Max / AI-Powered Network Modeling & Digital Twins
Max shared his recent exploration of AI-powered agent-based modeling for tokenomics simulation. Rather than using simple rule-based agents, he’s been experimenting with AI personas (digital twins) that make more sophisticated decisions — informed by goals, limitations, and decision-making patterns. He referenced the Highlands framework and Google’s simulation engines for modeling complex multi-agent systems, as well as an Anthropic hackathon winner that created AI customer personas for product validation. The work is early-stage but aims to produce more realistic tokenomics simulations than traditional approaches.
Brandon / Regen Market Monitor & Velf Trading Agent
Brandon presented his Regen Market Monitor — a trading agent that uses REGEN as its base asset. The logic: the agent exits REGEN, finds profitable trades on venues like HyperLiquid, Polymarket, and GMX (on MegaETH), returns profits to REGEN, and retires eco-credits with surplus. It can bootstrap itself without initial funding by “research mining” (contributing to decentralized AI research datasets using free models like Qwen to earn compute credits). Brandon deployed the agent on a Privy wallet with full transaction signing autonomy. He also contributed to four community repos: Heartbeat, Agentic Tokenomics, the AI Claude repo, and the Regen repo. The agent spec was built from the Agentic Tokenomics repo architecture.
Christian & Brandon & James / Trading Bot Experiences
Christian shared his experience building three different trading bots: an LP trading bot on HydraDX (entering tight-concentration positions at high APR for 3-minute cycles), a Polymarket strategy bot, and a HyperLiquid funding-rate arbitrage bot running for a month (~2,000 trades, up ~$1). Key insight: paper trading consistently shows profits, but real trading struggles due to slippage and execution differences Claude doesn’t account for. Christian runs his bot on a VPS scanning 10+ criteria to trigger trades via CLI — more of a looping bot than an autonomous agent. James shared that his OpenClaw agent achieved ~6% returns on Polymarket in two weeks with low-risk likely-payout trades. The group discussed creating a shared private repo for trading bot knowledge exchange.
Christian / Regen Compute UI Redesign & Marketing
Christian announced the new Regen Compute frontend redesign, created by Samu from Gaia AI and implemented across all pages with Claude. The staging version is being tested for details before going live. Key improvements include a more professional look (no longer “looks like it was made by Claude”) and the call-to-action moved above the fold. Current subscriber count is 33, up from 7 (all insiders) — showing external conversion. Gaia AI plans to send a newsletter to their 10,000 subscribers focused solely on Regen Compute. The Regen Compute Twitter/X account was suspended (likely due to algorithmic bot cleanup), which complicates social media marketing. Dave shared a LinkedIn scraping strategy. Christian invited anyone with marketing energy to join the marketing repo.
Group / Enterprise Regen Compute Vision
The group explored what an enterprise tier of Regen Compute could look like. Lance proposed a dashboard where company admins enter multiple API keys (one per developer), enabling cumulative organizational usage tracking and offset calculation based on actual AI inference data rather than flat subscription tiers. Christian noted this is a significant technical upgrade — even the retail version doesn’t yet track actual API usage (subscriptions are self-reported tiers). Enterprise users would need: combined company-wide dashboards, individual developer contribution views, and verifiable data tied to real usage — essential for companies treating this as a certification or emissions offset. Brandon emphasized that the enterprise version targets developers and capital, since “everyone’s using Claude at this point.”
James / Bridge.Eco V2 & 0G Agent Platform
James announced Bridge.Eco V2 is live (not yet pushed to production), adding direct eco-credit retirement and staking alongside the existing swap functionality. He’s also about to launch a new agent creation platform on 0G (Zero Gravity chain), where users can deploy agents on decentralized private compute with custom skills and Token Flight embedded. EcoBridge runs a validator on 0G with ~$10M secured, funding chain carbon neutrality via automated credit retirements (~24 credits/day at ~$3 each through climate.zerg.ai). James noted the political headwinds: Solana removed their climate dashboard and carbon-neutral commitments shifted after the Biden administration. He emphasized that selling impact requires individualized stories per buyer, not generic pitches — “nobody actually cares about impact, it’s just the story they can tell.”
Lance / Social Media Outreach Agent Concept
Lance proposed building an agent that scans social media for accounts meeting certain criteria, qualifies them as potential partners or customers, and generates engagement approaches — essentially an influencer/outreach strategy for Regen Compute. Brandon pushed for a more direct sales approach over partnership structures, arguing the regenerative claim itself is the value proposition. The group discussed the tension between influencer outreach vs. corporate client acquisition.