For this to work you need the public to understand why DARPA worked. Hiding details of failures makes that impossible (and again, completely open to corruption).
https://www.theweek.co.uk/951989/what-next-for-dominic-cummi...
Perhaps it hit a raw nerve or two.
This might be controversial, but it's very hard to have negotiations in good-faith with a company for services, and drive a hard bargain, if the company knows that the pricing offered to government can be FOI'd by a rival or other customer, to get granular price offered.
If you want to get the best value for money, and get below list price, you need the ability to have a commercial negotiation, with the confidence that granular pricing information (i.e. emails with discounted price lists for government customers) aren't becoming public.
That's not to say the total amount spent should be kept secret, but if exact breakdowns of unit pricing were going to be made public, it would likely cost the public more in the inability of government to negotiate around price with suppliers (or rather their unwillingness to enter into such negotiations)
Similarly, any kind of serious negotiation needs to have secrecy - it's very hard in a practical sense to have a negotiation with a party that has to (or might be forced to) publish everything. The number of startups (and even larger companies) that do everything as price-on-request should show industry's willingness to see the kind of price tarnsparency that FOI would expose. And that would give the taxpayer poorer value for money in the long term.
https://benjaminreinhardt.com/wddw
> The Advanced Research Projects Agency model is of an organization set up to maximize the agency and effectiveness of world-class program managers (PMs) who coordinate external research to midwife technology that wouldn’t otherwise happen (Programs.)
> The model has changed over time, but has still produced outlier results over time so it is worth paying attention to modern DARPA with more focus on informal process than formal process. PMs need specific characteristics to succeed: thinking for themselves, curiosity, low ego, vision, and ability to act under uncertainty. PMs also need to be trustworthy because the model depends on their ability to deploy funds quickly and redeploy them as needed. These PMs have temporary tenures, which enables idea turnover, aligns incentives, and enables DARPA to hire people it wouldn’t otherwise be able to. It’s worth deeply thinking about PM motivations because they are so core to the model.
> Organizationally, DARPA is tiny, flat, and opaque. It is set up to combine bottom-up and top down approaches through different-scale feedback loops. It is more ideas limited than money limited. DARPA’s project design and execution framework boils down to first showing that a precise technological vision is not impossible, then showing that it is possible, and finally making it possible. On top of many tacit tools, PMs execute on these steps by building focused networks and using seedling projects to derisk assumptions during a <$1.5m exploratory tranche before presenting a program design to an advisory group to the director known as the tech council.
> DARPA provides a critical ‘in-between’ role in the world. It facilitates cross-polination and derisks wacky ideas for both private, academic, and government organizations.
- The internet
- GPS
- Graphical user interface and mouse
- Onion routing
- Voice assistant
Hopefully this program executes well. There is clearly a big social payoff to betting big on credible people doing risky research with high potential impact.
https://www.itpro.co.uk/technology/34730/10-amazing-darpa-in...
What makes an (D)ARPA? What (besides funding) made it work?
Is secrecy and/or low accountability important? Defense/weapons focus? Was the cold war a necessary condition? Do they have an investment philosophy that could be copied? Managerial philosophy?
Is high level stuff even relevant or is it details like 5 year PM appointments and selection criteria? I always thought tenure-like jobs would be useful if you have high creativity/risk goals.
Thoughts? Any agencies (outside the US, also) that should also be considered shining example? Any failed attempts at creating a DARPA?
A couple things I'd pull out:
Program managers sit limited terms to prevent empire building
Program managers have a great deal of autonomy once funding is initially allocated
DARPA has a high appetite for risk. They're ok with 90% of projects failing to hit their goal.
DARPA research is performed in response to real challenges that service members and officers are faced with, and also in response to new capabilities that our adversaries develop.
I mean, just look at how they’ve fared in the last 18 months alone - how many unaccountable tens of billions did they pay out to their donors?
I had Tory ministers’ children approaching me about putting together business plans to help them get covid grants - same crowd who I used to help with R&D grants, which were also used to get high net worth individuals with money of dubious origin into the country. I anticipate I’ll be hearing from them again.
This kind of funding will not develop anything but the wealth divide.
https://metro.co.uk/2019/10/31/10000-grant-given-boris-johns...
As described, the risks I see are that, it has the incentives for a patronage slush fund, there will be selection bias to hedge political risk around people applying, and it will just become another grant vehicle for academics in the publish-or-perish regime, and instead of "making something people want," in the venture model, it means, "find things nobody else understands well enough to care about." To people outside of it, it will look a lot like corruption.
Government can invest in projects which have different types of return: projects which are only beneficial or more beneficial if given away for free, projects which will take a lot longer than a decade to come to fruition, projects which are important to government social objectives but just not big enough commercial markets to be interesting to VC and projects which the government will be the only customer for so might as well cut the VC middleman out. Even on a purely financial return basis, the VC captures only captures the part of the return the portfolio company can charge for: the government collects the tax receipts from everyone who benefits in the sector including competitors who copy, and doles out resources to the entire supply chain and consumers too/
Sure, making it secret certainly increases patronage slush fund potential but it's not like that isn't already there, or like VCs are perfectly efficient investment machines who never consider their network or biases before spending LP funds, or had much role in many of last century's significant inventions.
I have done a few “hold your nose” consultancy gigs for an outfit that specialises in the entirely legal, if morally dubious, process, of importing the family of high net worth individuals on entrepreneurship visas - and laundering their money into the U.K.
I was approached in 2016, after I cashed out of my business, by an acquaintance from school who had spent a few years as a Tory MP, and was now using his barristers chambers to run this scheme. He was the head of chambers’ nephew, and is still considered their golden goose, because this stuff is seriously lucrative.
I’d provide a business plan, and would be the U.K. national director. They’d be the investor director from overseas. I assigned all reporting etc. over to the agency, so I was just producing business plans and providing a U.K. taxpayer’s identity.
They’d put in their money (as it’s an investment the burdens of UWOs and so-forth don’t apply, as it’s counted as a liability, not an asset), the government would match it, and they then come to Mayfair, pay themselves a salary for three years, do an annual R&D tax credit and get a nice big cheque from HMRC, and then wind the whole thing up when it “unexpectedly fails”.
I did two of these before I got cold feet. A Russian oligarch’s son, a Kazakh oligarch’s wife. Never met them, only found out their IDs after the companies were incorporated.
Whole damn thing stinks to high hell, but it’s essentially the intended purpose of the scheme.
[1] - https://www.crunchbase.com/organization/in-q-tel/recent_inve...
[2] - https://en.wikipedia.org/wiki/In-Q-Tel [note] not every company is listed
How should government fund basic vs. moonshot research? Incentives. If they think they are smart enough to allocate funds via a committee, they need to be smart enough to figure out how to structure incentives.
Venture capital funding is a great way for product research and development to be outsourced---nobody is going to notice if the project fails but if it doesn't there is money to be had. Near-basic research has an extremely long time horizon (internet research started in the 1970s and didn't become publicly useful until the 1990s) and may never be profitable. Further, nobody wants the results of a long term research project until they already have it and can see the utility.
Isn't the difference that VCs are looking for short-to-medium term opportunities as they want a return on capital? Presumably gvt funding means riskier projects can be undertaken.
If you can keep things on the level, humans can achieve a lot, there are other goals than just making financial profit.
The internet is probably a cliche example, but it came out of DARPA as a resilient control network for the military. Nobody at the time envisaged the impact it would have - it was an interesting technical piece of work that scratched an itch. Funding it resulted in a huge economic and strategic benefit for the US. It makes sense for other countries to attempt to replicate this model.
Being able to strategically fund technically interesting (but highly adventurous) moon-shots is not really attempting to out-perform VCs - the goal is to evaluate and fund a portfolio of crazy (but credible) ideas, accepting they may fail, and that they won't run according to nice quarterly deliverable schedules, as things will go wrong.
Academic funding is very broken to begin with, as it focuses so hard on reducing the risk of projects that non-delivery (i.e. negative outcomes) are frowned upon or not tolerated (depending on the field) - this is a very unhealthy set of incentives. We should allow research to fail and return negative results, and not penalise the researchers. Not everything will work. Especially for early career researchers, they might be over-optimistic about how long something will take.
Researchers are trained to be so risk-averse that they are often seeking grant funding for work they've already done, in order to reduce (to near-zero) the risk of non-delivery. They then use that grant funding to tide over doing innovative research on the side, or doing the majority of the de-risking of it, in order ot be ready to bid for future funding for that idea, keeping the conveyor belt moving. But you have to find a way to get onto the conveyor belt and get started running on it.
Hopefully this research approach will embrace failure and delayed delivery for good reason, and focus instead on what can be learned, rather than on having low-skilled box-tickers focused on managing other people's projects - the two hallmaks of current govt funded research.
The governments should not be inventors, though, I agree. Government should regulate and protect without uncalled for interventions.
Publicly funded non foi sounds like a playground for scammers and corruption.
If there was an effective way to enlist VCs while aligning goals then they would opt for it.
Space exploration is a good use case here. The US government has found a way to align its goals with that of the private space industry and is therefore seeing tremendous efficiencies. However, because of a whole bunch of reasons (largely because there was a lot of basic open ended R&D that needed to be done) this alignment was probably not possible in the 60s and NASA was more successful.
Many moonshot projects require a very large amount of capital to get up and running, and run a high risk of failing. For a typical VC firm, they will rarely choose to invest in something like that if there are other opportunities that don't require as much up front capital, and with better odds of succeeding.
When you don't have LPs and/or shareholders to answer to, you can afford to take greater risks on more ambitious projects.
Equally, the UK has a gap in research funding. The private sector has had, literally, decades to close it. The gap is closed in many other economies by the govt, so this is a reasonable solution. The govt has also done a huge amount to involve the private sector, most of the approach has been about the private sector but this is targeted on an area where it is basically understood there is market failure.
There are always risks around politics, this has happened to an extent with Scottish Enterprise and the new SIB, but this problem isn't going to solve itself, and the path to oversight is relatively straightforward. Saying that has to be corruption just makes no sense when any funding is going to be under the same constraints as most other govt funding (I have no idea why people FOI is public scrutiny, it isn't, FOI would not help you uncover corruption to any degree...that isn't what it is for or how it is used, if you believe that the govt is corrupt, defund all services, that is it...that is the only logical conclusion that can be drawn from that belief, we have FOI and it isn't stopping corruption...so defund it all).
In Government? I chuckled.
> it will just become another grant vehicle for academics in the publish-or-perish regime
I can already see the selection committee, all from non-technical backgrounds, getting clear instructions on who to fund and who to reject based on the demographics the party wants to appeal to and existing donors.
DARPA lists biographies of its newest program managers. All but one has a doctorate in a relevant field; the lone exception has a master's degree and aerospace work. They turn over every few years too, so it's not like their experience is stale.
Different structures yield different results. Venture backed startups produce results different to internal R&D departments, for example. Open source/open culture projects yield different results than typical startups.
I don't think a VC fund would have invented the internet, not in the 60s. I don't think the worldwideweb could have been invented by a typical corporation. If it had been, it would have been more like facebook than the www. America Online and others were attempts at a proprietary FB-esque www. The web would have been different if they won. Linux is different from Windows. Different types of organisations make different kinds of things. If we only have one type of organisation, we limit the possible outcomes.
The vast majority of current VC success stories are neither here nor there in terms of public benefit. The next tiktok is a great return for VCs, but consumers are not likely to lack for social networking innovations either way. I think this is a sign of overinvestment in one category. They're fighting over the pie, more often that baking it now. This wasn't the case circa 2005.
Funding structure and organisational structure are kind of similar for these purposes.
Tesla's uniqueness is another case in point. Musk funded it himself. There isn't really a type of financing, besides self financing, that could have done that. Banks and hedge funds don't want high risk projects. VCs don't want projects that require lots of funding once successful. Despite being a superstar CEO, Musk's initial investment diluted him from 100% to nearly 0 at one point. Technically, the shares that he owns were earned back as CEO performance bonus. Not attractive to a VC... compared to a facebook.
I doubt that Tesla is the only such project possible, technically. Funding structures just make it less likely.
In much of the west (UK certainly), our economies have gotten very intangible. Software, Banking, financing, entertainment, patentable tech. Some of this is because the world presents those opportunities. Some of it is because of financial structures. We have bigger/better financing structures for some things more than others.
Silicon Valley developed a unique economy based on a pretty unique web of financiers, angels and VCs.
All that said, I have no idea if this agency is good/useful. Sounds dubious, honestly.
For me the issue is they intend to invest into 'projects', why phrase it like that? If it's research, then you're finding out new information, proving theories etc, so to deem it as high risk project is strange, just treat it as research, it's allowed to fail!
Another one is the FOI. This is just stupid, we should know what the money is being spent on, what the outcomes, things learnt etc. It doesn't have to be the costings down to how much a tea bag costs. Just the top level stuff. I'm happy for top secret projects to be exempt if there is serious national security concerns. But if the public aren't allowed to see it, then who is getting the benefit? Could very well be private companies (which isn't a bad thing, but we've paid for them to gain advantage...)
https://en.wikipedia.org/wiki/Defence_Evaluation_and_Researc... https://en.wikipedia.org/wiki/Qinetiq
However the tory gov is currently busy handing out questionable contracts to friends like a pez dispenser so not super excited about the no FOI part.
Who is going to benefit from any discovered new tech? Are the benefits going to be socialised, or is it just the risk?
If it was supporting pure research - finding great researchers and freeing them from the grant treadmill - I'd be supportive, but this doesn't sound like that.
Same here. I am all for more research but we want more transparency, not less.
> Ministers will announce plans tomorrow for an £800 million scientific research agency legally entitled to invest in projects that are likely to fail and which will be exempted from freedom of information laws.
This makes no sense to me. Why would they need to be exempt from freedom of information?
> It will be exempt from rules designed to prevent taxpayers’ money from being invested in projects with little chance of success. Ministers hope that the agency, to be set up next year, will produce next-generation technology.
This I completely support. From what I understand, the COVID vaccine wouldn't be here if it wasn't for some scientists working on some seemingly dead-end path for years.
Overall, I don't understand this initiative at all. The last thing I'd want is pulling funding away form fundamental research and putting it in some Instagram clone. I'd even be OK if this money went to some kind of James Bond/Quartermaster/DARPA thing instead.
So no-one can find out where the best part of a billion pounds went?
Perhaps that is the UK's reason, besides wanting to keep projects related to national security or industry classified from foreign eyes. US government research grants often have pundits pointing out how trivial and wasteful they seem at first glance, and it sometimes leads to grants being redirected or cancelled entirely.
Leaving this specific agency/project aside, what would be the right way to do this? How do we do public investment in high risk projects well?
FWIW, from my experience in UK public projects, I think the question of how to improve their varied success is complex and would take a long time to fairly cover. Though, off the top of my head:
* Pay market rates; non-private research seems to be struggling to attract and retain talent due to the lower wages, and without meaning to be unkind to anyone in particular, there is an evaporative effect where those that leave tend to be higher achievers that can find better offers elsewhere, so the organisation becomes increasingly mired and less functional over time.
* End the mentality of "promotion" meaning "to management", which eliminates at a stroke any meaningful technical contribution from the promotee. I think this one is improving lately, but it's taking a long time to undo the damage of the 1990s–2010s.
* Have advisers for scientific programmes at the level of policymaking, and *listen to them*. It's vexing enough from the outside, but an acquaintance employed in this capacity for the previous government described the scene from inside as "a shit show".
Look at what happens when the Daily Mail and/or opposition MPs find out that an IT boffin is being paid more than a nurse. They stoke unjustified outrage until the hospital is forced to outsource to Crapita to hide its costs.
(1) DARPA (and NASA) appear to be models. Secrecy and unaccountability are a big part of what makes them them. (2) The state of university grant making is currently terrible. A lot of what makes it terrible is accountability.
A agree with the premise that you need independence to pursue high risk stuff. The whole premise is investing in things that most people think it are stupid and will never work. That's just not compatible with standard accountability norms.
Meanwhile, as you say, contracts go to insiders regardless. It's not like FOI is preventing this corruption. I also feel that (at least on this thread) the British take on corruption is overblown. Was there ever a time or place where sleeping with the mayor of London was not good business? £100k seems cheap. Not saying this is good, just that I don't see the escalation that some here see.
On the one hand it is good that they recognise that big wins often involve big risks. On the other hand, it is important that they decide how to tell the difference between great potential and milking the cash cow.
As long as they have an independent ethics group who can make sure nothing untowards happens, it might be OK!
OTOH... IDK anything about VCT/EIS specifically. Any good examples success stories?
Removing a platform for people that say, in hindsight, ‘it was obvious that was going to fail’ is a step in the right direction.
If western societies insist on outsourcing R&D through misguided tax laws, they have no idea the amount of harm they are doing to the next generation of young people (who will be subservient to the innovating countries that were smarter in their policies).
I'm all for DARPA-like initiatives, but when it comes to taxpayers' money we should have decent accountability mechanisms.
Not all of DARPA's work is public, yet I'm sure that DARPA is fully accountable.
By excluding this new agency from FOI the UK government simply wants to be able to decide which projects to publicise and when, and which projects to keep secret, in the same way as DARPA operates.
A current example is the covid 19 vaccine contracts, which were multi billion dollar/euro/pound deals, with accelerated research and approvals, and so far only one contract was made public.
It is specially concerning when some vaccines were developed with public funding.
Government should absolutely have to justify how it allocates funds and which projects are taken on. This is a transparent attempt at avoiding the need to justify what is done with public money.
> Removing a platform for people that say, in hindsight, ‘it was obvious that was going to fail’ is a step in the right direction.
It really isn't. Covering failures up only makes the problem worse. Anything that does eventually leak will be blown much further out of proportion than if it has been clearly stated that it will probably fail from the start, but the benefits should it succeed outweigh the risks.
You would do even better to build a culture tolerant of failure. But the emergence of that culture depends on improving people's standard of living.
That this secret spending is coming from the same party that just months ago insisted there was no money to continue feeding children during a pandemic and who are also mired in corruption scandals regarding nepotism in government contracts just makes the idea look even more grotesque. Trying to do it during a pandemic you've just majorly fucked up the handling of makes it even less endearing.
Incidentally I happen to be in the middle of reading "The Dark Forest" by Liu Cixin. A core element of the book is how humanity is affected when standard of living is reduced and basic needs are unmet in order to meet technological research goals that will curtail an "undefeatable" foe. It generally leads to discontent towards the projects that are supposed to "save" them.
If the goal was to prevent the outsourcing of R&D then the government would be much better off taking a greater stake in universities and making sure both parties are compensated more from companies who go on to exploit that research.
Such as? Are "we" now going to say the state is better at innovating than the private sector? Under what conditions?
(I'm inclined to believe that it could work - it was after all the model that gave us Concorde - but there are also a lot of ways in which it could go wrong and I absolutely do not trust the government with a track record of bunging corrupt money at shell companies.)
The miniaturisation and cost reduction of electronics achieved as a result of continual investment by the phone industry is almost miraculous. A lot of innovation has happened to fit a supercomputer into your pocket.
The WWW example in a sibling comment is a great example of the contrasting sort. It benefits humanity as a whole, and would be very unlikely to be invented privately. We want more decentralised technology solutions - really standards rather than products - and government research have a different set of incentives that might make them better placed to deliver that sort of innovation.
That said, I work in telecomms. A large number of self interested telecomms companies have got together and agreed standard over the years (this is obviously not unique to telecomms), including making patents available (https://en.wikipedia.org/wiki/Reasonable_and_non-discriminat...). The technology is not decentralised, but it is interoperable.
At greater scale the State is definitely better at innovating compared to the private sector. Look at the internet itself and at www (CERN is basically a state project), and if we want to go even larger look at the war industry.
There's no way a "private entity" would have had the gall to build some rockets in order to bomb out London and its environs, and without the V2 rockets we probably wouldn't have had NASA and we wouldn't have had two private persons ~70 years later (Musk and Bezos) trying to build and innovate in private and with lots and lots of money what the Nazi regime did in 1943-1945 while being bombed out to hell and back by the Allies.
And then there's the entire nuclear industry which wouldn't have happened without Nagasaki and Hiroshima, both state-run (killing) projects.
That's why the title of the article is actually "Secrecy for high-risk tech research". It's 'high risk' but also aimed at being high rewards and potentially targeting sensitive domains so they don't want others to know what they are doing until after they have reaped the benefits.
I cannot read the whole article because of the paywall but the freedom of information laws do not prevent anything in terms of what to invest into. They only create a legal duty to provide information on demand.
The aim of this new agency has never been altruistic. It has always been discussed as a tool to further the UK's interests in the global competition. Clearly the government thinks that a level of secrecy, or at leat government control, is required.
Just for reference, the FOI already provides for the following exemptions:
22A Information obtained in the course of, or derived from, a programme of research.
24 Information for the purpose of safeguarding national security.
26 Information that would, or would be likely to, prejudice defence of the realm.
29 Information that would, or would be likely to, prejudice the economic or financial interests of the United Kingdom or of any part of it.
43 Information that constitutes a trade secret or would, or would be likely to, prejudice commercial interests.
I think that pretty much 100% of that new agency's work will fall within one of these exemptions so IMHO it makes sense to exempt it altogether and avoid having to deal with FOI requests in the first place. The agency along with the government will then have full discretion to make research and results public as they deemed useful and that will also allow them to work with private companies to license/commercialise stuff.
(Note that a lot of DARPA's work is also secret.)
There's a shocking number of people high up in UK tech influence circles who got there solely by networking, holding events and being on the peripheries of successful startups
Like Boris Johnson's lover.
Similar to Rohan Silva who advised Cameron; set up a co-working space, is on many boards, speaks at a lot of events
(Apologies if I just ruined the joke and you didn't really mean Carrie Symonds)
https://committees.parliament.uk/publications/4665/documents...
* drown them in documentation paperwork for the little money they get
* give money to private companies for free and without transparency
Corruption.
It makes me furious, and it should make you furious too, reader.
A solution is definitely needed for that, I don't think a government funded agency will do it though.
The last thing I want is for my taxes is to go into venture capital. Especially if it's run by the government.
I wonder what's the best place where to run to, to escape this trend. Maybe South America or South East Asia?
The only way out of this is to build better ways to keep the powerful accountable.
That's a terrible precedent.
is this supposed to be a double entendre ? I don't see how this follows from the parent comment.
Though of course all government will have secret projects in one way or another
