Depends on the government of course. A friend was tasked with studying some supposedly more eco-friendly fuels (biodiesel) the government had invested a lot in, and he disclosed his initial findings that they weren't eco-friendly at all, and had his funding pulled and his research buried.

You're comparing apples to oranges; Government not caring about negative/positive results and Google only funding things they care about.

Are you implying that government funds research when the topic isn't something they care about, or that Google only likes positive results?

No, I was responding to a comment that made a comparison between Google and government. I'm not saying the government never cares about the outcome of research, just that some (much? most?) of their funding goes to answering research questions where the government is agnostic.

The government is made up of people who are fundamentally no better or worse than anyone else.

Instead of accepting that anyone is an angel or devil, we should approach all results with a critical eye and push for reproducibility.

Google like the tobacco and oil companies and soda companies, etc largely/only want to fund research that furthers their business and public policy aims.

It's pretty simple. To equate the two, you'd have to say the USG only funds research that increases the size and scope of USG.

(Disclosure that I work at google)

Well, I mean there is that whole thing about how federal funding cannot go to gun violence research. So there are absolutely types of research that the government not just passively, but actively attempts to prevent.

I'm not saying that Google is being a good guy here, but I also don't think you can so easily make the argument that the (US) government is any better.

That's absolutely not the case. Government-funded research review processes are actually more prescriptive and restrictive and political than Google is. I know this first hand. It shouldn't be that way, but it is. Oddly, what ends up happening in a lot of government-funded research review systems is that those reviewing proposals are trying to guess "what kind of research would industry like to see?" so you get the worst of both worlds.

Also, Google is largely funding CS research -- this isn't the kind of thing where there's a double blind study going on. It's usually work on a new algorithm or protocol or architecture or ... -- things that are constructive engineering tasks that can be replicated by others when the researchers' papers are published.

> I think this is less of a concern with government funding, since in many cases the government doesn't care what you find — they just want you to do reliable research to see what the answers to relevant questions are.

I think that's a somewhat idealistic view. From my reading (which may be out of date, the book was published in 1999), even the basic research funding from governments often has a technological goal in mind, and many of the savvier grant applicants know what that goal is.

> For example, if you say "I'm going to research the effect of eating lots of eggs on cholesterol", the government won't care what outcome you find/publish. They funded you because they want the truth to come out about the question you seek to answer. There are some topics where the government "has a dog in the fight" — but much of their research funding is for areas where they are truly agnostic.

This isn't true at all for NSF and NIH grants, FWIW (and "I'm going to research the effect of eating lots of eggs on cholesterol" is not at all the level of specificity of proposals, which is part of this).

It's cause for criticism about how they approve grants (null results are science, too), but it's also partly a natural result of having limited money and many many requests for it.

There's different government grants. Basic research stuff isn't the same as a DARPA grant.