# SIG-DRG 2026-09-03 — Session Notes

Participants:
  - Anuraj R | UTC+0300
  - Maier
  - rafa (UTC+1)
  - shreeram
  - ree is done done
  - Spencer Norwick U-0700
  - wbnns ⚡

---

## 📖 The Reading
The specific title was not stated, but the discussion clearly centers on a paper about **robot swarms that use markings/pheromone-like signals in a physical environment**, involving an automated method for generating robot control software (referred to phonetically as "Habanero"). The environment described uses a photo-reactive floor flashed with UV, and the paper claims computer-generated control software outperformed human-written versions. The exact citation could not be determined from the transcript.

## 🧭 Overview
The group did a round-robin of reactions to a swarm-robotics paper, praising it as "really cool" while probing its limitations (an engineered/constrained environment, and a claim about automated control software that may only hold for simple finite-state machines). Discussion evolved into a generative brainstorm about how to apply the paper's ideas to an upcoming symposium/workshop, coalescing around Will's coined term "ferrabyte" (a digital pheromone). The session closed with logistics about the hands-on robot-kit workshop and a possible physical maze experiment in Berlin.

## 💡 Key Points & Themes

- **Maier — critique of the experiment:** Two shortcomings. (1) The paper ignores robots *changing* the environment (e.g., moving LEGO bricks), which he finds more interesting; this omission is what forces the engineered, constrained UV/photo-floor setup. (2) The "computers write control software better than humans" conclusion is overstated — what they optimized was a small-state **finite state machine**, which computers optimize easily; this approach likely wouldn't scale to more open control (e.g., using LLMs), and suits many-simple-robot swarms more than fewer sophisticated robots.
- **Maier — layers of protocol:** Robots must first agree on *what counts as a marking* (recognition), then agree on *what each marking means* (semantics), and only then tackle a third layer: distributed algorithms for coordinated action. Non-engineered, heterogeneous environments make defining and accurately interpreting environmental changes much harder.
- **rafa — biological/analog framing:** Recommended the book *Vehicles* by Valentino Braitenberg (short, ~140 pages, read five times), about simple sensor-driven vehicles that appear "alive" in aggregate. Proposed the need for a **global robot-interaction protocol** analogous to TCP/IP packet handling, and asked what an appropriate "stigmergic" (stigmatizing, per transcript) protocol looks like in the real world. Suggested pheromones might live on a **digital map overlay** rather than physically.
- **rafa — deeper questions:** Do pheromones live on the map itself or on a *mapping of the map* (e.g., a blockchain-hosted coordinate system, akin to but more specific than what3words)? Digital maps also degrade/diverge locally over time like physical ones. Also floated that **one robot might itself be a swarm of agents** (some agents acting like the robot's "organs"), so a robot may not evolve to be a monolithic agent.
- **shreeram:** Skeptical of the sweeping claim that "Habanero" beats humans in all cases — a determined human might still win. Noted the method is constrained to a fixed set of software modules; unclear from the paper exactly what those modules are and how they're grouped per task. Found it striking that such a constrained setup still yields good outcomes. Later raised teleoperation over wireless (e.g., a mining robot on an asteroid) where data/power is limited.
- **Spencer:** Interested in applying the ideas to the symposium. Suggested a rectangular built environment; robots could come **pre-wired with a shared protocol (like instinct in ants)** rather than negotiating one. Proposed a sliding weight too heavy for one robot but movable by two+ (a way to let robots change their environment, addressing Maier's point). For distributed/remote robots, aim for near-identical environments where agents navigate to the same pose and measure how closely different algorithms converge. Later proposed **battery life as an analog of an ant's lifetime**: send consecutive bots into a maze with limited time to leave "ferrabyte breadcrumbs," so later bots navigate faster; test 1 vs 2 vs 3 bots and randomized time limits.
- **Will (wbnns) — the "ferrabyte" concept:** Framed each robot/agent as leaving behind a **byte-code "pheromone signature" ("ferrabyte")** that other robots parse and act on, appending information recursively until a successful outcome, building up a collective **"ferramap."** Emphasized this could work in **non-visual contexts**, not just the hockey-table visual example. Suggested applications: security/capture-the-flag red-team/blue-team scenarios, unsolved math problems (agents given positive reinforcement to keep going), astronomy's massive datasets, and probabilistic future-scenario simulations where agents triangulate around likely outcomes.
- **Anuraj — implementation angle:** Had read the paper weeks earlier and thought about robot implementation. Works on robot mapping day-to-day; multi-robot mapping produces heavy data structures that must be processed offline on powerful computers. Explored (via Claude) lightweight coordination and landed on **CRDTs (Conflict-free Replicated Data Types)** as a candidate, possibly with a **decay factor** so robots "forget" over time. Considered blockchains but noted they grow large. Liked the "ferrabyte" term as a name for what they might build.
- **Convergent idea — digital pheromones:** Multiple participants aligned on abstracting physical markings into a shared digital layer that robots read/write, potentially enabling coordination even for globally distributed robots with few sensors.

## 🔀 Questions & Disagreements

- **Does the paper actually demonstrate "control software"?** Maier questioned whether optimizing a small finite-state machine really supports the claim that computers write control software better than humans.
- **Is the "beats humans in all cases" claim robust?** shreeram doubted it would hold against a truly determined human.
- **Where do pheromones/ferrabytes live?** rafa's open question: on the physical map, on the map itself as data, or in a separate mapping layer (e.g., blockchain-based coordinate system)?
- **Is a robot one agent or a swarm of agents?** rafa raised this as an open framing.
- **Simulation vs. reality gap:** Maier flagged that the paper mentioned one algorithm working better in simulation than the real world but said nothing further — he argued this failure mode is exactly what makes robotics hard and worth understanding; it also threatens the "shared digital world" idea unless done cleverly.
- **What is a "ferrabyte" precisely?** Maier stressed a byte can be data *or* a program (a robot could "eat" a ferrabyte and be reprogrammed, like ants reacting to pheromones); ideally it carries **spatial meaning** (two markers in different orders should mean different things). Digital pheromones could allow far more distinct signals (e.g., a thousand) than nature.
- **Symposium environment decision:** Spencer asked whether the group should decide soon between co-located robots vs. distributed robots — left unresolved (co-locating many robots may not be feasible).

## 🔗 References Mentioned

- *Vehicles* by Valentino Braitenberg (book recommended by rafa).
- "Habanero" — the paper's automated control-software generation method (name per transcript; spelling uncertain).
- TCP/IP — cited by rafa as an analogy for a global robot-interaction protocol.
- what3words-style three-word coordinate system — analogy for a shared coordinate system.
- CRDTs (Conflict-free Replicated Data Types) — suggested by Claude to Anuraj for lightweight distributed coordination.
- Blockchain — considered and largely set aside due to ever-growing storage.
- Claude (AI assistant) — used by Anuraj to explore implementation.
- Stigmergy / "stigmatizing protocol" — concept referenced by rafa (transcript wording unclear).
- A recently solved mathematical problem where an agent was reinforced to "keep going" — mentioned by Will (specific problem not named).

## ✅ Action Items & Next Time

- **Anuraj:** Explore how to actually implement digital pheromones ("ferrabytes"), particularly CRDTs with a possible decay/forgetting factor.
- **Group:** Keep thinking about the protocol layers (recognition, semantics, coordination) and design experiments once enough members have robots assembled and are comfortable with the platform.
- **Decision pending:** Whether the symposium environment will be co-located vs. distributed robots (raised by Spencer, not resolved).
- **Possible experiment:** A physical maze in Berlin (Anuraj and rafa), or a digital maze with robots propped up — time permitting.
- **Workshop logistics:** Attendees should buy and assemble the robot kit *before* the workshop (assembly takes time and shouldn't consume workshop sessions); detailed tutorials come with the kit; help available on Discord. The workshop will cover registering the robot on-chain, crypto payments, teleoperation, and a marketplace for tasks.
- **Next meeting:** In two weeks.

## ⭐ Memorable Quotes

- **Maier:** "That is exactly why robotics is difficult because the real world does not work like a simulation."
- **Spencer:** "Plus one for a terabyte. I hope that makes it into the symposium." *(referring to the coined term "ferrabyte")*
- **rafa:** "When you teleoperate, so the humans are the ants. Right? ... you can move the robot and you can see the ferrabytes or you can drop a ferrabyte, but you can't see the world."

*Note: The transcript's diarization heavily interleaved speakers (especially rafa, Maier, and Will talking over/across each other), so some attributions are reconstructed from context and may be imperfect. The "ferrabyte" term appears throughout as both "ferrabyte" and "terabyte" due to speech-to-text errors.*