# SIG-DRG 2026-10-01 — Session Notes

Participants:
  - Anuraj | UTC+0300
  - Spencer Norwick | UTC-7
  - Andre Comeau | UTC-5
  - Jenna Dixon | UTC-5
  - rafa (UTC+1)
  - Timber | UTC-7

---

## 📖 The Reading
Not clearly identified in the discussion. This session was a retrospective/planning meeting reviewing a hands-on remote robotics workshop run by the group (via the "Yakrobot" / protocol project), rather than a discussion of a specific written text.

## 🧭 Overview
The group used this session to debrief their first fully-remote robotics workshop, in which participants bought and assembled a camera/sensor robot kit, connected it via crypto payments, and controlled it with LLM agents (Claude Code / Codex). Participants shared their build experiences, discussed how to improve kit selection, the software/crypto UX, and facilitation, then brainstormed turning the setup into a repeatable, possibly revenue-generating product — including streamed "arena" games and a Halloween-themed maze event. The conversation closed on near-term goals: choosing a better kit, building a simulator, and drafting a pitch to seek grants/partnerships.

## 💡 Key Points & Themes

**Workshop retrospective**
- **Anuraj:** ~14 registrations, ~10 on the mailing list, with several dropping to attend other symposium workshops (e.g. a SIG/FPT-related one). Actual participants included Andre, Jenna, Spencer, Rafa, and Giovanni. He emphasized that running the workshop fully remotely was much harder than his prior in-person (robotics) workshops, where you can see participants' screens and help directly.
- **Goal:** make the workshop fully repeatable — a Yakrobot website section with region-specific kits (US / EU, possibly a Bangalore, India lab contributing to the protocol), documentation an agent can read, so others could run workshops in their own maker spaces.

**Build experience**
- **Andre:** found parts generally intuitive to fit (he works in construction), though some required backing off/loosening screws to seat others — a potential roadblock for less hands-on people. Building the robot was easier for him than setting up the Raspberry Pi and loading the program.
- **Spencer:** committed early, procrastinated, nearly backed out, but assembled it just in time and was very glad he did. Stressed that solo pre-building was valuable.
- **Jenna:** assembled the tumbler easily and enjoyed it; her first Frino kit arrived pre-assembled and broken, the replacement arrived after her window had closed, and she got stuck on a white camera cable (possibly broken). She noted "kits are an art form" and has since switched to a software/virtual rover to still participate.

**Kit quality & selection**
- **Rafa:** the kit was "okay" but a better kit and a proper bundled "protocol-institute kit" (one-click bundled purchase) would be an improvement; suggested an agent-driven "kit buying portal" (drop-shipping style, possibly including custom PVC parts). Framed a one-year objective: "can we find a better kit?"
- **Anuraj:** showed a new candidate kit — slightly cheaper (~€100–120 target), sturdier, with a separate chassis (not using the PCB as the chassis), a dedicated motor-control PCB that sits on the Raspberry Pi, and 3D-printable acrylic structural parts for easy replacement. Budget is the key constraint (Raspberry Pi prices rising due to memory shortage); wants it kid/family-friendly to broaden participation.
- **Spencer:** sympathetic to price but not sold on the ultrasonic stack; suggested a small LiDAR or better sensor could "disproportionately unlock a lot" of capability (e.g. SLAM).
- **Anuraj:** a ~$75 2D LiDAR is an option, or motors with encoders for odometry plus a monocular camera could enable some SLAM — a design choice to be made.

**Software / crypto UX**
- **Andre:** connecting to crypto forced him to learn robotics + Claude Code + crypto in one intense session, which he valued — but it adds real friction (MetaMask wallet, converting real money to crypto) and may discourage newcomers.
- **Anuraj:** plans to default the whole flow to a testnet (fake USDC, no ETH needed), with mainnet optional later; also building a desktop GUI for the gateway (buttons/API-key setup) since command-line steps are hard for many.
- **Rafa:** recommended abstracting away crypto for the *experience* layer (keep it for the nerdy/protocol layer); using Previa/Pervia + a Paymaster (sponsored gas) would solve ~90% of issues and makes chain choice irrelevant.

**Design-for-LLMs philosophy**
- **Rafa:** cater to people running AIs in Claude/agents; design products so documentation is fully accessible and includes instructions for agent exploration — maybe wikis matter less than agent-readable docs. He wants repos written for LLMs and continuously updated so a new agent can one-shot setup, and proposed building evals (random model attempts setup; loop to improve the repo).
- **Anuraj:** write the AGENTS.md so an LLM agent can read it and configure the robot.
- **Rafa:** "the bitter lesson" — he just pointed an agent at the repo and said "go at it"; it failed twice but better instructions/feedback should enable one-shot setups.

**Facilitation**
- **Rafa:** the small-group format worked well; this workshop would be very hard with more than ~half a dozen people because it needs participant initiative and a certain intimacy and mutual drive to get it "over the line."
- **Andre:** suggested breakout rooms for 1:1 help as an easy improvement for larger groups.

**Product / "arena" vision (the energizing thread)**
- **Spencer:** built a living-room "arena" (shoebox garage with front covered), sent his link to friends and had his teenage nephew drive the robot starting "in the dark" with only robot vision — discovering the environment and robot affordances. Set up games (tennis ball, "rocket league," a ramp to his table). Found it very generative; a friend told him "you could make money off of this." Proposed a viewer/queue lobby and a Halloween haunted-house maze event where people pay ~$5 to solve a spooky maze.
- **Rafa:** loved forking the interface (googly eyes, pink candy-crush-style UI). Pitched capture-the-flag/maze streaming, IRL video game mechanics (operator doesn't know robot affordances while viewers watch), multi-robot coordination, multiple drivers, simulation-based versions, and "portal" mechanics where solving one room teleports you to a robot in a different city. Suggested Christmas/Thanksgiving themes with a kit as a prize.
- **Anuraj:** compared the concept to the TV show *Legends of the Hidden Temple* (and a maze-race format); wants to stream head-to-head competitions, possibly tied to DevCon (Nov 3–6, India).
- **Andre:** as an RTS player, pitched controlling *swarms* of robots against a competitor by prompting an LLM.
- **Jenna:** proposed a software/virtual robot option for people short on time, and raised self-updating docs via a "space bot" that keeps a wiki up to date (noting human-maintained docs are "a lost cause").

**Resourcing / sustainability**
- Recurring "how will you make money?" question from people Anuraj has pitched (symposium, university).
- **Anuraj:** longer-term vision of a few small physical arenas/labs in Europe and North America where people run experiments and collaborate (e.g. with the FPT group).
- **Rafa:** suggested pooling community money (e.g. contribute $50 instead of Uber Eats) toward one person building a higher-def "community robot" with a local arena; and pitching for a grant — even ~$5,000 would go far toward a proper arena/"fighter robot." He has connections at University of Zurich and University of Michigan; Andre has autonomous-car contacts at University of Ottawa (pitch wasn't well-received yet; they focus on centralized, not decentralized, robotics) and built a full-scale robotics arena as a concrete finisher. Spencer offered to connect with University of Washington robotics labs.
- **Rafa:** modularity idea — standardize on one Raspberry Pi per SIG member that is compatible with multiple kits, since computers are getting expensive.
- **Rafa:** referenced an internet friend's project that uses a math engine to design Arduino/Lego robots from a prompt, generating a bundled parts list + Lego-style instructions — imagining "one-shotting a robot into existence" in a few years (not reliable enough yet).
- **Rafa:** recalled late-90s/early-2000s Lego robotics kits using a rollable mat with symbols/light-dark zones as a portable arena MVP.
- **Anuraj:** mentioned AprilTags/ArUco-style fiducial markers (printable on a normal printer) for camera calibration and robot localization as a cheap arena-building tool.

## 🔀 Questions & Disagreements
- **How much to kitify vs. DIY:** Spencer framed a tension between a "one-click order kit" and "source it all yourself." Rafa and Spencer agreed the right balance (and the value of personal "buy-in" / the "IKEA effect") will be found by iterating. No firm resolution.
- **Crypto's role:** Andre valued the forced crypto learning; Rafa wants it abstracted away from the end-user experience. Partly reconciled via testnet + Paymaster/sponsored gas, but the mainnet question ("do we even want mainnet?") was left open by Anuraj.
- **Sensor spend vs. budget:** Spencer wants better sensors (LiDAR/SLAM) even at higher cost; Anuraj is anchored to keeping the kit affordable/kid-friendly. Pooled-funding and community-robot ideas offered as possible middle ground.
- **What exactly is being pitched:** Rafa and Anuraj agreed a pitch is needed first, but acknowledged uncertainty — "Do we understand what we're doing? I don't know. But we know it's magical."

## 🔗 References Mentioned
- Claude / Claude Code, Codex (OpenAI) — agents used to program the robots
- MetaMask (crypto wallet)
- "Previa"/"Pervia" and a Paymaster / sponsored gas (account-abstraction tooling — exact name uncertain in transcript)
- Raspberry Pi (noted price increase due to memory shortage)
- AprilTags / ArUco-style fiducial markers (robot localization/calibration)
- LiDAR (~$75 2D unit); SLAM; wheel encoders / odometry; TT "yellow" motors
- Lego robotics kits (late-90s/early-2000s) and a prompt-driven Lego/Arduino robot-design project (Rafa's internet friend)
- *Legends of the Hidden Temple* (Nickelodeon TV show) — maze/obstacle inspiration
- *Factorio* — inspiration for a swarm-based strategy simulation (Rafa's proposed future project)
- Oculus Rift VR haunted house (Spencer's prior college event)
- Universities: Washington (Spencer), Michigan & Zurich (Rafa), Ottawa Autonomous Car Research Center (Andre)
- DevCon (India, ~Nov 3–6); the protocol symposium; a lab in Bangalore, India
- Kits referenced: the "Frino"/"free no" kit used in the workshop; a "tumbler"; a new candidate kit (EU link to be posted)
- A possible SIG/FPT group (exact name unclear from transcript)

## ✅ Action Items & Next Time
- **Anuraj:** ordered a new candidate kit (sturdier, separate chassis, modular motor-control PCB, 3D-printable parts); expects delivery in ~10 days and will test it. To post the kit link (EU, plus a US equivalent) in today's thread.
- **Anuraj:** build a desktop GUI for the gateway; set up a testnet-based flow (fake USDC, no ETH needed); write an AGENTS.md so agents can configure the robot.
- **Jenna:** to look up the kit on the US-equivalent site and check the one-star reviews for common failure points.
- **Group:** draft a standardized **pitch** (deciding both *who* to pitch and *what*) toward partnerships/grants — Rafa to potentially mobilize Zurich/Michigan contacts, Spencer to explore University of Washington, Andre to revisit Ottawa/Toronto contacts.
- **Group:** design a **games event** — Spencer and Rafa keen on a Halloween-themed maze/haunted-house arena (Halloween is ~1 month out); aim to spin something up that can also be shown at DevCon (Nov 3–6). Spencer offered to build the Halloween maze and help spruce up the interface/pipeline.
- **Spencer:** proposed a dedicated **ideation meeting** to gather "reach" ideas, then distill to what's buildable in ~a month and chargeable.
- **Anuraj:** to build a **simulator** once a kit's CAD model exists (same code drives real and simulated robot); Rafa open to contributing funds toward a shared higher-spec "community robot" + local arena.
- **Next session:** in ~2 weeks; continue developing the kit choice, the pitch, and the games event. Members interested can buy/test the new kit.

## ⭐ Memorable Quotes
- **Jenna:** "Kits are an art form, and a good kit is so great. And a bad kit, you know, it's just tough."
- **Rafa:** "What it means to be a centaur in today's world where you're not doing it alone."
- **Rafa:** "Do we understand what we're doing? I don't know. But we know it's magical."