Catch up on stories from the past week (and beyond) at the Slashdot story archive


Forgot your password?

Comment: Re:Get rid of it (Score 1) 387

by umafuckit (#49785385) Attached to: Obama Asks Congress To Renew 'Patriot Act' Snooping

I submit that there is not a single human being, alive or dead, that can stay true to their promise of integrity AND be in the highest power office in the world. its not possible, its not do-able and we should stop expecting it. abs power corrupts absolutely, we all know this and we can see it, first-hand.

The president doesn't have absolute power, of course, and this quote is getting tired to my mind. I think the way Washington works makes it very hard to maintain integrity, so I agree with you. However, I also think that Washington doesn't have to work the way that it does and were it to work differently it would be easier to maintain integrity. Cleaning the money and professional lobbying out of politics would be a great first step to increasing integrity of elected officials. Maybe reforms to the way votes are counted would also help.

Comment: Re:What a guy (Score 1) 387

by umafuckit (#49785351) Attached to: Obama Asks Congress To Renew 'Patriot Act' Snooping

Regarding Obama personally: Perhaps the presidency changed him, or perhaps his campaign was a lie to co-opt the enthusiasm of the masses. I don't think we'll ever really know.

A large chunk of his first campaign was smoke and mirrors. It was defined by soaring speeches promoting "change". It's easy to interpret these in your own personal way. You then get stung when you fantasy vision inevitably doesn't come true. Of course he also failed to do shit that he promised he would, like shut Gitmo.

Comment: Re:Tighten up peer review especially STATISTICS (Score 2) 397

by umafuckit (#49774509) Attached to: Can Bad Scientific Practice Be Fixed?

Much of the problem comes from studies being published whose data is not robust because the sample size is too small to be meaningfully significant. This needs to be headlined in the abstract if it is published at all; the best magazines should refuse anything without a decent sample size, whilst the ones further down the food chain should have statisticans on hand to ask hard questions.

This is too simplistic. In some fields you can only ever get small sample sizes because collecting data is too difficult or expensive. One example is human electrophysiology studies of brain activity: you have to get quite lucky to find the right patients. Further, the term "meaningfully significant" relates to some very thorny issues. Statistical significance is conventionally defined using a p-value and this says nothing about the size of the effect. In fact, if I do a study with a HUGE sample size then I will able to detect very small effect sizes. So I can generate a very small p-value and show what is commonly (and dubiously) called a "highly significant result." Yet, if I look at the size of the effect it may be tiny. In other words, the "highly statistically significant" result may have little or no practical significance. The upshot is that you need to look at the whole study and not over-intepret p-values, or get hung up on sample sizes (although these, of course, do matter).

Comment: Can it be fixed? No. Can we circumvent it? Yes. (Score 4, Insightful) 397

by umafuckit (#49774079) Attached to: Can Bad Scientific Practice Be Fixed?
There will always be shitty studies out there. With the proliferation of these pseudo-journals there will be even more bad science out there. This science is a waste of time of money but I don't think it poses much of a direct threat to progress. The bulk of the wrong studies are likely also the obviously bad and unintersting studies. These are the studies that nobody reads. The quantity of genuinely significant work (stuff that pushes forward a field) is tiny. When something that looks like this comes out it is immediately mobbed: people rush to reproduce the results and/or use the new techniques. If it's wrong we'll know very soon. In practice there is always an attempt to replicate the important stuff, even though the publish or perish nature of science means that pure replication studies are rarely carried out and instead are dressed up as a minor extension of preceeding work. The lesson is that it's dangerous to treat a single study as definitive. Wait for the field to catch up and, where appropriate, wait for the meta-studies.

Comment: Re:Funny, that spin... (Score 1) 413

by umafuckit (#49768937) Attached to: What AI Experts Think About the Existential Risk of AI

The summary really emphasizes the minority opinion, "and only a slight majority said it would be a net positive." As if "only a slight majority" is not the majority opinion.

This isn't an election result or referendum where 50% is a magic number. I think it's correct to emphasize whether a majority is slight or not, as it conveys the size of the effect and that's of significance. For instance, if 60% of scientists thought some effect was real then this would be a majority opinion but it would indicate that the issue was under debate. We interpret the degree of consensus differently if told that 95% of scientists agree.

Comment: Re:Original M3800 Model Linux User Here (Score 1) 133

It's a seriously nice laptop - after a year or so of constant use I still really like it but it's not without certain issues, especially for Linux users.

The sort of reasons you list are why I no longer bother running a Linux laptop. In the three or four times I've done it, there have always been issues of this sort and I don't want to spend time trying to fix them. I now use Linux on desktop but my laptop is a Mac. I'm no Apple fanboy (e.g. I prefer Android), but I do appreciate have a *nix latop with features that "just work." My main gripe with OS X is that installing the more unusual Python packages (sometimes compiling those from source) is really annoying and time consuming.

Comment: Re:random breakage (Score 1) 141

by umafuckit (#49623343) Attached to: Microsoft: No More 'Patch Tuesday' For Windows 10 Home Users

> Home users will receive updates as they come out, rather than queueing them all up on "patch Tuesday."

So random breakage, then, rather than breakage on a particular weekday. Sucks to be a home user.

I'd be surprised if you can't still set the auto-update to work once a week (or whatever you want). It's just that from now on the patches will be sent out constantly.

Comment: Re:39/100 is the new passing grade. (Score 1) 174

Is there a valid reason we accept studies that have not been reproduced at least one more time to truly vet them before the community?

I don't think it really works like this. In practice, most studies are totally uninteresting and their only purpose in life is either to not ever be read or to seed more uninteresting studies. Nothing of value is lost if they're wrong (and probably they often are). The studies that do matter are replicated because they're interesting enough that other people try to use them. So if a study discovers an interesting new effect or develops an interesting new tool then other (good) researchers jump on to the badwagon. If the original study was wrong then it'll be obvious pretty quickly. In addition, within every field it's an open secret which high profile papers are actually bollocks. It's usually obvious by just reading them. These papers are generally not contested by others for political reasons, but they are ignored because the field knows them to be crap.

The major difference between bonds and bond traders is that the bonds will eventually mature.