Rachel Coldicutt

[in science funding] “almost anyone can block things … stop things … but no one can get anything done"

Graham Stringer saying there is less freedom for ARIA in the draft bill than would be had in a university department. "And there is no mission statement for ARIA, one of the strengths of DARPA/ARPA when it was set up was that it had the mission to make America a safer place."

Stringer saying the bill is a bit woolly (I paraphrase) and notes that there is nothing about defense.

Cummings offers his take on the ARIA mission: “To accelerate scientific discovery far beyond what is strategically normal, and to seek advantage in some fields in the UK. I would keep it broad and vague like that."

Cummings saying that the move from ARPA to DARPA was retrograde step, and that he wouldn’t give it a defense focus and “I certainly wouldn’t try and import a bunch of buzzwords like AI and quantum and net zero and things like that … I would stick to a broad mission"

Cummings saying he would not commit to a single discipline in the same way but leave the ARIA team to spend time "problem finding … things like creating the Internet … are not obvious … it will take time to find the often very odd people who are pursuing them"

Katherine Fletcher saying it sounds like a skunkworks, but “what do you actually want them to *do*? You keep saying ‘go and sample’ but are we sampling in highly theoretical aspects … or long-term goals, or are you just boiling the ocean with 800m quid and hoping for the best?"

DC: “I don’t think views from me on what sort of thing it should be looking at have any real value, it should be going off to people on the far edges of the science world.” Mentions Licklider and the intergalactic communications network. “Those things are hidden."

Katherine Fletcher: “What I’m bothered about is the fallability of humans in this system. If you’ve got a tiny leadership team of a director and 4 trustees, you’re effectively looking for 5 Sir Patrick Vallances…"

"how are you going to stop a small number of people getting the wool pulled over their eyes?”

You’ve got to pick the 5 people very carefully … a very flat organisational structure and they have good taste in finding people, and there’s no alternative to this fundamental problem. You have to have someone in charge who has good taste in scientific ideas and … researchers"

Cummings now listing the historic precedents — essentially, some white men who have had great tastes and hired some other white men — and says there is “no alternative”. Katherine Fletcher picks this up. “All the people you’ve mentioned are blokes."

Fletcher: “Taste is one of those subjective things we need to be very careful about when concentrating power in a small number of hands”

Fletcher asking Cummings what he thinks about collaboration. Cummings saying academia is now overly focussed on individual and papers, "which is why Demis Hassabis had to go to California and talk to people like Peter Thiel to get DeepMind going…"

“… because the British funding system does not think in that way. [ARIA should be able to say] here is a Licklider-type person, have £10m quid, knock yourself out… I don’t think it should be limited to that"

Cummings saying missions and moonshots are not the only way of doing things. Says the Licklider vision was not a moonshot but more ephemeral than that “and much more powerful for being that”. (Brett Victor klaxon. Cummings loves Brett Victor.)

Dawn Butler now asking who should lead ARIA and why. DC offering 3 people who he would think of: he offers up Michael Nielsen, Steve Zou (sp?), Tim Gowers.

Oh he’s remembered a woman. “A brilliant physicist called Chiara.”

Cummings say he won’t be involved and isn’t seeking to be. He says, “If you pick the right people, what could I possibly contribute to it?"

Dawn Butler asking whether Cummings considers eugenics to be a science. Cummings is wary, says he’s not sure what she means, and that Britain has a great advantage in genomics and “strongly thinks that that fundamental scientfic research should be funded here”.

DC: “If you’re going to put it inside UKRI then don’t do it at all” and says that Ottoline Leyser agrees with him.

Aaron Bell now saying that Manhattan Project etc stem from “existential crises” and asking, “Without a crisis, will ARIA work?” DC saying “extreme freedom” is often only granted in response to existential crises.

“One of the most obvious lessons of last year… is that it makes the case for creating an institution like ARIA.” He’s now talking about infrastructure to deal with pandemics and that should be the impetus for creating something like ARPA. The logic here not v clearly articulated

Oh hang on, I *think* he means that there needs to be an horizon scanning function with long-range thinking capacity, as well as fast-paced procurement and problem solving for emergency situations. Not entirely clear though.

“It was interesting when I was talking to some of the old timers who were funded by Licklider and Robert Bell… [they said] it seems like great funders are much rarer than Nobel prize winners … it seems like ppl like General Groves … are very very odd people …"

“if you just get some bog standard Vice Chancellor then it won’t work”

On misfits and wierdos: “2020 was proof that if you don’t have people from scientific and tech backgrounds who understand how to think rationally and quantatively about extreme uncertainly, you have disastrous decisions"

Says that the “strongest supporters for this are not very high status people … if you had Newton or Darwin or Turing turning up with their ideas now, aged 21, everyone in the fund system would say ‘you’re mad, of course we’re not fnding you”.

Carol Monaghan saying Newton’s ideas were sneered at when he first presented them. Saying there must be checks and balances, “how do we avoid extreme freedom leading to extreme cronyism?"