tarrence ⛩️

@0xSero Have you tried Forge?

I still stand by Droid being the best agent harness out there. I’ve tested everything under the sun. #1 Droid #2 Pi #3 Amp #4 OpenCode #5 Codex CLI I am still working on a few reviews but performance wise this has been my experience.

Congratulations Sorters! During our first week together we minted over 100k $NUMS Based on your activity we are making a few changes 1. You will now unlock two Power Ups at stages 6 and 12 2. We are increasing the burn rate from 70% to 90% Remember, order is its own reward ✌️

Sheesh @Await_0x @numsgg Well sorted

Your sorting terminal is now online. Here is what you need to know. 🧵

Your job is simple

My application was accepted! Have you checked yours? 🔢 @numsgg nums.gg/?ref=tarrence

Your application has been accepted. ✌️

Dami Game Jam submissions always on point

huge agents meet onchain games

Slop Cannon

Dojo Game Jam VIII is two weeks out! 📅 March 6-8 💰 $15K in prizes Build with agents, Dojo, and the new Godot SDK. Here's everything you need to know 🧵

wen alice or heeeeee is on the summit. do this. revange for @tarrence

Awesome! Let us know if there is anything we can do to help. To answer your questions: 1) Yes this is for Controller accounts and wouldn't work for other account implementations 2) Yes you can use any RPC / it doesn't have a dependency on keychain once the session is signed 3) Currently it is only contract + entrypoint but would love to add calldata level policies at some point in the future too

Nice! We’re exploring whether this could compose with our @starknet agent stack (MCP call builders + ERC-8004 identity) (cc @lordOfAFew): github.com/keep-starknet-… Trying to understand compatibility... 1) Is Controller CLI Controller-account-only, or could non-Controller account contracts be supported? 2) If a session is already registered, can "execute" continue during API/keychain degradation (RPC-only path) or is Cartridge infra required at runtime? This is crutial... 3)policy depth: contract+entrypoint only, or calldata-level constraints too (recipient/amount/max spend)? If there’s overlap, happy to test a proof of concept

Your ClawdBot can now execute Starknet transactions. So can any AI agent, script, or backend service. Trade onchain markets, automate game rewards, execute deployments, update contract state. All with the same security model protecting users in Controller-powered apps.

We have this hooked up to our internal clawdbot (Joystick) and he's been grinding @LootSurvivor for income 🤖

@aaronburnett I replaced cable with starlink and it is both faster and more reliable

By the time V2mini has reached 95% of the Starlink constellation (~60% now), I expect we’ll begin to see the early stages of this phenomenon in cable internet. Not just satellite internet.

0/ We spend more waking hours in digital spaces than physical ones. In those spaces we’re guests not owners. We build identities on accounts we don’t control. We create value that flows to a few digital landlords. We have the tools to take it back. Guest thread by @tarrence

The dread beasts of Death Mountain are rarer than ever - driven into the deepest caverns by relentless hunters. If you're still chasing a complete @LootSurvivor collection, the hunt just expanded. @spaghettiont built BeastDex to track, trade, and collect what you're missing. 👇

The crazy thing about skills.sh is just how few people are using it There are literally a few thousand people on the coding agent frontier

@banteg @fiddyresearch You convinced me and codex is light years ahead but does require more effort to configure and unlock the productivity gains

@fiddyresearch fiddy we are the only people out here seeing through the claude facade

I’ve come to a point where if I see anyone singing praises of claude code, I’m just very suspicious of any claim they make. I’ve tried claude code so many times and found it simply unusable. It’s bizarre to the point where I’ve been starting to doubt my own assessment: how can it be that everyone loves this when I only see issues? It writes bad code, its bad at long running tasks, the cli ux is horrible, buggy, it introduces vulnerabilities way too easily, it will eagerly read your secrets. Sure it has taste, that I agree. But beyond that I found it genuinely unusable. I’ve given it three attempts. Maybe fourth time is the charm but I’ll just wait for 4.6.