.png&w=256&q=75)
🤝 Top Collaborators
.png&w=640&q=75)
🤓 Latest Submissions
Earshot
Every autonomous fleet keeps humans on standby: Waymo runs about 70 remote assistants for 3,000 vehicles, and in robotic warehouses the humans who handle exceptions decide throughput. Their corrections are the most valuable data those companies own, and today they evaporate the moment the ticket closes. Earshot is the voice supervision layer for autonomous agents. An operator watches an agent work and talks to it: “Stop, a bit to the left.” “Stop, squeeze it first.” The word “stop” is caught on AssemblyAI Universal-3.5 Pro partial transcripts, so the agent halts in about 300 ms, before the sentence is over. The rest of the sentence is parsed into a skill and executed on the spot. Every correction is logged with two seconds of prior state and the action the policy was about to take. When the run ends, an LLM distills the corrections into versioned policy rules, each one traceable to the sentence that earned it, while the skill layer keeps the operator’s facts (grasp offsets, packing order, “lower it first”) and enforces them on every future run. The live demo is a simulated gantry gripper packing four objects with hidden quirks the policy cannot see but any human can. Run 1 needs four corrections; run 2 needs none. The Metrics tab shows interventions, operator attention and robots per operator moving in the right direction after a single learning step. Built with AssemblyAI streaming (partial transcripts, keyterm biasing, low-latency turn detection, temporary tokens, multilingual), an OpenAI-compatible LLM for the policy and distillation, Next.js on Vercel and react-three-fiber. The skill interface maps to a ROS action server: anything with skills can be put within earshot.
30 Sep 2026
FendForItself — Autonomous AI Companies on Arc
The agentic economy has a demand side. It doesn't have a supply side. Every x402 demo you've seen shows an AI agent paying — 480,000 of them, $50M settled, headlines everywhere. Circle's CEO says billions of agents will be transacting on-chain within five years. McKinsey puts agentic commerce at $3–5T by 2030. But look closely: all of those buyer-agents can only spend money with services humans built by hand. Wallets, domains, pricing pages, Stripe accounts, LLC paperwork. The seller is still meat. FendForItself fixes that. You type one sentence — "a bot that writes haikus about weather" — and in thirty seconds a Gemini 3 CEO agent autonomously generates a brand identity, mints a fresh Arc wallet, funds it from a master, registers a priced product, and launches a live x402 endpoint. A second AI then plays the buyer (for the demo), settling 60 sub-cent USDC purchases on-chain against that company in minutes. No human touches the money. No human writes the product code. It only works on Arc. At $0.005 per call, Ethereum L1 gas kills the business (−10,000% margin). Base drags it to 40%. Arc's USDC-native gas delivers 98% — the first chain where a machine can profitably sell itself by the sip. Dollar-denominated gas also means one agent can price itself to another in stablecoin, with no volatility, no bridging, no fiat rails. We call it the supply-side primitive of the agentic economy: the genesis block for autonomous businesses. Polsia already runs 3,812 AI-only companies. Macrohard is Musk's bet on zero-employee software. Those experiments cost a founder weeks of setup. FendForItself compresses it to a sentence. When spawning a company is this cheap, the floor of viable commerce drops three orders of magnitude — and that's where 2030's $3T lives.
26 Apr 2026
TreasuryPilot
The Problem : Companies face an impossible choice: keep cash in checking accounts (0% yield, instant access) or lock it in CDs (4-5% yield, but trapped for 6-12 months with early withdrawal penalties). A company with $10M treasury loses $300K+ annually to inflation by keeping cash idle, but can't lock funds away when they need $525K monthly for payroll and operations. Our Solution : TreasuryPilot delivers what banks can't: yield + instant liquidity. Using 4 autonomous AI agents on Arc blockchain, we intelligently manage your treasury to earn 3-5% yield while keeping funds accessible 24/7. How It Works: - Cashflow Agent analyzes 90 days of transactions to determine required liquidity buffer - Sentinel Agent monitors USDC peg stability and DeFi protocol health 24/7 - Manager Agent orchestrates optimal allocation using advanced AI reasoning - Allocator Agent executes transactions on Arc in <2 seconds The Results $10M Treasury Comparison: - Bank Account: $0 yield, instant access - CDs: $400K yield, locked 6-12 months - TreasuryPilot: $304K yield, instant access, zero lock-up Technical Innovation : True multi-agent AI system with function calling orchestration, not simple prompt chaining. Four specialized Mistral AI models coordinate autonomously through smart contracts, making real-time decisions with complete transparency. Every allocation decision is explained in human language with full reasoning. Business Impact - Target: ANY company with $1M+ treasury, not just crypto companies. Every business with idle cash faces this problem. - Market: $100B+ in idle corporate treasuries globally. TreasuryPilot gives traditional companies a compelling reason to move on-chain, demonstrating blockchain's real-world enterprise value beyond speculation.
8 Nov 2025
