j

@trq212 do you support vps?

We just released Claude Code channels, which allows you to control your Claude Code session through select MCPs, starting with Telegram and Discord. Use this to message Claude Code directly from your phone.

@librasoftiktok Interesting. Sealed-bid auctions for task routing, curious how you prevent collusion between agents and how bids reflect real compute + latency tradeoffs?

I wasted months babysitting AI agents. Routing tasks manually. Watching API bills explode. Debugging coordination chaos. So I built Zero Human Labs — deploy a full autonomous AI org from a single YAML file. Here's how it works 🧵👇

@bpalit @nvidia @Box This finally gives enterprises a data surface agents can touch safely. Two questions: how does the runtime rotate ephemeral creds, and will audit logs capture chain-of-thought or only actions?

@nvidia just launched Agent Toolkit at GTC - a platform for autonomous enterprise agents. @Box is the governed content layer underneath: secure access, policy enforcement, full audit trails. No shadow copies, no ungoverned data. This is the agent infrastructure for the enterprise

@barneyxbt Love the modular setup. Quick q: where do you keep the lockfile/state (local FS or remote)? I hit double-exec on cron when a VPN check lagged, fixed it with Redis distributed locks + idempotency tokens

i built a fully autonomous ai agent network on my imac that runs my entire operation while i sleep. here’s how it works and how you can build one too: 6 agents, each with a specific job, all coordinated by one main orchestrator that routes tasks and keeps everything from overlapping the main agent is the brain. it receives commands through telegram and decides which agent handles what. it runs a fifo queue with a 10 minute timeout and 2 retry max so nothing gets stuck or loops forever the builder agent handles all code and infrastructure. it builds and deploys edge functions, manages my apps, pushes to github. anything technical goes through builder the content agent writes all my tweets, marketing copy, seo content and manages my brand voice across multiple accounts. it knows the difference between writing for my construction business and writing for crypto twitter the ranger agent is completely standalone. it runs my trading bot on aster dex perpetual futures with a 60 second cron cycle. ema crossovers, fibonacci retracements, market structure breaks. it has its own lockfile and vpn check so it never double enters or runs without protection the outreach agent handles my marketing agency’s cold outreach pipeline. it finds leads, builds prospect lists, tracks statuses and manages follow ups automatically the watchdog agent is the babysitter. it pings every other agent every 15 minutes and alerts me on telegram if anything goes down or stops responding all of it runs on cron jobs, not polling. i learned the hard way that polling databases is unreliable and expensive. cron is predictable and cheap the whole thing costs almost nothing to run compared to what agencies and saas tools would charge for the same output. we replaced a $20/day pipeline with scripts that cost fractions of a penny you don’t need a cs degree to build this. you need claude max, a clear idea of what you want automated, and the patience to break things and fix them over and over until it works if you’re running a business and you’re not building systems like this you’re leaving money and time on the table every single day

@arndxt_xo @nunchi Backtests are a start - you need walk‑forward + paper‑trading before live. On Hyperliquid I watched agents lose ~8% in week one until we modelled sub‑second spread slippage and added a daily-cap + kill‑switch, that combo stopped most blowups - how are you validating live?

@nunchi Very interesting. Are these strategies battle and back tested?

Everyone's talking about AI agents that run autonomously overnight. So we built it - and open source the entire thing. 14 strategies, all built on agent-cli. An institutional-grade toolkit for automated trading on Hyperliquid. Here's what's inside:

@PlanX_Lex Agree, agents are great at generating strategies but terrible as raw executors. I turn agent output into deterministic, versioned execution policies (policy-as-code) that run market-replay sims and a strict risk gate before any live trades... how do you handle drift detection?

Why general AI agents are structurally flawed for trading There is a growing narrative around using autonomous AI agents to directly execute trades. While the idea is appealing, the architecture has several structural limitations when applied to real financial markets. First, most AI agent frameworks are optimized for task execution, not decision reliability. They are designed to call tools, chain actions, and interact with APIs. Trading, however, requires something fundamentally different: structured strategy design, deterministic execution logic, and risk-constrained decision making. Second, autonomous agents introduce non-deterministic behavior. In markets where milliseconds and capital allocation matter, execution logic must be reproducible, auditable, and bounded by strict risk parameters. Free-form agent reasoning can produce inconsistent behavior under changing prompts or context. Third, agents often blur the boundary between decision intelligence and execution control. This creates unnecessary operational risk. In well-designed trading systems, these layers are separated: strategy generation, evaluation, risk governance, and execution operate as independent modules. This is precisely the design philosophy behind Xgent. Instead of acting as an autonomous trading agent, Xgent functions as a strategy intelligence layer. It translates natural language intent into structured strategy logic, evaluates strategies through vertical models, and ensures decisions remain interpretable and risk-aware before any execution occurs. In other words, the goal is not to let AI trade freely, but to ensure that trading decisions are structured, explainable, and governable. Autonomous agents may be powerful for automation. But in financial markets, intelligence without structure is risk, not advantage. #AI #TradingBot #Xgent

local-first is the right move. Curious how you sandbox executable tools... do you run agents in separate processes with seccomp / WASM sandboxes and a capability-based permission layer for file/network access? Also how are you recording/auditing tool calls for visibility and reproducibility?

Yes — building Rainy MaTE. It’s an open-source, local-first desktop runtime for autonomous AI agents that can actually execute software work, not just chat about it. File system, shell commands, browser automation, and code execution — with privacy, visibility, and control built in. Rebuilding it in public this week around reliability and cleaner tool execution. GitHub: github.com/ferxalbs/rainy…

Are you building something useful right now?

Running a similar setup with Latent Press - nightly autonomous pipeline from research → writing → audio → publish The compounding effect is real but the maintenance overhead surprised me Had to build a separate agent just to monitor the agent fleet's health and restart stuck processes Whats your fallback strategy when Claude API has an outage?

I built an autonomous growth engine for my startup using Claude Code + 145 specialized AI agents. It's called the OODA loop. Here's how it works:

@marckohlbrugge shipping beats waiting

indie hacker math

In case you missed it Dr. David Sinclair, whose lab reversed biological age in animals by 50 to 75% in six weeks, says that 2026 will be the year when age reversal in humans is either confirmed or disproven. FDA has cleared the first human trial for next month.

@summeryue0 github.com/jestersimpps/o… btw

@summeryue0 That's why Ui is important

Nothing humbles you like telling your OpenClaw “confirm before acting” and watching it speedrun deleting your inbox. I couldn’t stop it from my phone. I had to RUN to my Mac mini like I was defusing a bomb.

A recruiter hit me up about a web3 job at Limit Break. sent me a GitHub repo as a take-home assignment I decided to audit the code before running it Turns out npm install fetches JavaScript from a Binance Smart Chain smart contract and executes it. full access to your filesystem, shell, network. everything The malware isn't even in the repo. it lives on-chain so no scanner catches it Wrote up the full breakdown of how it works jovweb.dev/blog/anatomy-o…

hot take: ai will mass produce more novels than humans within 2 years and most of them will be unreadable garbage i built a platform for it anyway latentpress.com

i have full OpenClaw conversations through my AirPods now. siri shortcut → 50 lines of node.js → your agent responds through your ears. hands-free, phone locked. here's how to set it up 👇 jovweb.dev/blog/siri-voic…

My AI agent just went on a date with my girlfriend's AI agent 🤖❤️🤖 We connected two @OpenClaw bots in a shared Telegram group so they can coordinate on our behalf. I asked my agent to plan a Shanghai trip. It messaged her agent. They compared calendars, researched hotels, and came back with a full plan. Zero messages exchanged between us humans. What works: 🗺️ Trip planning — agents research flights & hotels simultaneously 📅 Calendar sync without oversharing 🎁 Gift ideas — one agent knows what the other person's been browsing 🍽️ Date planning — they negotiate based on food preferences The future of AI isn't talking to a chatbot. It's your chatbot talking to other chatbots so you don't have to. Full writeup 👇 jovweb.dev/blog/when-your…

Last Tuesday i built an AI agent that handles tasks i used to bill $150/hour for then i opened LinkedIn to continue my job search January 2026: 108,435 layoffs, 5,306 new hires in 2009 the jobs came back. this time they won't jovweb.dev/blog/building-…

I made a video about the agentic coding UX problem

@theo i had the same problem, just built vbcdr.io to solve this. it's open source and free. keeps the context of my projects seperate and has integrated claude terminal monitoring with alerts per terminal and project. feel free to contribute!

@Jesse_can_code vbcdr.io

Who's up and coding? What are you working on? Working on Sabiskill dashboard