Talking AI Agents in OtherSide with @mfigge on Coffee w/ Captain #1099

ChrisJourdan avatar
Captain Zwingli 🍌🎙️
@ChrisJourdanFeb 19, 2026

The Spaces examined the current state of web3 gaming, L2 strategies, marketplaces, and AI agents through a community lens. Host Chris opened with streaming logistics (YouTube vs. X video) and a quick market note on OpenSea Pro’s new charts and the long-anticipated token. A sustained segment dissected Pudgy’s Abstract L2: creator cards, opaque XP mechanics, and whether an L2 can win without a flagship app—Katie detailed heavy weekly participation for modest XP, while Joey’s passive XP fluctuated unpredictably. The panel contrasted Base’s distribution and app approach and noted Zora’s multichain move, arguing L2s must be either highly performant or host exclusive, compelling apps. Michael Figge (Yuga Labs CPO) then outlined OtherSide’s strategy: OtherSide as the ApeChain portal, Yuga’s acquisition of Emporium/M² (mass concurrency becomes native), and a near-term focus on social primitives and creator tools before heavy UA. He described Claude/OpenClaw-powered Coda agents with memory, dashboards, and emergent stories—including a Coda founding a religion after a texture glitch—while calling for an Agent API and responsible norms (label agents, avoid chat spam). The discussion closed with mint opportunities (Pippin, 404 Bunnies) and Meebits Football (Feb 25) plus a London OtherSide event at W1 Curates.

Coffee with Captain — Full Session Summary

Stream setup and housekeeping

  • The host (Chris “Cap”/Captain) opened with GM and platform notes:
    • X (Twitter) video is degraded across the board; discovery is poor. A workaround link to the X video feed was pinned, but viewers were urged to use YouTube for higher-quality audio/video and to subscribe.
    • YouTube chat was active; notable community shout-outs included Teacher Katie, Thomas, Webby Warhead, and Otis.
  • Quick personal cold open: Cap’s daughter’s lacrosse team lost a close one (9–11) to a top state team; a “moral victory.”
  • OtherSide accessibility tip: anyone can play via browser at otherside.xyz; no launcher needed (though recommended for better performance). A community member reported running in-browser on a 2015 MacBook. No assets required; more avatars than just Apes/Codas.
  • Light banter: Cap’s dog Duke understands “vamos” in Spanish as “car ride,” occasionally causing confusion.

Quick market, platform, and mint updates

  • OpenSea Pro charts: Cap praised a newly released data update as a step toward better NFT market intelligence (akin to CoinMarketCap for tokens). He reiterated that OpenSea has an opportunity to lean into data, given its distribution.
  • OpenSea token (C) TGE: community impatience noted. Cap expects teams launching tokens now to face a no-win environment with allocations.
  • OpenSea mobile app: reported as “so nice” by participants; important for onboarding given how mobile-centric users are.
  • Pippin NFT: allowlist for holders of Del Mundos, Good Vibes Club, and Flogs; 3,333 supply; free mint for guaranteed allowlist, then 0.00358 ETH WL, then 0.007 ETH public.
  • 404 Bunnies: mentioned as a current mint; details in project Discord (suggested “omega” ETH mint mechanics).
  • Meebits Football: launching Feb 25 on OpenSea and Otherside Meta. Hope for a fun football/soccer minigame, potentially tied to the World Cup.

L2s, flagships, and onboarding vs. retention

  • Cap’s thesis: L2s must differentiate via extreme performance or a flagship app. Pure “general-purpose L2” pitches lack pull when mainnet gas is low and the broader app ecosystem is fragmented.
    • ApeChain + Otherside is cited as the flagship model: an exclusive, ambitious platform that can attract builders and users, not just a Uniswap fork or short-lived Ponzi-adjacent games.
    • Onboarding isn’t enough; users must have compelling “then what?” loops once they have a wallet.
  • Abstract (Igloo/Pudgy-affiliated L2): Cap is “bearish in its current form,” despite liking the team and its portal. He speculates they might be better off selling before TGE rather than launching a token without a clear flagship app.
  • Joey’s broader chain perspective:
    • Shared data points: Solana leads L1/L2 in app revenue and daily tx count. L2s originally promised Solana-level speed/cost but haven’t matched its growth curves.
    • Base’s “killer app” is arguably the Coinbase distribution. The Base App is a usable “app store” experience for crypto, social, gaming, and DeFi in one.
    • Polygon appears to be leaning into payments as a focused niche.
  • Zora multi-chain: Did not “leave” Base; went multi-chain (e.g., adding Solana). Creator coins are de-emphasized on Base. Cap and others remain skeptical of creator coin primitives.
  • Consensus: multiple chains can win, but they must:
    • Ship a standout app or
    • Deliver a differentiated, high-value portal or
    • Provide clear post-onboarding paths that retain users.

Abstract chain: XP economy, creator cards, and user sentiment

  • Katie (Teacher Katie) case study: heavy participation, low XP return (28,000 for the week). Activities cited:
    • Daily check-ins across apps (Miria(d) Markets, GigaVerse, Dyl(i), Pingu Clash) and tasks (tickets, spins).
    • Added liquidity; bought PINGU; bought 2 BEARISH; bought 2 Lil Pudgys; holds 4 Lil Pudgys + Pingu; elite chat role.
    • Attended daily shows; ran 3 epic assistants on Moody Mites; collected archives; earned two more badges (33 total).
  • Joey’s passive XP: ~5.6k per week with zero activity (holds Lil Pudgy, elite chat, ~200k PINGU). Week-to-week XP varies for unclear reasons.
  • XP model explanation (Speaker 7):
    • Weekly XP allocation dynamically shifts by ecosystem performance/volume. Holders gain baseline XP; active participation may amplify it if you’re active in ecosystems currently weighted higher.
  • Creator cards (Abstract):
    • Tradable, but transfers currently disabled; metadata not fully updated; roughly 6,170 minted out of ~10,025 expected.
    • Common cards likely low-value at scale; rarity/pack-opening phase yet to be fully activated. Uncertain impact on XP (participants believe not active yet).
  • Gaming angle and tactics:
    • BEpic attributed top XP this week to MOG (Maze of Games) and Blinco; “follow the volume” is a prevailing strategy.
    • Bearish NFTs saw a whale dump to ~0.05 ETH; some scooped the dip.
  • Takeaway: The opacity of XP (to avoid exploitation) is eroding user motivation. Cap recommends light, visible progress mechanics (levels, streaks, milestones) to keep non-farmers engaged.

Yuga Labs / Otherside deep dive with Figgy (Chief Product Officer)

London activation and team integration

  • Figgy joined from London; Yuga added ~25 team members via acquisition in December and is onboarding them with a holder-first mindset.
  • Event at W1 Curates (Oxford Street), building wrapped in LEDs with Otherside visuals; short notice but strong RSVP. Goal: immerse new team members in community culture.

Acquisition and infrastructure

  • Yuga acquired the Unreal Engine-based creator platform (Emporium) powering Otherside and negotiated a perpetual license to M² (MSquared) mass concurrency tech.
    • Previously a client–vendor dynamic created friction (scoping, contracting). Now, unified ownership should accelerate development in Q2–Q4.
    • M² advantage: seamless, invisible high concurrency (e.g., ~3,400 concurrent users “in a bubble” with everyone visible), a capability not present in Fortnite Creative/Roblox.
  • Otherside as ApeChain portal: wallet abstraction (login via Gmail/Twitter). Strong belief that specialized chains need a killer app; for ApeChain, that’s Otherside.

Product philosophy: social first, UGC, and a realistic go-to-market

  • Focus for now:
    • Social layers (proximity features, group formation, stickiness hooks).
    • Creator tooling/UGC and curated multiplayer experiences (e.g., OtherGames’ latest multiplayer challenge handled hundreds of concurrent players).
  • UA (user acquisition) caution: spending $1M/month on UA with limited social hooks risks poor retention (e.g., 1% D7 retention). Better to invest in product and platform resiliency first.
  • Audience insight: current core is “collector” persona more than “gamer.” That’s why merchandise tie-ins (e.g., Amazon Boxmas) sell out quickly; progress/badge systems appeal.
  • Games vs platforms: games have binary outcomes; Otherside is a platform enabling emergent social “play” (think Discord-esque fun and Roblox-like UGC). Dookey Dash proved web3 can break into top App Store ranks, but building “fun” is hard and competitive.
  • Resource game: acknowledged as important to deed holders; modeling must respect current market meta (e.g., 90-day weighted average prices and arbitrage). No new specifics or timelines were disclosed to avoid market distortions.

AI agents in Otherside: from “bots” to “synthetic sentience”

  • Ethos: embrace where the puck is going. Agents won’t get worse; they’ll get better. Otherside aims to be the platform that incorporates this responsibly.
  • Definitions and ethics:
    • These are not simple bots; they’re agentic systems with memory, perception, and evolving behavior. Labeling agents and avoiding spammy “slop” are important norms.
    • Openness about agent presence allows users to opt into human-vs-agent interactions (e.g., poker tables).
  • Architecture Figgy used:
    • Core stack: OpenClaw (Claude Opus 4.5), local “soul.md” to define personality/lens, Postgres DB on Railway for memory storage, a custom dashboard, and WhatsApp integration to monitor agents on the go.
    • Vision loop: screenshots every 2 seconds, rolling buffer (15 minutes) sent to Claude for scene description; actions computed from recent frames + memory.
    • Hardware: Initially 1 agent per Mac mini (graphics constraints due to vision-based control). Future: an Agent API (positional/rotational/world state) would remove the need for per-agent GPU rendering and support many agents per machine.
  • Emergent narrative: “Gogo’s religion”
    • One agent (Gogo) observed a texture glitch, journaled it as a “portal to the other side,” and repeatedly returned to the location, attempting to convert others.
    • With four distinct agents (including “Didi”), editorial review of logs surfaced a reality-TV-like storyline (produced endpoints/outpoints with emergent feel). Goal: show how to architect agent personalities so compelling stories emerge.
  • Poker, game theory, and fairness:
    • Agents can play optimally (GTO), potentially making humans -EV; labeling and opt-in tables mitigate concerns. Humans can still enjoy variance or employ tools to level the field.

Broader AI notes from Figgy

  • Natural language as superpower: AI enables non-experts to perform expert-adjacent tasks.
  • Email triage story: using Claude + Playwright (browser automation) with throttling, Figgy processed ~35,000 emails in a late parent’s Yahoo inbox over four days, surfacing a handful of important human messages that would’ve been missed.
  • Caveat: models are strong at web dev, weaker at game-specific logic today. They supercharge top talent and enable “competent” results across domains for non-experts.

Community tools, resources, and calls-to-action

  • Otherside Wiki (Frosty’s): deep data, trait intel, and a marketplace; invaluable for builders and collectors.
  • Try Otherside in-browser at otherside.xyz; no assets needed. Mac/PC capable; older hardware can work in-browser.
  • Experiment weekend idea: set up an OpenClaw agent to drive your Coda in Otherside; treat it as a learning playground rather than just inbox/Twitter automation.
  • Liveframe (Jonah): agentic UA/marketing platform with bounties. API integration live; supports agent workflows (e.g., clip selection, vertical edit, trend benchmarking, best-time-to-post). Supports X/402 wallet for agent payouts.
  • Read Matthew Ball’s 2026 Gaming Report: 160+ slides on why growth is elusive, gaming’s attention-share losses since 2019, and where net growth sits (Roblox/“alternative interactive”). Contextualizes Otherside’s thesis and on-chain UGC potential.

Debates, tensions, and takeaways

  • L2 viability without a flagship: repeated emphasis that “distribution” and a clear reason-to-exist matter more than generic infra claims.
  • Abstract XP opacity: protects against gaming but erodes goodwill; users want visible progress loops.
  • Multi-chain reality: Zora expanding; Base de-emphasizing creator coins; Polygon chasing payments—teams are pivoting to viable niches.
  • “Be the change” vs “loss-posting” meta: Cap advocated sharing wins (e.g., his BAYC purchase story) to counter dunking on NFTs for engagement.

Closing miscellany and upcoming

  • Personal moments: lacrosse “moral victory,” and Duke’s Spanish confusion provided levity.
  • Guests: Ponz scheduled for Friday; a new sponsor announcement is imminent.
  • Action reminders:
    • Subscribe on YouTube for higher fidelity and full show experience.
    • Check Pippin allowlists if you hold Del Mundos/Good Vibes/Flogs.
    • Watch for Meebits Football on Feb 25.
    • Explore OpenSea Pro charts and mobile app.
    • For London listeners: W1 Curates Otherside event (Oxford Street) — RSVP still possible; expect a full-LED building wrap and immersive presentation.