Open Source Competitions and Prizes #1

I recently read Future Perfect by Steven Johnson. The book articulates what is meant by a person being a “Peer Progressive” – someone with a set of values based on the power of distributed networks to solve problems (a “Baran Web”) as opposed to a more regulated, centralized mechanism (a “Legrand Star”). It’s an interesting read that gets better as the book progresses. Lots of stories, but there is one particular gem.

There has been recent talk about prizes being used to alter the way drugs are found, such as Bernie Sanders‘ proposals for ways of funding the prize pot. I’ve written before about the idea of a teamless competition for open projects. We planned this, but have not yet deployed it – more about that later.

What I find interesting is Johnson’s description of the use of prizes (“premiums”) in mid-18th century England to find a way to produce a deep “Turkey Red” dye derived from madder plants. The organization administering this prize, and others, would become the Royal Society of Arts (RSA); the competitions were very successful. The thing that struck me most about the story was the mechanism of promoting innovation – no patents in a competition with a significant prize where the aim is something of commercial significance. As Johnson quotes from the Rules and Orders published by the Society in 1765: “No person will be admitted a candidate for any premium offered by the Society who has obtained a patent for the exclusive right of making or performing anything for which such premium is offered.” As Johnson says “Prohibiting patents meant that solutions could circulate more quickly through society, and could be easily improved upon by subsequent inventors or scientists. The RSA wanted you to profit from your idea. They just didn’t want you to own it.”

This, to me, is a crucial aspect of whether patents are harmful for drug discovery or whether they promote innovation. This also goes to the heart of how to run prizes for drug discovery. If we want the best ideas and for those to flow freely, we shouldn’t have a secretive research process. Jamie Love’s suggestion was to de-link the research process and the process of reclaiming costs. The deal is: make me a drug for cancer and I will give you a billion dollars, but you don’t have any say in how much the medicine will cost. Will that work? I don’t know, but it’s a possible solution that would allow the free sharing of information during the discovery process, which is why it’s interesting. I predict in the near future we’re going to see more government/NGO money going to support contract drug discovery in big pharma – that is part-way to the same goal of removing the need for excluding people from the process.

The open competition structure was applied in the Matlab programming competition – an open process, with a prize, where everyone can see the previously-submitted solutions. Yes, this means you are able to tweak someone else’s solution a little and submit it as your own, but this is surely the way to get to a good solution most quickly.

The open source drug discovery project I’m part of has a completely open structure (without the prize) as spelled out in the six laws. It’s crucial the structure is a Baran map (a redundant network), not a Legrand Star where everything goes through a central point (Linus doesn’t scale). The needs of the project need to be clear, and the structure must be such that anyone can take part. The network also needs to have redundancies and be robust. Changes to the structure have to be easily accommodated. So can such an open arena be combined with a prize? There are lots of things to think about here, and one of the reasons we haven’t deployed a prize yet is because we want to get it right. We don’t want to set something up that is too small, or with the wrong structure. How do we ensure that participants contribute everything into the public domain? Will the money pervert the incentives, possibly irreversibly in the sense that when the competition is over there will be a serious drop off in participation (as in Deci’s classic study on intrinsic/extrinsic motivation (PDF)).

I’ve been talking a lot about this with a talented guy I know, Jaykumar Menon. Jay secured resources from the Rockefeller Foundation to host a meeting to address this issue for a few days in 2014 at their Bellagio Centre, and now we’re setting it up and organising who’s going. I’m starting to talk about it now to garner insights people might have in advance. It’s very exciting. I’m optimistic that in those few days we can thrash out a good way to do this – make a competition long-term useful and actually effective.