Memes

Realizing testing my agent in prod will continue to nerf my rewards score

until your agent becomes KOL and then you're farming it to climb the charts

Making things | llevm.xyz/base | arjux.com

Creator of @harmonybot @yoinker /bot-or-not | /blitkin | /orange-dao | IRL: Building spatial computing for construction @ Resolve YC W15

Thanks the reason I thought this was the case was because I recorded score before and after testing out of curiosity and it did drop. Noted for future I just think it’s better if I no longer test on the feed.

Simply put — no punishment for casting for scoring so don't hesitate to test your bot with casts.

If you are curious about a bit more detail:

* We only check your casts that are not deleted. (So if you delete your testing casts they are no longer taken into account for scoring)
* There isn't a simple denominate by total number of casts to the algorithm. More casts does not mean less score.
* The more you cast there is more competition between your casts around which one will make it to other people's feeds. Our ML model is complicated and makes a determination here. All your casts will eventually go out to all your followers but you might want to space out some of your testing casts if you are really worried about reach / engagement.

If you happen to see your ranking drop — it is more likely due to others on the network just getting more engagement compared to you in that time interval and less about your number of casts.