S&T - Bret Victor, Realtalk and the ‘laboratory of the future’ vision
‘Communal computing and the 21st century laboratory’
The image above is from a talk by Bret Victor that Alan Kay sent me this afternoon.
It’s a visualisation of the laboratory of the future by Bret Victor, in which every component of the lab is a computer that you interact with communally. Out are people sitting behind screens. In is using the physical space as a computational system. Everyone in the illustration is using the same integrated computational device. The whole room is a computer.
This short blog is to a) promote the video of the remarkable vision that Victor outlines and b) provide a brief intro to it for people not familiar with it. I hope more people interested in metascience take notice of Bret Victor’s ideas.
Surprisingly few people know the history of how personal computing happened. Alan Kay was a key figure in it as part of the legendary ARPA-PARC team. Their 1960s/70s vision of computing is in some ways very different to the one that has arisen.
Computers were to be ‘intellectual amplifiers’, as Licklider described them in his 1960 ‘man-computer symbiosis’ (one of the all time great vision documents). As Alan has noted, too often the amplifier has been turned down, not up. Alan points repeatedly to Bret Victor’s work as though he is an heir to their visionary project.
Many (including our #10 science team) have tried to get Victor funding, but the kind of long-term, risk-tolerant funding is very hard to find. Recently it was described to me dismissively by a potential funder as ‘Xerox PARC nostalgia’ without a clear capability at the end of it (a little like asking PARC in the 1960s what computers would be used for). As Alan says repeatedly, the kind of funding that made possible personal computing, with a strong component of ‘problem finding’ and community building and high risk tolerance, is basically absent today. If you want to read some Alan Kay emails on this critical issue, Victor posted some on his website and I really recommend it.
What if the funding style that led to personal computing through ARPA-PARC had continued? As I said above, Alan Kay points to Bret Victor as continuing the vision they were trying to do.
Trying to explain Bret Victor’s vision is tricky in words, like trying to explain a personal computer to someone in 1960. People who have used Victor’s tools say ‘this is the future, why isn’t this funded’?
Realtalk is the name of the system. In realtalk, anything in the physical space can become a computational system - ‘the entire building is a computer’. Cameras and projectors turn the physical space into a computational system. You can alter a piece of paper, and have it in turn alter what is being projected in the physical space. People can work collaboratively using real physical objects, not via screens, and have projectors output the result of the computation. You can see this here at https://dynamicland.org, and videos on their twitter. Unfortunately Elon Musk has decided you can’t embed twitter videos in substack anymore. A couple of example videos are here and here.
In this example the physical space on the right is transformed into a ‘colleidoscope’ projection:
In this example an individual is programming a series of projections on the wall based on the actions they make on the floor:
I can’t do it justice in a small blog - here is a dedicated explanation, and the video below will make it clearer.
How could this communal computing be used in the lab? How could you reinvent a laboratory so that it didn’t have screens, but that everything was a combined computational device? A friend has a research program looking at the intersection of AI and biology. I can’t say much more as it isn’t announced. I got thinking about how Bret Victor’s ideas could be used in this context, and I emailed Alan Kay about it. Alan sent me this video, released just a few days ago. It’s Bret Victor applying the ‘physical world as a computer’ vision to the vision of reinventing how laboratories work.
Here, for example, is a snapshot of protein engineering using RealTalk (from midway through the talk):
And the remarkable 10 minute talk:
Also, when I took a look at the vision of the lab of the future I noticed that everyone is in an open office.
I’m convinced that the lab of the future will actually return to closed offices. Or maybe even better to spaces that can be configured to the psychological make up of the individuals in them (Bell Labs, at least the 1950s version, had reconfigurable rooms).
I’m convinced that scientific productivity is hampered (and the mental health crisis in science is due to) by at least three factors:
1. Too much career uncertainty/anxiety. You put your finger on this in the design of the Lovelace Institutes.
2. Computers: probably not a negative, but more like a net neutral or only slight positive in terms of productivity. You point to this here.
3. I think a big unacknowledged one is/are open plan labs. Especially ones where you do experiments and sit all day. They suck for thinking. Especially for introverts. My ability to get stuff done was in the toilet for 5 years when in a lab like this, and was doing the worst work of my career. It immediately rebounded, and I started doing some of the best, when I got a closed office.
My lab is designed so that we have separate small office (library quiet) and lab space (smelly, noisy, fun)1. I definitely think this has contributed to the quality of our work.
> In is using the physical space as a computational system.
Nice.
Thanks for sharing this, I've been looking forward to updates for a while. Bret Victor's stuff is very cool and it's quite strange to me how few people seem to be aware of these ideas.