Hacker News Reader

Comment by dluan | Hacker News Reader

I'm going to argue that direct public funding is the best way, and by that I mean the historical analogy of direct patronage or crowdfunding research today (e.g. experiment.com).

Today's 'big science' grant funding model is so far removed from the average citizen or constituent, that by the time they see or feel the impact, it's hardly recognizable anymore. Because it's behind paywalls, or is too technical, or locked in IP, or anti-collaboration.

What we need to solve this problem is a direct connection with research and science. If you were motivated to go out and seek an answer for yourself, then perhaps you'd be less likely to lean on witnesses, testimony, or experts that you don't understand. The reality is that before the internet, people never really had that ability to go out and seek the truth for themselves. And I mean 'Truth' with a capital T, within significance, reproducible, and transparent.

At least that's best thing short of everyone becoming a scientist, or doing science independently.

View on HN

> I'm going to argue that direct public funding is the best way, and by that I mean the historical analogy of direct patronage or crowdfunding research today (e.g. experiment.com).

Sadly, you'll be hard-pressed to do direct public funding in anything that is not publicly-attractive. You'll get a gazillion bazillion dollars for research on canabinoids, and barely a dime on some boring ray-bending materials research problem that will eventually turn out to be easy to integrate on an organic substrate, opening the way for silicon-on-organic optoelectronic devices.

The "average citizen or constituent" has a level of scientific education that would prove at least as, if not more toxic than commercial interests. The current (absolutely terrible!) grant funding model is indeed too technical for the "average citizen or constituent"; a frickin' k-12 textbook is too technical for the average citizen, of course high-level research is too technical! That's why it's high-level!

> If you were motivated to go out and seek an answer for yourself, then perhaps you'd be less likely to lean on witnesses, testimony, or experts that you don't understand.

That's how we ended up with the anti-vaccines craze. People are motivated to go out and seek an answer for themselves, but they're also lazy and seek comforting answers.

Couldn't the money gathered be split up?

Like, 75% of the money goes to the funded project, 5% to the crowdfunding company and 20% is given to some less popular research.

Problem is, how to distribute those 20%.

Also, these platforms would be less attractive to the original researchers, because they lose 20%.

I upvoted and I also want to say that your response was most convincing/seemingly logical.

I wonder if letting people directly control priorities for some percentage of research spending (10%?) would encourage more science spending in general though. Like even if that 10% is flushed down the toilet, maybe it's counteracted by a much larger budget over all.

> Sadly, you'll be hard-pressed to do direct public funding in anything that is not publicly-attractive.

I think this problem can be solved exactly as it is solved in academia. Since technical things are indeed too technical, even many researches do not understand the precise value of a particular research. And this is why a commissions are created and peer-review is used. So the money from ordinary people should go to this commissions instead.

Upd: of course such a commission should be able to explain/promote the corresponding research to the public. This will be their work, too.

This may be a new up and coming model.

Take a look at this, http://www.lifespan.io/campaigns/sens-mitochondrial-repair-p...

A crowdfunded mitochondrial repair project, fully funded ($34K) with 14 days to go.

Well, you'd get my money for a ray-bending materials research problem that will eventually turn out to be easy to integrate on an organic substrate, opening the way for silicon-on-organic optoelectronic devices.

But in all seriousness, there are different types of research, no? Some research leads to new technology and some is verification, repeatability and just plain knowledge discovery.

There might be some way of rigging up the system so that people who fund potentially patent creating research could get a cut of the royalties.

And the reproducibility area would be great for a large number of people especially if such things are kept track of ("This particular study has only been done once"). If we can get people to trust repeated studies more than not repeated that site will help better clarify emergent research trends/topics.

Maybe in the long term - though I very much doubt that patronage was unbiased, nor will crowdfunding. Imagine the fundraising power of anti-vaxxers, for example, in funding utterly junk science.

I like the idea of experiment.com, but right now...it's just not there. I want it to be, but it's just not. None of the projects you link even approach "Pay a grad student for a year", let alone supporting an assistant professor and their lab.

I'd also like to note that you're taking "close to the average citizen or constituent" as an inherently good thing, and I'm not necessarily convinced that's true, or at least so true it doesn't need examination. What's worth research is not necessary what captures the public imagination. Heck, its not even what captures the imagination of other scientists - I have no idea what's worth funding and studying in high energy physics or theoretical CS, and I very much doubt they know the same for my field.

What about http://experiment.com/curebatten?

Still, big science does not equal good science. In most cases, the opposite is true. I find it's such a huge stereotype, even within research, that scientists forget about the creative, entrepreneurial, or hustling spirit of doing science that they default to pointing to have's and have-not's. "He has a big grant, so I how can I compete", and that zero-sum attitude is so counter productive.

> an inherently good thing

Yes, I vehemently believe it's a good thing. If we ever approach the levels of science literacy needed to create the adverse effects you're hinting at, then that would be fantastic. And, not knowing what's worth funding isn't a citizen's fault, that's on the scientist. If scientists were more vocal about their work, new opportunities, and big challenges, then of course the public would be able to identify what's worth funding. This is because communicating science is inherently a good thing, and when done well, it can motivate people to learn new things or act differently. Look at what pointing a bomb at the direction of the moon did in the 60s.

With respect to the problem in the article, these regulators and policy-makers need a big rocket to point at a single direction. When people are disconnected with what's out there, then there's no context for when science gets taken for a ride with a different agenda.

How frequent are studies being funded at that level? Based on the numbers on your website, that study is almost half your total volume.

Again, using the numbers on your site:

5,615 projects created and 364 funded is a 6.5% success rate. If I'm going to be pouring energy in to single digit success rates, they need to hit pretty hard - and while you do have some successes, right now their long-tail oddities. If experiment.com takes off and you regularly get R01-sized projects, that's awesome. But it's just not there yet.

And while "big science" does not equal good science, there are some things that just have to be paid for. I'm not talking about massive grants. And I've got experience doing the whole scrappy underdog thing. It's not a zero-sum attitude - it's that the effort for crowdfunding at the moment isn't necessarily worth what comes out of it. I think your problem is that you're conceptualizing your critics as proponents of "Big Science". I funded by dissertation with grants that aren't even approaching big science-type grants, but they brought in way more money than most of the projects on experiments.com, and required, combined, four pages worth of proposals.

> Yes, I vehemently believe it's a good thing. If we ever approach the levels of science literacy needed to create the adverse effects you're hinting at, then that would be fantastic. And, not knowing what's worth funding isn't a citizen's fault, that's on the scientist. If scientists were more vocal about their work, new opportunities, and big challenges, then of course the public would be able to identify what's worth funding.

Would they? [Citation Needed]. I'm about as science literate as the average citizen could possibly be expected to be, and as I said, I've got no idea what's worth funding in fields well past my own.

As I've said, I really like the idea of experiments.com. I've considered it in the past. I'll keep an eye on it in the future. But right now, suggesting it as an alternative to traditional funding streams, or even industry funding is, at this point, just not realistic.

Sure! Point taken. What do you think it would take for it to work?

By the way the number you want is 860 projects launched. The 5,615 figure is people who've started working on a project but haven't launched. So closer to ~42% success rate.

That is, I'll admit, a much better success rate. Still not as high as I'd like, but I was honestly surprised by my ~7% figure - I didn't figure it was actually that bad.

I can't give you a specific figure, but off the cuff? I'd like to see decent numbers of projects regularly getting funded at the ~ 50K level. That would have a couple benefits:

- We're now talking about a year's funding for a postdoc/grad student/significant % effort for a prof. This is I think a fundamental hurdle. We can't pretend that crowdsourcing is a viable funding stream if someone else has to keep the payroll going.

- 50K is also about the level, at least in my field, where we can talk about long-term investment purchases. Equipment, computers, larger datasets...the kinds of things that are hard to put on a single project.

- That's also about the point where spending time on crowdfunding campaigns likely won't irk chairs, tenure committees, etc. A couple years back I asked the folks in the fiscal office at my department about how a ~3-5K crowdsourcing project would work, and I got serious side eye.

- It's also about where modest but decent pilot grants come in in my field generally.

That's great and very concrete, thanks. It's funny, at those levels is also where more friction starts to get introduced, because suddenly the system rears its ugly head and tries to co-opt it. If you were to regularly bring in $50k, suddenly the university wants more than half of it and it dawns on you that you don't quite need the university as much as they need you.

The extreme is a Jack Horner type who would eschew academia (but oh, the creative control he has!), but the thing I like about the size that we are right now is that we're flying under the radar of the tier-1 schools, and we're just funding and seeing lots of cool stuff. I'll happily duel university administrators in the future, but for now we get to focus on small pockets of really impactful science.

For the record, you can prevent the university carrying off half of the grant by stating as an official policy that you only pay X% overhead, where X is some much more reasonable number like 15.

Many private companies do this, and as long as it's something that's stated clearly and publicly, grants offices often go along.

> Today's 'big science' grant funding model is so far removed from the average citizen or constituent, that by the time they see or feel the impact, it's hardly recognizable anymore. Because it's behind paywalls, or is too technical, or locked in IP, or anti-collaboration.

Fundamental research often has no direct impact on the public. If it does, it's often 10 or 20 years down the line. Fundamentally research is, by it's nature, removed from the impact it has. Nobody knows what the final results will be.

This isn't conducive to crowdfunding.

Here are three projects funded by normal human beings. I'd love to hear your thoughts on impact.

http://experiment.com/snails http://experiment.com/dicty http://experiment.com/vanek

I mention this above, but my first thought on "impact" is that they're really neat experiments, but it's very much in the domain of "enough preliminary data to keep a grad student busy for a bit and justify a pilot grant submission". A great stepping stone, but...

I'll be honest, I looked at crowdfunding when I was working on my dissertation. It was more work than an industry grant (especially when you include wooing the public) for a fraction of the money.

Crowd sourced scientific funding is a fine way of supplementing existing infrastructure and it's good for funding small side-projects that augment existing grants / research groups. It's also a good way for amatuers to get a hobby project funded.

But it's not an adequate alternative to more traditional forms of scientific funding, for a lot of reasons.

First, the amounts are just too small. Even a very successful crowd funder would have an incredibly difficult time paying for a single full time employee, even at grad student rates and even before administrative overhead / tuition. Which is essentially $10-$15/hr.

Second, if you look at sites like experiment.com, it's rare for a project to be funded without a very clear picture for what questions are being asked, how they are being asked, and exactly how they will be answered. And none of the projects I found asked to recoup that cost. This is particularly problematic for fundamental research, where scope and intent are not well-defined almost by definition.

(Being a cynic, I'd be unsurprised if a lot of the people doing the crowd funding thing in small dollar amounts are mostly just in it for the sex appeal it will have on their next grant report, rather than e.g. actually needing the money. There are often way easier ways to scrounge up a few thousand bucks at a major research university. Heck, were I did grad school, you could get grants in the small thousands just for being a grad student with a pulse, and doubling that money was about as easy.)

> sex appeal

Actually, it's the other way around. Granting agency budgets are so constrained that funders are looking for heuristics to de-risk their funding decisions.

You're fooling yourself (or someone is fooling you). The small numbers just don't make sense. $1k-10k is such a miniscule portion of the total cost of even a grad student+. No PI with even the vaguest hint of a conscience is using uncertain 1k-10k increments as anything other than purely supplementary funding; if one project doesn't work out your student is done. In fact, I doubt that most universities would even let a professor take on/keep a student with that sort of funding scheme. And it certainly wouldn't be enough for a tenure case.

You do realize that the average CS professor in the US spends many thousands of dollars per year, just on conference travel, just for themselves, right? $5k or even $10k for a side project really isn't substantive.

+ which in the sciences is usually in the mid five to low six figure range, before the cost of lab materials, travel, etc.

Of course, no one has yet funded their entire PhD this way. It is however a fantastic way to fund a conference trip. There've been quite a few successful projects funding conference presentations or travel.

In my experience speaking with program officers and grant reviewers, particularly NSF, the public or 'broader impact' section of a grant application is starting to grow in importance, because like I said above, it's become a way for grantees to stand apart from the growing number of applicants. So having more to put in that section certainly helps, and I do know that it's come up with a few tenure committees (successfully). It just depends on who, where you are, what field, etc.

CSE is always a bit of an outlier, given the lower costs but still. I realize your hostility is towards the notion of someone being entirely alternatively funded, which we've never claimed to do.

> and I do know that it's come up with a few tenure committees

If you're saying what I think you are, then that's exactly what I meant by the sex appeal comment. Regardless of how it's talked about, people aren't treating it as an alternative; rather, it's a particularly sexy supplement (because it's new and because it addresses dissimination / public awareness / education broader impact criteria that their more traditional sources of funding care about). So even though they could have picked up 2k in travel funding from a random pool of university cash or just funded everything out of an existing grant, they do the crowd funding thing instead for the sex appeal factor. But it's also possible that's not the case, they really need the money, and I've just spent too much time around the wrong types of academics.

> I realize your hostility is towards the notion of someone being entirely alternatively funded, which we've never claimed to do.

I don't mean to sound hostile. Maybe incredulous. Your earlier posts certainly sounded like you were proposing an alternative :-)

Anyhow, good luck.

To partially address this, those of us in the US should support the Fair Access to Science and Technology Research bill[0, 1]. FASTR would essentially mandate all journal articles resulting from federally funded research to be published in an open repository online after a short embargo period, as in NIH PubMed Central.

To effect real change, we'll need to see senior researchers lead the way in publishing in open journals, and there just isn't buy-in among that crowd in some fields outside of math, physics, and CS.

In addition, I think researchers need to do a better job personally engaging the public regarding their research. Fully addressing the gap between researchers and the general public may well require an educational component.

0: http://www.sparc.arl.org/advocacy/national/fastr/faq 1: https://www.congress.gov/bill/114th-congress/senate-bill/779

"I'm going to argue that direct public funding is the best way, and by that I mean the historical analogy of direct patronage or crowdfunding research today (e.g. experiment.com)."

I'd say there isn't a best way. The money will "corrupt"; I scare quote it because when it's just an unfixable constant you're sort of stacking the cognitive deck calling it "corruption", which usually implies a conscious choice. Therefore, the best solution is to spread your bets; government funding, corporate funding, crowd funding, startups, that thing YC is about to do, any and every other funding model you can get your hands on.

You can stop arguing about "which" is corrupted, because the answer is "all of them". But hopefully in the crossfire of all the different "corruptions" the truth will emerge. And if it still doesn't get out, well, we took our best shot.

Direct public funding is a horrible idea.

First off, it would greatly misallocate research funding. The people who'd get the most funding interest would be those who can pitch their research to the public the best, to the expense of those who'd struggle to communicate their research. You'd see AI get good funding while things like compiler or parallelization research would be rather poor, for example--even though the latter would conceivably have more immediate impact on peoples' lives.

The other problem is that there is a loud crowd of people who aren't interested in objective research, but want researchers to prove their facts. The anti-vaxxers or anti-GMO people would fund study after study until they found one that "proved" their claims. They already do this today, but trying to move to a purely crowdfunded research model would reduce the research that provides counterexamples to avoid the implied p-hacking.

Research today has many problems. That most journals are pay-to-access, in my opinion, is actually among the least important. Let's face it: very few people who aren't researchers are going to care to actually try reading 10+ pages of dense scientific text as opposed to a journalist's lay summary; there's also very few disciplines were amateurs without access to research infrastructure are capable of doing publishable research (admittedly, CS and math are disciplines where an amateur can do it just about anywhere, and I suspect most commenters here are most interested in these disciplines as opposed to agronomy or microbiology).

Bigger problems are underreporting of negative results (e.g., we found arsenic-based life! Actually, now it looks like those results were wrong...), the desiccation of research funding, the lack of interest in "fundamental" research, reputational bias in grant funding and even research paper acceptance (it's generally possible to guess which group is publishing a paper, even if it's a blinded submission), etc. None of these are helped by crowdfunding, and, generally, they'd actually make them even worse.

Reply on news.ycombinator.com