🛠️ unstuck

Unstuck is a desktop app for connecting people seeking help with peers who can help.

Think: a junior developer stuck on something, asking for help, and jumping into a screen-sharing session with peer or mentor.

There are some smarts in how the broadcast is routed (to limit the number of people interrupted).

There's more context and a demo video on our website.

History

The first inklings of Unstuck started as a sketch back when I was in high school. (So, 2005-ish) I remember sketching the design for foldable-screen tablet for students, featuring an app in the toolbar that let you announce you're stuck on something (say, "Grade 10 Math - Unit 5 - Assignment 2") and find someone in the school who could help. (Note: smartphones did not exist yet)

It's probably worth mentioning that my school was "different": we had no classrooms or all-day lectures; instead we spent most of our time on independent study, working through "unit guides," completing assignments at our own pace. We would seek one-on-one help from teachers (during their frequent office-hours), or, ask a friend who was just a little ahead (and often better at explaining things than a teacher). Barely anyone had cellphones back then, so there was quite a bit of time spent hunting down our would-be mentors. Hence the idea.

Fast forward several years...

I was a lecturer and mentor at Lighthouse Labs, teaching web development. Students would work through the curriculum in a large common area, and when someone ran into trouble, they would submit a request through the course website, notifying the mentors; one would grab the request and help out the student. It worked well enough. The technical implementation was embarrassingly flaky for a technology school, but they did eventually rewrite it to be better.

At some point around that time, James and I were occasionally delivering Clojure related workshops and wanted a similar system, so we made one. We eventually added remote audio, video and screen-sharing; but only used it for our own purposes.

In 2022, Mike joined in, we polished the UI and UX and tried to have a go of making it a business. We made it into a desktop app (which was our first serious attempt at using Electron).

My favourite little side-quest in the Unstuck polishing efforts was the audio design. Sound is an oft overlooked part of user experience design. For most technologies, it doesn't matter, but for anything in telecommunications, it's critical! We take for granted the "ring tone", "dial tone" and "busy signal" - but without them, the phone system would be unusable. Somebody had to invent that experience! For Unstuck, we had a similar challenge. We had to indicate to the user various statuses: asking for help (and thus, could have someone connect at any time), connecting, temporarily disconnecting, etc. I sketched out the logic, prototyped some sounds on the piano and worked with a sound engineer I found on Fiverr to come up with the final sounds.

Here is a sequence of sounds, for a student making a request, waiting, having a teacher start connecting, and then end the call (I will say though, without the visual context of the app, it doesn't make much sense):

We had two major business models in mind:

1. Unstuck for internal peer-to-peer use (say, a tech company, using it to facilitate peer-to-peer help between their staff; or an online community for learning a certain tool; or a university)
   1. Unstuck as an on-demand external expert, that could (virtually) pop by your computer in an instant (say, to help with an Excel formula, or setting up an Airtable database)

We started with the first model - to lukewarm reception. Unstuck just wasn't 10x better than "post a message on slack, wait for an answer, then meet with Zoom."

I tried it with Clojure Camp - hoping to experiment with Peer Learning - but it suffered from a critical mass problem: if you have a small worldwide group of potential askers, mentors would need to be available 24-7, with requests only once in a blue moon. ...and if they miss a request, then it sucks very hard for the asker. At a small scale, scheduling one-on-one times worked much better. (What I really need is a large group of students to force Unstuck on... for science!)

We started preparing for the second business model, planning to raise money (because it would require having a bunch of experts on standby to deal with the above-mentioned critical-mass problem and a marketing push, to make that problem as short-lived as possible), but we ran out of steam. (...partly due to a bad experience trying to bring some new partners on board).

In the meantime, we've been using it at Bloom as a replacement for remote pairing apps like Tuple or Pop (even though Unstuck is not meant for that).

The rise of LLMs has filled a lot of the gaps that Unstuck was solving, maybe lowering the demand and opportunity... but it could also be framed as expanding the opportunity:

• there are more people than ever poking around with various technical things, and getting stuck when they hit the limits of LLMs (where a human touch can help)
• Unstuck could integrate LLMs and desktop-control models, to act as a first line of support (and generally, build a solid UX for on-demand desktop help)

I think there's still something there. The dream of an "instant digital fixer" is probably possible.

...but it will take energy that I just can't summon at the moment.