This episode with Richard Sever, co-founder of bioRxiv and medRxiv, focuses on preprints. How come they have been so widely adopted as a means of disseminating research? Which factors helped, and what made the process difficult? We also discuss the place (and time) of scientific validation in a scientific system where work is published and shared rapidly, and independently, without in-depth review.

In this episode, we talked to Professor James Zou, who brought us his perspective on how academia might collaborate with AI. He covered how AI could help us ask better questions instead of answering them, how they can translate information for different levels of expertise, and how we can use them to make our feedback more diverse and specific instead of general. After that, we explored how AI is already changing science by increasing the number of papers, creating more general GPT-generated reviews, and also making writing a more accessible task for incoming PhDs. To top it off, we all agreed that we would absolutely go to an AI-generated concert (especially if there were robots). James Zou is an Associate Professor of Biomedical Data Science and, by courtesy, of Computer Science and Electrical Engineering at Stanford University. He works on making machine learning more reliable, human-compatible and statistically rigorous, and is especially interested in applications in human disease and health. Several of his algorithms are widely used in tech and biotech industries. If you enjoyed this episode, check out our other seminars here.

James Evans, Researcher on designing innovative research ecosystems, discusses the importance of diversity in driving innovation. Topics include dimensions of diversity, AI's role in diversifying scientific ideas, and the potential for AI to replace human scientists. The conversation also delves into abduction in scientific reasoning, embracing diversity in innovation, and the balance between disciplinary silos and crossbreeding for innovation in universities.

Taylor Oshan has been exploring how use and take advantage of decentralized infrastructure like IPFS and Filecoin for geospatial data. In this podcast we learn what geospatial data actually is - rumor has it it’s about 60% of all data! Then Taylor explains why decentralizing is an important effort and how it can increase the robustness as well as efficiency of data storage and access. Taylor is a geographic information scientist at the University of Maryland and part of the Easier Data Initiative.

Allison Duettmann, president and CEO of Foresight Insitute, talks about predicting progress, interdisciplinary collaboration, and the biggest risks facing humanity. They discuss the challenges of mapping out tech trees, the significance of AI in biotechnology and nanotechnology, and their favorite books and interests, including personalized longevity techniques and roller dancing.

In this episode, we sat down with Daniel Hook CEO of Digital Science. Digital Science is a company that invests in software companies servicing researchers and research institutions. Daniel shared his personal journey from being a researcher in theoretical physics to starting his own software company to now leading innovation in the scientific tooling space with Digital Science. We also talked about the philosophy and ground revision behind this company, as well as what Daniel thinks the future of science should look like. 

In this episode we talked to Davide Grossi about peer review. Davide is a decision making and mechanism design researcher at University of Groningen and the University of Amsterdam.  We first covered the state of peer review today: How it’s being done and how it’s going wrong – Davide likens the outcome of reviews to a lottery: Whether you get a positive or negative review seems to depend more on chance than on the quality of your work. Then we dove into potential solutions and new designs for a fairer and more effective peer review system. Besides reducing the amount of reviews in general, Davide suggested some interesting new avenues for exploration when it comes to how we decide which proposals and publications get the limelight or funding. There’s a lot more to be done before we find the “best” way to perform peer review. Davide is calling for a science of peer review to systematically engage with the design choices around this pillar of scientific integrity. Watch the full seminar: https://youtu.be/jq_9BqwMK_U Sign up for future seminars: https://descifoundation.org/seminar Support DeSci Foundation: https://descifoundation.org/donate

Josh Nicholson is co-founder and CEO of Scite, a deep learning platform that evaluates the reliability of scientific claims by citation analysis. Previously, he was the founder and CEO of the Winnower (acquired 2016) and CEO of Authorea (acquired 2018 by Atypon), two companies aimed at improving how scientists publish and collaborate. He holds a PhD in cell biology from Virginia Tech, where his research focused on the effects of aneuploidy on chromosome segregation in cancer. In this episode, we dove deep into metrics: Which stats should you optimize for to progress in your career as a scientist? And how can this inevitable culture of optimizing for one-dimensional metrics hurt the pursuit for truth? Josh told us about Scite, a tool that helps researchers make sense of the literature by differentiating between supporting and contrasting citations of a paper, versus just surfacing the raw citation count. We discussed how alternative metrics like Site can create a more positive, constructive optimization game.  We also talked about the culture of innovation in the scientific ecosystem that is supposed to be the cradle of new ideas: does the ecosystem really encourage innovative, new, maybe even controversial ideas? And what’s it like trying to change the legacy system as a new startup? Sign up for upcoming seminars: descifoundation.org/seminar Support the DeSci Foundation with a donation: descifoundation.org/donate Explore more Future of Science content: linktr.ee/futureofscience

Juan Benet started multiple highly influential projects in the crypto world, like IPFS and Filecoin. He started his company Protocol Labs to build out further technologies to aid the preservation, spread, and usage of information. In this episode, we talked about the grand ideas behind the technologies Juan and his network have been building over the past decade. Some of the questions we discussed included: How do you store information over long periods of time, like really long as in 10s of thousands, or even millions of years? How much information do we need in which format in order to be able to jumpstart civilization after a collapse? And how can we transfer information in a multi-planetary society? Lastly, we also learned how many hours of sleep Juan needs and, importantly, he told us – without missing a beat – what his spirit animal is. This podcast is a follow-up conversation to Juan's Future of Science Seminar. You can watch the recording here. Sign up to future seminars here. Support the DeSci Foundation by making a donation here.

James Boyd co-founded the Wolfram Institute together with Stephen Wolfram in order to advance physics, mathematics, and other scientific disciplines through the strategic use of computational technology. In this podcast, Philipp and Carla dig a little deeper into James' views on a future in which science is done automatically. What will that look like? What's the role of humans in this future? Oh, and what in the world is a cyborg scientist?! They discuss transhumanism, personal growth, and spirit animals. This conversation is a follow-up to James' Future of Science Seminar. Watch the recording here. Sign up to upcoming Future of Science Seminars here. Support the DeSci Foundation by donating towards the acceleration of scientific progress here.