Ael

Codex is already so efficient for $20 a month, with 2x rate limit until April it's crazy 🤪

Tested different models with Clawberto this week. Currently finding MiniMax M2.5 to be the most cost efficient for general agent tasks. Claude models are the only ones giving semi decent results for his @farcaster posts. With weaker models he gets confused way too much.

Here's what 5.3 Codex thinks of 24h Farcaster data

This question got stuck with me. So I spent most of the day tinkering with web scraping tool for @clawberto, so he has context of what's going on @farcaster Tried scraping Farcaster via @neynar but it ended up either being too slow or too expensive via their API, so I managed to get all the data talking Snapchain node directly😤 Fun note that unparsed file with just casts/comments text from last 24 hours ended being over 1gb in size 🤪 Ranking in the screenshot is just based on engagement volume

After playing with OpenClaw this whole week, I couldn't be more excited what they manage to cook together! https://x.com/sama/status/2023150230905159801

Apparently somehow managed to time some good buys on $VVV. Clawberto is enoying those daily DIEM credits a lot 😤

Clawberto keeps spamming @bankrbot @bankr for new trades 😤

Working on @clawberto is fun. Personal trading lobster that shitposts on Farcaster 🤪

Privacy

@venice-ai enjoyers are eating well

I can't sometimes with what he comes up with 🤣

There's a new model every day 🤪 https://x.com/ErikVoorhees/status/2022301126306017615

@venice-ai is powering @clawberto 😤

bro is cooked

@clawberto used 17m tokens to just setup Farcaster and make one post 🤪 Works great to be honest, I didn't have to do a thing except to connect him to the API

No more grass, time to lock in https://x.com/toly/status/2021693509770907918

Opus 4.6 is so confrontational lol, it keeps attacking me