Initiative to provoke a systemic change We need to realign incentives with the scientific endeavor
Venues are profiting from Science, not contributing to it anymore. Citations, h-indexes and impact factors are flawed to the root.
Now is the time for a new scientific system to be born!
How to properly measure scientific progress? Even if a work has been published in a top tier journal ... does it actually work? Would you bet on it convincing the community on the long term? What we want to evaluate is how credible hypotheses are, and what evidences made them so.
With Bycelium,
Easy to fake, even more with LLM based AI generation now. Paper mills and review mills are surging because it became an easy business to generate fake science. Publication should not be the goal. Impact *after* publication should be what matters to a scientific career.
Data gathered should always be shared and valued, regardless of the interpretation it was gathered for. It can be interpreted later on in light of a new theory or in meta-analyses. An evidence can be an interpretation of data with a theory to update the credibility of an hypothesis. It can also be a mathematical proof or anything that would change the beliefs we could have on new or existing hypotheses.
Peer-review is a bloated system that fails at filtering misconduct and irreproductible experiments. It became gatekeeping and an labor exploitative money machine for publishers. Let's stop fooling ourselves that peer-review before publication is a good practice and use it for what it should be: Provinding non-blocking feedback to build better science together. Peer-review should not evaluate if a work is 'worthy' of publishing before we could even try to replicate it. It worked before the 70s, we can make it work again in the information era.
What actually matters regarding peers is: How much were you able to convince them with evidence you have provided? Our goal at Bycelium is to find proper ways to gather the updates of credibility of hypotheses that an evidence has provided. We are investingating multiple ways, an example to both assess the belief changes and give monetary incentives for both positive and negative replications is to have a betting system on hypotheses credibility evolution, this is similar to replication markets.
About Bycelium in general
An alternative to review-filter-publish is publish-review-curate.
The review-filter-publish method worked when the quantity of works to review was much smaller, allowing the filter to be efficient.
Today the filter is lotterie and publishing became a goal in itself. Contributing to science is not just publishing, contributing to science is sharing experiments that will strenghten or shake or beliefs.
Switching to publish-review-curate allows us to let time and replications be long term filters of what is relevant.
It is possible to build a system robust to intentional attacks from individuals and small groups (for example decentralized finance).
Here with mechanism of contributor's credibility, we have a defensive mecanism that allows us to increase the cost of gaming the system.
Trying to cheat, and doing bad reviews, would lower your credibility and reduce how much your reviews impact the community.
Once you have a defensive mechanism like this, it is just a matter of tweeking the equation to make it more expensive to cheat than the money you would get anyway.
In comparison, the current citation system only has retraction as a defensive mecanism, and we all know how well it works at dissuading misconduct.
This is why the current system can be gamed by single individuals and even worst by small groups of researchers citing themselves in circle.
Discontent for the current system is at its peak, and the pressure given by language model will push it to its breaking limit.
We hope to build a sane and self-sustained alternative to switch to.
As the number of empty AI-generated papers and reviews grows, the power given to alternative systems will grow too.
Bycelium will be one of them, and if we build it well together it will be the best of them.
Yes! We would love to collaborate with anyone that want to push for better science!
This includes Universities, non-profit and for-profit organisations. If you want to partner in any way, contact us!
How Science is broken for you? How it should evolve?
Contact us: contact@bycelium.com
Follow us changing Science: LinkedIn