Scott what do you think is going to be a successor to the scientific paper/academic journal? What do you think of platforms like distill.pub and Fermat's Library and the impact they might have?
I'd seen Fermat's Library but haven't used it. I hadn't heard of distill.pub before your comment.
Honestly, I don't see scientific papers going away anytime soon. What's the alternative to an individual or group setting out clearly in writing what they discovered, the evidence that it's true, and the background and context to the discovery, putting the writeup where interested people can easily find it, and then taking full responsibility for it, so others can cite and build on the work? As far as I'm concerned, that's all a "scientific paper" means---regardless of whether it appears in a prestigious journal or just the arXiv or somewhere else on the web, and whether it otherwise follows normal academic writing conventions or flouts them. Even an answer on a website like MathOverflow, were it sufficiently well-written and sourced, could as far as I'm concerned be cited as a research paper and given academic credit.
But while "papers," for some definition, are probably here to stay, I'm optimistic about deep reform to the system of journals. The first step, of course, should be for academics to shake off the yoke of the truly brazen predatory publishers like Elsevier. Many of us have already pledged never again to review for or submit to those publishers, at least until they fundamentally change their practices. For more about this, see my review of "The Access Principle": https://www.scottaaronson.com/writings/journal.html
As the next step, we should break free even of society journals, insofar as they put research results behind a paywall. Everything---certainly if it was funded by the taxpayers---should be freely accessible on the web. (In math, CS, and physics, we already put essentially all our papers on the arXiv, where they're freely available, and have been doing that for ~25 years. But the fact that the paywalls even exist still rankles---and other fields, like biology, have yet to catch up.)
Ultimately, we might converge on the model of journals as "arXiv overlays": that is, stamps of approval that particular arXiv preprints have been peer-reviewed (which is pretty much the only service that journals now provide anyway). Or maybe we'll even handle peer review in some other way entirely. E.g., people keep trying to experiment with peer reviews being public, but it keeps failing---possibly because, lo and behold, most academics don't want their frank commentary on the importance of each other's work to be made public with their names on it! :-)
As it happens, my friend and colleague Michael Nielsen used to work in quantum computing, but now spends full time at Y Combinator thinking about the future of scientific communication. He wrote a book that had lots of interesting ideas for how we could improve things, and surely there are many other good ideas waiting to be proposed, possibly taking advantage of recent or near-future technologies. While the concept of a "paper" (or treatise, or monograph, or note, or other unit of written research) strikes me as mostly determined by the nature of science itself, as far as I'm concerned almost everything else is up for grabs.
Honestly, I don't see scientific papers going away anytime soon. What's the alternative to an individual or group setting out clearly in writing what they discovered, the evidence that it's true, and the background and context to the discovery, putting the writeup where interested people can easily find it, and then taking full responsibility for it, so others can cite and build on the work? As far as I'm concerned, that's all a "scientific paper" means---regardless of whether it appears in a prestigious journal or just the arXiv or somewhere else on the web, and whether it otherwise follows normal academic writing conventions or flouts them. Even an answer on a website like MathOverflow, were it sufficiently well-written and sourced, could as far as I'm concerned be cited as a research paper and given academic credit.
But while "papers," for some definition, are probably here to stay, I'm optimistic about deep reform to the system of journals. The first step, of course, should be for academics to shake off the yoke of the truly brazen predatory publishers like Elsevier. Many of us have already pledged never again to review for or submit to those publishers, at least until they fundamentally change their practices. For more about this, see my review of "The Access Principle": https://www.scottaaronson.com/writings/journal.html
As the next step, we should break free even of society journals, insofar as they put research results behind a paywall. Everything---certainly if it was funded by the taxpayers---should be freely accessible on the web. (In math, CS, and physics, we already put essentially all our papers on the arXiv, where they're freely available, and have been doing that for ~25 years. But the fact that the paywalls even exist still rankles---and other fields, like biology, have yet to catch up.)
Ultimately, we might converge on the model of journals as "arXiv overlays": that is, stamps of approval that particular arXiv preprints have been peer-reviewed (which is pretty much the only service that journals now provide anyway). Or maybe we'll even handle peer review in some other way entirely. E.g., people keep trying to experiment with peer reviews being public, but it keeps failing---possibly because, lo and behold, most academics don't want their frank commentary on the importance of each other's work to be made public with their names on it! :-)
As it happens, my friend and colleague Michael Nielsen used to work in quantum computing, but now spends full time at Y Combinator thinking about the future of scientific communication. He wrote a book that had lots of interesting ideas for how we could improve things, and surely there are many other good ideas waiting to be proposed, possibly taking advantage of recent or near-future technologies. While the concept of a "paper" (or treatise, or monograph, or note, or other unit of written research) strikes me as mostly determined by the nature of science itself, as far as I'm concerned almost everything else is up for grabs.