Skip to content
Pico

Updates

The enclosure has gone through two design studies and the v0 digital prototype. Nothing has been printed yet, and the firmware and agent connection have not started.

Progress log

  1. Reading list expanded

    Added six lab papers and four HRI 2024 papers covering robots that ask for help, overreliance on AI and physical presence.

  2. Parts checked against local listings

    Picked target modules available in Bangladesh and listed the measurements to take before printing.

  3. Pico v0 assembly prototype

    Removable rear cover and tray, display clips, button carrier, ESP32 sled and 20 printable STL files with interference checks.

  4. Compact body study

    Shrunk the body to 100 × 104 × 100 mm, about 44% less desk space than the first layout.

  5. First layout study

    First millimetre-scale Blender layout of the enclosure and components.

Hardware revisions

Pico first layout study render
Study: first layout
Pico body study render with a face on screen
Study: compact body
Pico v0 assembled render
v0: assembly prototype
Pico v0 with rear cover removed showing electronics
v0: rear service access
Exploded view of Pico v0 printed parts labelled A to P
Pico v0 printed parts. Print the fit tests J, K, L, M and P before the full shell.

Reading list

These papers inform parts of Pico's design. None of them tests an approval device for coding agents.

Ten papers collected on 26 September 2026.
PaperSourceWhy it matters for Pico
Robots That Ask For Help: Uncertainty Alignment for Large Language Model PlannersRen et al., CoRL 2023When an AI planner should stop and ask a person.
To Trust or to Think: Cognitive Forcing Functions Can Reduce Overreliance on AIBuçinca, Malaya and Gajos, CSCW 2021Making people think before accepting AI output.
Tangible Bits: Towards Seamless Interfaces between People, Bits, and AtomsIshii and Ullmer, CHI 1997Physical controls and ambient displays.
Effects of Nonverbal Communication on Efficiency and Robustness in Human-Robot TeamworkBreazeal et al., IROS 2005Expressive signals help people track robot state.
The Benefits of Interactions with Physically Present Robots over Video-Displayed AgentsBainbridge et al., Int. J. Social Robotics 2011Physical presence versus an on-screen agent.
Asking Easy Questions: A User-Friendly Approach to Active Reward LearningBıyık et al., CoRL 2019Questions people can answer quickly.
Understanding Large-Language Model (LLM)-powered Human-Robot InteractionKim, Lee and Mutlu, HRI 2024Physical LLM robot versus text and voice agents.
Reactive or Proactive? How Robots Should Explain FailuresLeMasurier et al., HRI 2024How to present the error state.
When Do People Want an Explanation from a Robot?Wachowiak et al., HRI 2024When to show more context on screen.
Dimensional Design of Emotive Sounds for RobotsWolfe, Su and Wang, HRI 2024Designing notification sounds.