What is BitDAO and why is it relevant to science?
Stable, fast-growing DAOs like BitDAO could be the future of Open Science. Metaresearchers take note!
Okay, I admit it, I’m obsessed with BitDAO. But for all the right reasons! I’m a die-hard lover of science, skepticism, and empiricism. Metaresearch has shown us that modern science has serious problems, and Web3 technologies are offering a solution.
BitDAO can make science fun again!
Sure, a new technology is only as useful as the real-world problem it solves. But science has some big problems these days, and blockchain technologies like DAOs can solve them. I’ll explain.
Science’s big problem
Science is biased. It’s biased by its incentive structure, and that bias is becoming inescapably obvious.
In recent years I’ve become enamored with metaresearch (the science of science). This field began in 1966 and is advancing rapidly due to the digitization of large academic databases by indexing services like Google Scholar and PubMed.
“Science is the best thing that has happened to human beings … but we can do it better.” — John Ioannidis
Depending on your perspective, metaresearch is either navel-gazing or effectively introspective. Regardless, like meditation or psychotherapy, it's shown a glaring spotlight on the tools we use to interpret the world around us. And — like meditation and psychotherapy —what we’ve observed is not pretty.
The best data we have shows that science is biased by misaligned incentive structures, authorship bias, funding bias, and sloppy statistics. John Ioannidis neatly summarized the mess in his elegant 2005 PLOS essay “Why Most Published Research Findings Are False”. (Nerds might also savor the statistical analysis in the New England Journal of Medicine of chocolate consumption among Nobel laureates.)
This isn’t a big surprise, more of an open secret. In a 2016 Vox survey, 270 scientists identified the biggest problem facing science today as perverse incentive structures. For instance:
- Science is collaborative and team-based, however, credit and authorship are linearly hierarchical. Women, people of color, international ghostwriters, and researchers with no degrees are often the first to be omitted from formal credit.
- Science can only disprove hypotheses, yet somehow negative results are career suicide.
- Funding is shrinking and skewed toward a select group of advantaged applicants. Roughly 32% of US NIH grant applications were approved in 2000, but this was reduced to 19% in 2021.
- Publication-based promotions, the “infamous publish or perish” system, systematically rewards sloppy or unethical behaviors, and disincentivizes good teaching.
Simply put, we need to restructure.
As I keep saying one can do logic without empiricism, but one must never do empiricism without logic. — Nassim Nicholas Taleb
Given that many of our species' big problems have been solved with empiricism, we have significant incentives to resolve the problems with science, because cleaner science means longer lives.
So how do we fix science?
Blockchain has many more uses than cryptocurrency
Like many other people in my age group, I associated blockchain with crypto, and then wealthy white males making billions, and then (unsurprisingly because I’m not a wealthy white male) lost interest.
Okay, I’ll be honest, I tried to buy several thousand dollars worth of Bitcoin in 2009, got deterred by the technical challenges, and have regretted giving up on it too quickly ever since. Maybe I should have listened to the demographic I just disparaged… But, since I wasn’t the one profiting off early adoption, I ended up in the pragmatist bucket of the blockchain Diffusion of Innovation curve and waited for it to become useful.
I knew the technology had more utility — many of my friends do not fit that profile and have been happily building insightful products with blockchain for years. The emerging ecosystem of these products is collectively referred to as Web3, and it’s proving remarkably valuable. For instance:
- Blockchain is especially useful for preventing small-scale graft across a supply chain. It could be used to prevent pharmaceutical fraud in developing nations or track shipments subject to cargo theft like televisions.
- Blockchain can secure data access. Block-chain-based smart contracts could be used to enable patient ownership of electronic health records or manage collective data-sharing.
- Blockchain can enable complex micropayments. It’s too expensive to pay out small amounts. The need for validators and accounting makes the distribution of micropayments prohibitively expensive which is why sites like Medium have the 100-article threshold. But blockchain is ideal for smart contracts — when collaborating parties have competing interests. Smart contracts allow distributed value and ownership. For instance, ride-share apps where every driver has a small share in ownership as they contribute to its success.
- Blockchain can distribute collective resources. Anyone who has ever tried to split a bill at a restaurant, turn in a group school project, or shared a house with more than one roommate knows that humans don’t always contribute equally in a collective. But blockchain enables distributed autonomous organizations (DAOs) which are great for large collectives because they can flatten traditionally hierarchal Web2 while enabling everyone to profit transparently.
So rather than pick on a technology that is still developing, I’d like to make the point that blockchain is growing up and that people should start paying more attention to it because it's about to become a part of all of our lives.
Scientists in particular.
Blockchain can solve innumerable problems with science
There are countless ways that collaborative benefit structures would benefit research. I’m going to list two in hopes that the conversation continues and expands.
Smart contracts can enable reviewers to profit from peer review. A major problem with science is the incentivization of peer-review. Peer review is the real reason people trust journals like Science and Nature. In order to publish an article there, people who really know what they are talking about (your scientific peers), have to agree that you created an adequate experiment and found something innovative and useful. Not everyone understands the nuances of statistics or can tell when someone using arcane technical jargon is bluffing. Technologists like me who work in many disciplines are reliant on specialists who can take a deep-dive to validate niche research and we often use the reputation of a journal to establish that has been done.
The problem is that peer review is often unpaid which means researchers have to fit it around all their other tasks, and a good reviewer is only rewarded with more work.
If scientific journals used smart contracts, a reviewer could receive compensation for the time they spend contributing to the formation of a solid journal article. This incentivizes collaboration and more formative remarks, rather than competition and critical or hasty reviews.
DAOs could enable collaborative scientific efforts. One of the coolest scientific projects in recent years was the reproducibility project conducted by the Open Science Collaboration and published in Science in 2015. Researchers repeated 100 studies from the Psychology canon and found that while 97% of the original studies had statistically significant results only 36% of repeated experiments did, and for these, the effect size was reduced.
This kind of replication is invaluable to science, but it’s penalized under current scientific structures. The nine principal investigators in the Collaborative voluntarily gave up the benefits of having a Science publication listed in their name in order to advance “alignment between scientific values and scientific practices”.
Traditional authorship structures are radically misaligned with projects like these, forcing scientists to choose between truly groundbreaking projects and the publish-or-perish mandate. But blockchain could reverse this by enabling more nuanced metrics to track contribution.
The Human Genome Project is another project that required the contribution of hundreds of anonymous scientists. Although the double helix was first discovered in 1953, the human genome was not fully sequenced until 2001. It was a public international collaboration spanning 13 years at 39 different institutions. The project was only successful because of data-sharing under the, now-legendary, Bermuda Principles. Collaborating scientists agreed to make all results public within open databases, ideally within 24 hours — “no delays, no exceptions”.
However, despite the promise of the Human Genome Project and its undeniable success, researchers have struggled to maintain this standard. Modern genomics has become balkanized as collaborating sites, and private entities, still have competing interests. Still one of our best models for data-sharing, genomic research shows us how the core incentive structure of science is flawed, wastes resources, and hinders advancement.
Projects like these require complex, multidisciplinary interactions between international scientists working with transparency, pooled resources, and shared incentive structures. The only technology that can adequately address these needs is DAOs. Fortunately, we have working DAO models right now, that can be copied and used.
What science can learn from BitDAO
BitDAO is a collective of 3,400 builders and stakeholders on the decentralized web. It currently administers over $638 million distributed across autonomous entities and labs. (Sound familiar?)
The collective was created by ByBit, a derivatives exchange founded in China but now headquartered in Singapore.
If we want to turn our business from billions to trillions, we cannot exist in a company-form, but a ‘social phenomena’ form. — Ben Zhou, ByBit founder
In 2021 ByBit decided to apportion 0.025% of its total trading volume, $9.3 billion today, and an estimated $1 billion annually, into BitDAO a decentralized funding organization.
In June 2021 it raised an additional $230 million in funding with some of the biggest names in the game: Peter Theil, the Founders Fund, Pantera Capital, and Dragonfly Capital.
The BitDAO token (BIT) is a governance token with voting rights, meaning the organization is truly autonomous of its parent. The DAO is designed to give Web3 builders funding, R&D, and liquidity while crowdsourcing assessment of new projects. While the DAO enables innovation in the space, it also profits when its child companies do. The treasury swaps tokens with the projects it builds, making it possible for investors holding BIT tokens to both control investments, and derisk their holdings by distributing capital across the entire sector.
Innovation points out paths that are possible; replication points out paths that are likely; progress relies on both. —Open Science Collaboration
The reason I’m describing BitDAO in depth is I think this self-replicating model can be directly… replicated… in science.
Technology and science are two sides of the same coin — they share core characteristics such as risk, experimentation, interdisciplinary teams, and complex funding mechanisms.
Like the Human Genome Project, BitDAO represents a transnational organization with transparency and distributed collective incentives. The collaborative work of open-source code closely mimics that of genomic data-sharing. Like emerging research disciplines, BitDAO encodes for itself to both scale and clone. Both require contribution-tracking and resource-pooling.
The public experiment of BitDAO is an invaluable lesson to scientists of what we could accomplish if we thought outside the box, and were willing to take the leap.
I think it's time for decentralized research.
Originally published at https://www.todreamalife.com on May 17, 2022.
