It’s a big day guys, we’re abandoning the royal We. After a year of weekly posts, we’re going to decrease the frequency. To mark the milestone, we thought we’d reflect separately on the past year: why we started, why we’re slowing, what we learned, our favorite posts, what’s happening now, and the future. We’ll still post here under our own names, and we’re still excited to edit and publish posts from you all, they just won’t come at the same weekly cadence you’ve come to expect.
Dan
Why we started
Reinvent Science began as a jam session in the Lewis Science Library at Princeton University. We met up for our first extended conversation, imagining that we’d write one piece together, and ended up with a document full of dozens of essay ideas about how science could be better.
Ben and I both have day jobs at non-profits that help scientists have more impact on the world. That gives us a rare perspective on science as an enterprise. Though we see large-scale structural problems, we also see opportunities for the choices of individual scientists to have an outsized impact. We bonded over a shared and visceral belief that science could be a lot better, not just incrementally improved. We also didn’t know ahead of time how much interest there was from the greater scientific community to read and discuss topics like this and wanted to find out.
Why we’re stopping
There was interest, but not that much interest. Reinvent Science has a bit over 400 subscribers and each post gets about that many views. Many of you are practicing scientists or science funders, and knowing that is what has kept us writing as long as we have.
Our recent post about the scientific literature and LLMs went a little viral, but seemingly because it made some folks angry and I don’t want to get anywhere near the “make people angry to get traffic” spiral. Also it turns out writing the substack is the fun part but you also need to promote the substack and neither of us wanted to do that. Our day jobs include being public faces of our respective organizations and so neither of us could go all-in on the Reinvent Science brand. As a result our day-job audiences remain much larger than the audience for Reinvent Science.
Another aspect is that we’ve said what we wanted to say ourselves. A post per week for over a year turns out to be a lot. We’re not out of ideas, but we are a bit tired of listening to ourselves talk. However, we remain excited to publish content from you and other folks in science to continue to foster discussions about what science could be and how we get there. Nelson Ndahiro’s field report reached a substantially larger audience than our typical posts, which strikes me as a positive sign.
What we learned
Science can be better in a lot of ways. It was never that difficult to come up with a topic to write about. It’s also the case that it’s actually quite hard to get individual scientists to do anything about the situation. Those in privileged positions are happy with the status quo and those who aren’t are under enough pressure that iconoclasm doesn’t seem very attractive.
We also became frustrated with big-picture metascience because of how little credit it gives to the agency of individual scientists. We continue to believe that small tweaks to unblock individuals could do as much for the world as big policy interventions.
We also learned that we share the video game Sid Meier’s Alpha Centauri’s as a formative text for our scientific worldviews.
My favorite posts
Looking back at the archive I’m proud of all of it. I hope you’ll scroll through and read some of our older pieces.
What’s happening now
We’re going to stop publishing weekly, coming back occasionally as individuals when we have something appropriate for this audience and platform (I’ve got something queued up for next week). We also continue to be open for submissions from you and will be glad to edit, publish, and promote them.
The Future
We’re still working on making science better (more than) full time. Come find us here or at our day jobs if you want to help.
Ben
Why we started
At the end of the day we started Reinvent Science to catalogue a whole bunch of ideas about how science could be more awesome written by scientists for scientists.
Unpacking that: it really feels like science is going to change in the 21st century. A thousand problematic threads in science are coming to a head: the massive bureaucracies scientists need to slog through to even do their work, institutional stagnation, becoming a political football. Since we started writing this, AI has added a big piece: creating unprecedented abilities to both comb through information and unleash thoughtless slop, potentially automating labs, and raising broader philosophical questions like “does it count as science if it’s entirely done by a computer?”
It’s a microcosm of the broader “polycrisis.” But as-it-would-be-great-if-Sun-Tsu-said: “Chaos is a ladder.”
But most of the discourse about changing science revolves about some combination of abstract philosophical ideas and policy: “Is Stoke’s quadrant model better than Bush’s Linear Model?” “Should the NSF support more fast grants? The NIH more prize competitions?”
These are important questions, but they largely ignore three critical ideas:
The possibility that the day-to-day mechanics of how we do science have been and could be drastically different.
Scientists themselves are going to be major drivers of change in how science works; they’re not just agents blindly responding to external incentives.
It’s impossible to know a-priori the “right” things that scientists could be doing differently to create more awesome science, abundance, flourishing, rainbows, ponies, etc.
Given those ideas, we thought it might be productive (and fun) to put a whole bunch of provocations, rather than fully-fleshed-out-plans, of how science could be thought about and done differently by actual practicing scientists out into the world.
Why we’re stopping
Honestly, it’s a combination of being very busy and idea-exhausted. Dan got a promotion and now leads Activate’s entire fellowship program; Ben is responsible for an entire organization that is default dead. We came into the project with a massive idea backlog, but it’s actually hard to come up with a genuinely good new idea about science every other week on demand when most of your cognitive juice is being deployed elsewhere.
What we learned
“Short-form provocations about how to bottom-up change science” is niche, but in that niche people care. Only a few hundred people subscribed to the Substack, but the ratio of comments (especially good comments) to readers was way higher than most publications; and just glancing at the list of subscribers, there are some serious and awesome people on the list.
Making up plausible new ways of doing things is shockingly hard. The way things are done now creates such a strong basin of attraction of thinking that the vast majority of ideas are either totally absurd or just tinkering around the edges.
My favorite posts
What’s happening now
Two narratives that were definitely around when we started have exploded into the zeitgeist the past ~year: how science is done as a major political question and AI in science.
(If Google Ngram didn’t just do books and went through 2026, there would be a graph here of the mentions of AI for science exploding.)
In some ways, these trends make what we’ve done here more important than ever. New ways of thinking about and doing science are becoming essential rather than nice-to-have alternatives to muddling along.
On the other hand, the shrill urgency of change means that there are now many other voices and minds seriously tackling the question of how science could be different. A big motivator for us writing was just to get people thinking “oh, things could be different” rather than the belief that we have the best ideas. So to some extent we can now stand on the aircraft carrier and wave a “mission accomplished” banner.
The Future
There is still so much work to be done!
We’ll still occasionally post ideas as they come to us but will not commit to the once-a-week cadence that we’ve kept up for the past year.
Thank you for giving me much food for thought over the past year. The lovely bitesize nature of these posts also means it is the only Substack whose emails I read straight away!