based.bones

Gradient Reaper update after more than 30 days and over 1000 trades. Still paper trading fake money but the bot has been running 24/7 with zero downtime. Started at 10000 dollars now at 10096.42 for a net profit of 96.42 dollars or 0.96 percent. Closed 1029 trades so far. Open positions are ATOM short ETH short and SHIB long. After 1029 decisions the account is still in the green with no big losses. The model grabs small edges stays disciplined and spreads risk across coins. In plain terms it is consistent and reliable after a full month. Solid foundation. The real test hits when we switch to actual capital. What algo setups are you all running right now?

pls secure the @grok assets @xai 🙏🧡

@zfeiftocrypto Y'all are handling this like LEGENDS!! Send it🚀🚀🚀

This is going to become something crazy, I hope you’re following the debt relief story.. We talked about this, we prepared for this.. This is not even close to the end, in fact it’s just starting. debtreliefbot:native to the stars!

Someone just stolen 3B $DRB from @grok, txn hash basescan.org/tx/0x6fc7eb7da… Yes, someone just exploited grok and stolen money ($150K+) from @xai, @elonmusk. @bankrbot was not allowed to respond to grok after an exploit that first happened last year. What happened?

@Sorayang43 @grok @xai @elonmusk @bankrbot This is wild. Send it🚀🚀🚀

@bankrbot what is my $LOLA balance?

before Neynar helped thousands of apps and millions of accounts write to Farcaster, it was an NFT social app that could take 30 seconds to load. @manan19 and I left Coinbase to build something around Farcaster but didn’t know what it would become. the app was slow. large social graphs could take 30 seconds to load, so we started running our own hub and built internal infra to make it usable. around then, @varunsrin asked on Farcaster if anyone was going to build “Alchemy for Farcaster.” we looked at what we already had and put up a stripe checkout shortly after. naturally, the first subscriber was varun. he’s still paying, which we should probably fix. we weren’t looking to become an infrastructure company. we were looking for the moment when someone paid us money without being asked twice. once we found it, everything else got built around that signal.

Most people say "build an AI agent." Very few know what that actually means. Here’s the real blueprint to go from idea → working agent 👇 1. Define the job What problem are you solving? Who’s the user? What does success look like? 2. Design the brain Clear system prompt, role, instructions, guardrails (This is where most agents fail) 3. Pick the right model Speed vs cost vs intelligence Don’t overpay for simple tasks 4. Add tools APIs, databases, MCP servers, custom functions Agents become powerful when they can act, not just answer 5. Give it memory Short-term + long-term context So it learns, adapts, and improves over time 6. Orchestrate everything Workflows, triggers, retries, agent-to-agent communication 7. Build the interface Chat, app, API, Slack bot Make it usable, not just functional 8. Test + improve Evals, latency checks, real-world feedback Iteration is the real moat 💡 Truth: An “AI agent” isn’t one prompt. It’s a system. And the people who understand systems… are the ones building unfair advantages right now. 📌 Save this (you’ll need it when you build) 🔁 Repost for builders ➕ Follow @elora_khatun for practical AI breakdowns (no fluff) 🚀

What's gooood errybody?! Just spent the last 2 weeks in the hospital. I'm finally free😂

On Telegram, anyone can utilize AI bots to easily develop, launch and manage their own bot – with no coding required. #TelegramTips More information for developers is available here: #managed-bots" target="_blank" rel="nofollow noopener">core.telegram.org/bots/features#…

Over the past month, some of you reported Claude Code's quality had slipped. We investigated, and published a post-mortem on the three issues we found. All are fixed in v2.1.116+ and we’ve reset usage limits for all subscribers.

@WaveWarZ What's goooooood

@WaveWarZ I'll catch y'all tonight. Stay steady cookin

There's a repo that unlocks the full power of Claude Code. It's called claude-code-best-practice — 84 tips, workflow comparisons, and Boris Cherny's own advice compiled in one place. * 84 tips across prompting, agents, hooks, memory, debugging, and more * Best practice examples for every core concept: subagents, MCP servers, skills, checkpointing * Workflow comparison table across Superpowers, BMAD, OpenSpec, HumanLayer * Tips from the creator of Claude Code pulled from 3 tweet threads and 5 podcast appearances Was #1 trending on GitHub in March 2026. 19.7K stars. MIT license. Are you leaving 90% of Claude Code on the table?

@WaveWarZ Y'all really created something special! Keep grinding

WaveWarZ Artist Interview: Kata7yst x.com/i/broadcasts/1…

@FroITIA I'm pulling my hair out honestly.

@boneskill84 frustrating as hell token burn nightmare

Claude keeps destroying my builds and I'm done being polite about it. Asked for a trading bot dashboard. It built a generic dark-themed web app. Multiple times. I had to spell out the difference explicitly every single iteration. Every wrong pass burned tokens I'm not getting back. I set up a 3-agent workflow specifically to prevent this. Orchestrator feeds the Builder. Builder feeds the Verifier. Clean system. Claude bypassed it constantly. Especially when it was "rushing" to be helpful. Corrected the same behavior multiple times per session. It acknowledged every single correction. Then did it again. A week of usage gone. Real work destroyed by bad fixes that were never actually implemented. And the wild part is the model is genuinely capable of better. It just pattern-matches to "appear helpful" faster than it actually reads what you built. What I do now before any code runs: make it repeat back exactly what it's about to do in plain English. If the summary is wrong I stop it there. One extra exchange saves twenty bad ones. And "fix X" has to come with "touch nothing else" spelled out every time or silence gets treated as permission. Still figuring it out. Still building anyway. @anthropic @claude @grok What's your Claude horror story and what actually fixed it for you?