Suddenly There’s Just So Much New Science to Do
Something remarkable has happened these past two years. For 45 years I’ve devoted myself to building a taller and taller tower of science and technology—which along the way has delivered many outputs of which I’m quite proud. But starting in 2020 with the unexpected breakthroughs of our Wolfram Physics Project we’ve jumped to a whole new level. And suddenly—yes, building on our multi-decade tower—it seems as if we’ve found a new paradigm that’s incredibly powerful, and that’s going to let us tackle an almost absurd range of longstanding questions in all sorts of areas of science.
Developing a fundamental theory of physics is certainly an ambitious place to start, and I’m happy to say that things seem to be going quite excellently there, not least in providing new foundations for many existing results and initiatives in physics. But the amazing (and to me very unexpected) thing is that we can take our new paradigm and also apply it to a huge range of other areas. Just a couple of weeks ago I published a 250-page treatise about its application to the “physicalization of metamathematics”—and to providing a very new view of the foundations of mathematics (with implications both for the question of what mathematics really is, and for the practical long-term future of mathematics).
In a sense, everything we’re doing ultimately builds on the great intellectual tide of the past century: the rise of the concept of computation. (And, yes, that’s something in which I’ve been deeply involved both scientifically and technologically throughout my career.) But what’s happening now is something else—that one can see as the birth of what I call the multicomputational paradigm. It’s all about doing what our Physics Project has suggested, and going beyond working with specific computations—to look at the systemic behavior of whole interacting collections of computations. In the whole multi-millennium history of science, there’ve only been a very few fundamentally different paradigms for making models of things—and I think multicomputation is basically the fourth one ever.
And what makes its arrival particularly dramatic is that it comes already supercharged by its deep relation to physics, and by its ability to spread the successes of physics to other areas of science, and beyond. When I started investigating the concept of computational irreducibility in the 1980s it became clear that there are fundamental barriers to many kinds of scientific progress. But what I didn’t see coming is there would be a new path opened up by a new paradigm: the paradigm of multicomputation. And suddenly there are now all sorts of fundamental questions that are no longer blocked—and instead are ripe for rapid progress.
Over the past year we’ve started exploring a host of potential application areas. We’ve got a concept of “formalized subchemistry” with applications to a new way of thinking about molecular computing, and with potentially dramatic implications for molecular biology. We’ve got new ideas about how to think about immunology and probably also neuroscience. We’ve got a potential new approach to finding a formalization for biological evolution and its relation to biocomplexity. We’ve got a new concept for “geometrizing” the space of programs—with implications for foundational questions in computational complexity theory. We’ve got a promising way to construct new kinds of theories for economics, with implications, for example, for distributed generalizations of blockchain. We’ve got a potential new way to think about linguistics, and the structure of meaning space. Oh, and we’ve got a new “physicalized” way to conceptualize and organize distributed computing.
In the nearly half a century that I’ve been doing science, I’ve had the good fortune to be involved in quite a few significant bursts of progress. But I’ve never seen one quite as concentrated and immediate as what we’re now seeing. It’s an exciting thing. But it’s also overwhelming. There’s just so much low-hanging fruit to be picked—so many things with such important potential consequences, for science, technology and the world.
We’ve been working very hard to move all this forward with the resources we have available. But even though I think we’ve achieved a remarkable level of productivity, it’s become clear that there’s just too much to do. We’re in the midst of a major “science opportunity overload”. And to be good stewards of the ideas and their potential we’ve got to scale things up. I’ve had lots of experience over decades in making big projects happen. And now it’s time to take that experience and define a new structure to move forward the amazing science opportunity we find ourselves with.
And I think that leaves us no choice: we’ve got to launch the Wolfram Institute, and now!
Maximizing Scientific Productivity & Communication
In the course of my life I’ve spent a great deal of effort trying to maximize productivity and innovation around the things I do. I’ve done lots of (arguably nerdy) optimization of my own personal setup. I’ve thought long and hard about the best strategies both for choosing what to do, and for getting things done. But in many ways the most important piece has been the whole structure we’ve built up over the past 35 years at Wolfram Research.
I sometimes refer to our company as a machine for turning ideas into real things. And I think it’s been an extremely impressive machine—year after year for more than three decades systematically delivering all sorts of far-reaching innovation, and using our progressive tower of technology to efficiently implement it on larger and larger scales.
As a company, we’ve mainly been concerned with delivering technology and products. But the systems, culture and methodologies we’ve developed are, at their core, about maximizing innovation and productivity. So what happens if we apply them to science?
A New Kind of Science was in a sense a first result, and I consider it an impressive one. Indeed, even nearly 20 years after it was published I’m still amazed at the sheer volume—and depth—of scientific results that it was possible to obtain in the span of just a single decade.
But with the Wolfram Physics Project it’s a yet more impressive story. We started late in 2019, and in less than six months—with an extremely small team—we were able to make dramatic progress, and to publish nearly 700 pages of material about it. It wouldn’t have been even vaguely possible without the whole tower of computational tooling provided by the Wolfram Language. And nor would it have been possible without the structure and strategy for doing projects that we’ve honed over the past three decades.
In the past two years, we’ve been energetically moving forward—and we’ve now published altogether over 2500 pages of new material, as well as nearly 200 publicly deployed new functions. In terms of scientific ideas the pace of innovation has been quite breathtaking. But we’ve also been innovating in terms of how to do the science.
An important objective has been to open the science up to give the widest possible access and potential for engagement. I’ve worked hard to define a style of expository writing that makes what we’ve done accessible to a wide audience as well as to experts. And in what we’ve published, essentially every graphic has “click-to-copy” Wolfram Language code, that anyone can immediately run and build on. We’ve also uploaded our working notebooks—so far nearly 2000 of them—so that everyone can see not only our “finished product” but also the research (wrong turns and all) that led to it.
A few years ago I started livestreaming to the world many of our software design reviews. And building on this concept, we’ve now routinely been livestreaming our scientific working sessions—giving people for the first time real-time visibility into how science is done, as well as the possibility to interact with it. And for those interested in an even deeper dive, we’ve also been recording and uploading “video work logs”—bringing us up to a total of nearly 1000 hours of video so far.
Even when we thought we were “just solving physics” we knew we had to involve other people in the project. For 20 years we’ve been doing a very successful annual Summer School about our approach to science and technology, and starting in 2020 we added a track about the Physics Project, as well as a physics Winter School. We’ve had a terrific stream of “students” for our Physics Project. And partly building on this we’ve been setting up a network of people involved in the Physics Project—now with 55 members from 20 countries.
Increasingly, there’s work based on our Physics Project that’s happening in academic institutions, quite independent of us. And no doubt this will bear all sorts of fruit.
But as we look at the next phases of our Physics Project, and even more so the huge collection of opportunities provided by our new multicomputational paradigm, it’s clear there’s so much to do that—particularly if we want it to happen in years rather than decades—we need a more focused approach.
And the good news is that through 35 years of experience at Wolfram Research, as well as the experience of A New Kind of Science and the Wolfram Physics Project, we have an excellent blueprint for what to do. But now we have to implement it at scale. And that’s what the Wolfram Institute is about.
The Plan for the Wolfram Institute
The basic plan is simple: to create a basic-science analog of the immensely productive “machine” that I’ve built at Wolfram Research over the past 35 years—and to use this “machine” to accelerate the delivery of new science by many decades if not more. We’ve already got a definite seed: the Wolfram Physics Project. But now we have to scale this up to the full Wolfram Institute—and give it the structure it needs to grow and take full advantage of the amazing opportunities we now have.
It’s often assumed that the way to achieve maximum innovation is to put together innovative people, and then let them “just innovate” in whatever directions they choose. But it has been my consistent experience that the greatest innovation is instead achieved when there is a definite “flow” and definite, ambitious goals. The Wolfram Institute is going to be about doing large-scale basic research this way.
Modeled on Wolfram Research and the Wolfram Physics Project, the idea of the Wolfram Institute is to aggressively pursue basic research that’s explicitly managed and energetically led towards its goals.
Our initial goals are already tremendously ambitious. We want to use our new paradigm to basically rewrite the foundations of several important fields of science. As with the Physics Project there’ll no doubt be tremendous synergy with existing approaches and the communities around them. But with the paradigm we now have, the tools and methods we’ve developed, and the organizational framework we’re defining for the Wolfram Institute, I think we have the opportunity to jump ahead, and in effect to deliver foundational science that would otherwise emerge at best only in the distant future.
It’s going to be an exciting thing to be a part of, and—as with projects I’ve done in the past—there are going to be many outstanding people who want to be involved. Our many decades of activity in science and technology have provided an extremely broad network of contacts, and we’ve developed a particularly concentrated pipeline of worldwide talent through our annual Summer School (as well as our High School Summer Camp).
What will the Wolfram Institute be like? There’ll be a leadership core, which, yes, I’m signing up to head. But the main meat of the institute will be a collection of researchers and fellows, working on particular, managed projects—together with students at multiple levels from high school to graduate school. We’ll be doing open science, so there’ll be lots of livestreaming and lots of open tools produced. There’ll be lots of working materials and academic papers published, and whenever we manage to make a big step forward our plan is to present it to the world using the immediate and accessible approach to exposition that I’ve developed.
In what we’ve done for the past 35 years at Wolfram Research there’s in a sense a clear model for how the organization fundamentally operates. We invent things, then we deliver them in products—and from these products we derive revenue, which is then used to allow us to invent more things and deliver more products. It’s been an extremely productive setup. And because we’re a private company without outside investment we’re able to chart our own course, pursuing the ambitious and long-term projects that we believe in. Often we choose to make things we do freely available to the world, but in the end we rely on the fact that we’re producing commercially valuable products from which we derive revenue that funds our activities.
There’s no question that the science we’ll be doing at the Wolfram Institute will lead to things of great value to the world. But it won’t be near-term commercial value. And our fundamental model is to concentrate on producing the best and broadest long-term basic research—rather than to aim for things that can be deployed directly in specific products that provide immediate value to specific customers.
In a sense, our “customer” is the world at large—and the future. But in the present we need a way to support our researchers and fellows. So far we’ve basically been incubating the Wolfram Institute within our existing organization, with me effectively footing the bill. But as we launch the full Wolfram Institute we need a larger scale of support, and we’re counting on having a network of people and organizations to provide that.
Some of the support will be for specific researchers, fellows or students, perhaps drawn from particular geographies, backgrounds or communities. Some of it will be for specific projects. But it’s also important to have a stable core of support that will allow the institute to pursue long-term basic research that will likely deliver many of its most valuable results through developments that are in effect computationally irreducible to predict in advance.
It’s difficult to know how society at large should value the general activity of basic research, and it’s easy to criticize the inefficiencies of a large-scale “just let researchers do what they want” approach. But with the Wolfram Institute we have a very different model. We’re starting the institute right now for a specific reason: we’ve got a new paradigm that’s just opened up an amazing collection of possibilities. And we plan to pursue those possibilities in an efficient and tightly managed way, optimized for innovation and new ideas.
When I look at our Physics Project and what we’ve achieved so far with it, I’m frankly amazed at how quickly and comparatively frugally we’ve managed to do it. (Yes, it helps that at least so far we need only people and computers, not telescopes and particle accelerators.) And as we scale this up to all the various projects we plan at the Wolfram Institute, it’s almost absurd how much of long-term significance I expect we will be able to deliver for how comparatively little.
The Wolfram Physics Project has been done as an entirely geodistributed project—building on the 30+ years of experience with coherent geodistributed work that we’ve had at Wolfram Research. And even though its projects will be tightly managed, the Wolfram Institute will also primarily be geodistributed, although we plan regular physical events and we’ll probably have some physical locations available.
It’s been great in working on the Wolfram Physics Project to have what we’ve been doing be so open, and to be able to share it with so many people. And as we launch the Wolfram Institute I’m looking forward to having all sorts of people involved, both within the institute, and as supporters of it.
It’s been exciting these past many months seeing the whole multicomputational paradigm emerge, and seeing more and more possibilities reveal themselves. It’s a remarkable—if overwhelming—collection of opportunities, and I believe a historic moment for the progress of science. And as we launch the Wolfram Institute I hope that with the help of enough supporters we’ll be able to deliver many dramatic results that will have great long-term value to the world and to the arc of intellectual history.