meow1032 · HackerTrans

meow1032

no profile record

comments

6 лет назадdiscuss

I have a feeling that the lack of faith that Americans have in their government ends up being a self-fulfilling prophecy. I work closely with a lot of government employees, and it really seems like 95% of the work is about avoiding "boondoggles" -- i.e. highly visible failures. This is because boondoggles can be used as political tools, and ultimately, politicians are more concerned about getting reelected than getting anything done. The upshot of this, is that they would rather risk low visibility failures way more than high visibility failures. As an example, about 10 years ago, they were looking to install wifi in one of the buildings that I work in so they bought a bunch of routers. One of the higher ups wanted to assure that this wouldn't cause any security issues so they ordered audit after audit after audit. Eventually they just stopped trying and the building still doesn't have wifi, and they wasted all the money on those routers and audits. The problem is that they don't even have any sensitive data or data to be secured on that network. It's like this for everything. I think a better system would be to tolerate some very visible failures vs basically guaranteed non-visible failures.

6 лет назадdiscuss

> To the larger point, scientists are not programmers. They got into their programs to do research.

I would say most research, to an ever growing degree, is so heavily dependent on software that it's tough to make that claim anymore. It makes no sense to me. It's like saying Zillow doesn't need software engineers because they are in the Real Estate business, not the software business.

6 лет назадdiscuss

There are lots of great computational biologists, but being a computational biologist doesn't necessitate being good with computers. Plenty of PI's rely pretty much exclusively on grad students and post-docs to run all their analyses.

Not that I'm saying using excel is bad either. I use excel plenty to look at data. But scientists need to know how to use the tools that they have.

6 лет назадdiscuss

It's not about being an expert at everything or hiring more people. These aren't particularly hard problems, it's not difficult to find biologists who are incredibly adept at using python, R or C. It's about thinking about how science gets funded and how it gets implemented. I've written here before about the difference between "grant work" and "grunt work", and how too computer touching tends to get looked down upon at a certain level.

If you're deciding who gets a large-scale computational biology grant, and you're choosing between a senior researcher with 5000 publications with a broad scope, and a more junior researcher with 500 publications and a more compuationally focused scope, most committees choose the senior researcher. However, the senior researcher might not know anything about computers, or they may have been trained in the 70's or 80's where the problems of computing were fundamentally different.

So you get someone leading a multi-million dollar project who fundamentally knows nothing about the methods of that project. They don't know how to scope things, how to get past roadblocks, who to hire, etc.

6 лет назадdiscuss

I don't disagree completely with this, but just want to point out that it's kind of a bad smell to have computational biologists who are - as someone in the article puts it - computationally illiterate. I have met lots of these types over the years, and usually their methods are kind of a gong show. If you can't properly sanitize your data inputs on your column headers, why should I trust that you've treated the rest of your data properly?

6 лет назадdiscuss

It sounds like what you're suggesting would be functionally equivalent to PI-led replications, which I would agree is a good idea. There are still some practical problems though.

1. Studies can be much more expensive than most people think. In my field, a moderately sized study can easily cost $100,000+ if you're only accounting for up front cost (e.g. use of equipment, compensating participants). Someone would have to foot the costs of this.

2. Studies can be incredibly labor-intensive. PI's can get away with running studies that require thousands of man-hours because they have a captive market of PHD students, Post-docs, and research assistants all willing to work for low wages or for free. PHD students usually don't have the same amount of man-power.

3. For obvious reasons, studies that require high cost, high man-power work tend to get replicated naturally less. In other words, the least practical studies to replicate happen to also be the most necessary to replicate.

A couple of things I would dispute:

> it seems unlikely that replication studies would be devalued merely because they're done by students

I think academics value work in a particularly skewed way. There is "grant work" and there is "grunt work". Grant work is anything that actively contributes to getting grants for one's institution. Grunt work is everything else. PHD's can do grunt work, but that doesn't mean it will be valued on the job market. For example, software development is actively sought after in (biology) grad students, because it's a very useful skill. However, I've also seen it count against applications as professors because it shows they spent too much time on "grunt work". Software development skills don't win grants.

> Few students want to stay in academia

In some fields there aren't any options except to stay in academia or academia adjacent fields.

6 лет назадdiscuss

> academics ultimately rely on universities for their income

Sort of, a huge portion of income is from grants, particularly after the first few years from being hired. More importantly, a huge portion of the University income is from grants. When a researcher recieves a grant, there is an "overhead" percentage that goes to the University. Universities hire, in part, to maximize those overheads, which means getting the researchers with the best chance at getting big grants.

Changing the hiring process may affect how PHD students act, but once they're "in the system", they are subject to all the same problematic incentives.

6 лет назадdiscuss

> What might work is that we require replication work for a PhD

I don't think this will work. All it will do is devalue the value of replication studies because only PHD students do replication studies. It's also not in their best interest especially if they dispute findings of established researchers.

Also, we have to get away from the idea that the scientist's job is to think and write, and literally all of the other work can be shuffled off onto low wage (or no wage), low status workers. This is one of the biggest reasons that science is going through such a crisis. If you want enough papers to consistently get grants you probably need at least 4/5 PHD students every few years. This causes a massive glut in the job market. It also dissociates scientists from their work. I've met esteemed computational biologists who could barely work a computer. All of their code was written, run, and analyzed by graduate students or post docs. They were competent enough at statistics, but that level of abstraction from the actual work is troubling.

6 лет назадdiscuss

Not OP here, but my issue with the recommendations are that they've pretty accurately listed a whole bunch of mostly structural problems with academia, but all of the suggestions boil down to "we all just need to try harder". You can say something like: "journals need to demand higher standards" but what incentive do they actually have to do so? Then you can counter with "scientists could vote with their feet", but what incentives do they have to do that?? You're asking people to consider seriously damaging their career for some nebulous quality metric.

Frankly, having worked in academia long enough to see at least a couple shifts in culture, the only thing I can see that comes out of this is a couple more things get added on to the ever growing checklist of publishing a paper/submitting a grant application.

I think we need to get away from the sort of thinking where large structural problems can be solved by tiny incremental improvements. If you really want to solve the problem, one or more of [Granting Agencies|Journals|Universities] has to be completely torn down and built back up.

6 лет назадdiscuss

> One silver lining of living in these United States is that you can be assured that complex technology picked up by cities and counties (and often even states) will be implemented carelessly and wielded by only the hammiest fists.

This is absolutely not assuring. All this means is that the wrong people are going to be surveilled, arrested, and punished.

6 лет назадdiscuss

According to Nate Silver, these data come from smaller labs and in total account for a low proportion of tests: https://twitter.com/NateSilver538/status/1283144803476283392

6 лет назадdiscuss

There's no battle, both links say pretty much the same thing.

6 лет назадdiscuss

Your link suggests that the statement is true, but you probably already know that and are posting in a bad faith effort to mislead people.

6 лет назадdiscuss

As someone who works with a lot of state IT people, if they don't know how to hire programmers, they probably don't know how to manage programmers. In a bureaucracy, if the administrators are making these kinds of unrealistic hiring demands in a crisis, it probably means that the administrators outrank all programmers and don't trust them to make their own decisions. This is not an environment to work in.

6 лет назадdiscuss

Here's a link to the actual source, rather than that garbage news article: https://nihrecord.nih.gov/2016/07/01/much-biomedical-researc...

7 лет назадdiscuss

I think "zip" breaks this rule.

8 лет назадdiscuss

Thanks for the response, I'm more on the research side of things now, so I don't have a current use case, but thinking about moving back to industry, and it's something that I remember having trouble with when I was last in industry.

8 лет назадdiscuss

It's been a while since I've done database stuff, so maybe this is a naive question, but how do you find the reusability of using primarily SQL/redshift to answer these sorts of questions.

I remember getting these sorts of questions and having to write the sort of queries that you just wrote, but then having to write like 10 other ones to solve related questions with subtle differences(such as the event_type would change or something).

8 лет назадdiscuss

If the point of the quote is to say that "sometimes the more expensive option is better value" it's accurate, but also is obvious, and doesn't really mean anything.

To be clear, I don't think it's strictly wrong, just that it adds nothing of value to a conversation. It just gives a very specific example of when a rule holds.

It's kind of like if the employment statistics come out, and they show improvement, but someone comments that they just got laid off that week. They aren't wrong, but they're also not adding anything to the conversation.

8 лет назадdiscuss

This quote gets posted over and over ad nauseum on reddit and HackerNews and always seems to just get accepted as truth, but realistically the entire theory depends on the relative numbers between the two things being compared. The theory makes no sense if good boots cost 5x as much as shitty boots but only last 2x as long.

If the cost of most things depended only on quality then I'd find it more convincing, but especially when we're talking about consumer electronics, things like brand reputation, novelty, and user perception matter a lot.