling human behavior to maximize engagement metrics is somehow in the human's best interests, and not just another system designed to benefit the few at the expense of the many.

https://warpcast.com/earth2travis/0x7012ac3f

I agree with you, there needs to be some kind of human supervision too, you can't just change numbers in an algo.