Metakgp Demo Day
Author: Vikrant
Status: 75% done, first review, public review
Date: 8/12/17

(For the sake of clarity, this document makes assertions based on personal judgement. I estimate 10-30% of them are not true in some way, but that’s a good enough signal)


Since Vivek and I left campus, there has been a drop in project development with a large user-facing impact. 
  • The highest impact projects – wiki, mftp, mfqp – were all started near the founding period. These projects have >= 1000 unique users / month in the relevant periods.
  • The highest impact projects since then – mcmp, naarad, gyft, dashboard – have seen sporadic development and adoption. Although these are used by a few people, they are relatively incremental, or not in a state where they can be widely used, or not known/useful to many people.
  • Many “non-incremental” project ideas that are most exciting – Android app, youtube caching, KGP auth, DC client – are not being built. We have a plan to fix some of that, this proposal is meant to work alongside that one.
  • Exciting non-incremental projects are probably being built in other places in KGP (eg. I heard something about a blockchain effort), but we’ve either failed to attract those people’s attention, or failed to show real benefits of participating in this community.
  • Compared to most orgs in KGP, we spend more time talking about meta-level strategy and planning (see eg. all the wiki pages outlining how to do things well), but less time actually doing things, and our overall doing/planning ratio is lower than most.
  • We have ~400 people on Slack, but that has no perceivable impact on project output beyond the most active 25 or so members.
  • This state has persisted for over a year now, despite a lot of planning and a little bit of doing (as described above) to fix it.

Proposed solution

Hold a Demo Day once every two weeks, and award a ₹ 1000 prize to a winner randomly chosen from one of the participants, paid by me and Anonymous.

You can demo anything written/made by any number of people, as long as:
  • Many (think 100s or 1000s) of people could potentially use it (ie. not hobby projects written for very narrow use cases)
  • It works; it’s freely available, and will be up for the foreseeable future (ie. not something only shown for Demo Day)
  • The part being demoed is substantially[0] new or different from anything submitted anywhere else (ie. don’t resubmit old projects or things also submitted elsewhere, or done as part of society/department/hall/curriculum work, unless you substantially improved it since then)


Obvious high-impact things you can demo: 
  • features for software projects
  • data visualisations
  • substantial improvements to the wiki

Unorthodox high-impact things you can demo: 
  • a workshop or hackday you successfully organised (for metakgp or on your own)
  • artwork (including posters, videos)
  • administrative decisions you successfully lobbied (eg. getting physical copies of question papers from the exam control centre)

The only caveat is that you should explain how you did it, what you learned, and how others can benefit from it. In this spirit, submissions are also generally expected to be FOSS or CC or similar.


If we create a public tradition of celebrating high impact projects, hopefully people will feel good about it to overcome whatever obstacles are currently in their way, and do it more.

Demo Day is a good faith[1] attempt to counterfactually[2] incentivise high impact projects being worked on and talked about. If successful, over time it will become a healthy tradition.

How will it work?

Submissions will be screened by an Arbiter (initially, me) before the demo, to make sure they meet the qualifying criteria. Ideally we want to accept as many submissions as possible every week, and when we cannot accept a submission, it’s best to explain how the project can be modified so it qualifies. Of course, unaccepted submissions are welcome to be demoed anyway, this will not affect their qualifying for an acceptance the next time.

Accepted submissions will then have around 3 minutes to demo, at a place and time decided in advance. Demos should be broadcasted live so people can watch remotely.

All demos meeting the qualifying criteria will be entered into a lottery, and a winner will be randomly chosen by a computer program.

Good incentives

  • When people have concrete goals or deadlines, they end up at least paying attention to them, and often work to meet them. Demo Day provides a highly visible convergence point for people to work on interesting, high impact projects.
  • The financial reward is substantial enough to make people think about what they can demo, and change their behavior to start working on it more.
  • There’s a social reward in showing off something you built in front of many people. Being able to demo something often marks you out among other interesting people as someone who gets things done.
  • The demo aspect makes things very concrete, which makes it easier for outsiders to observe the tangible impact of being part of the metakgp community, and might encourage them to join. If people already doing interesting projects choose to demo them here, we learn about their existence.
  • The demo is a test of work – if you haven’t built anything, you don’t win these rewards. Talk is cheap.