Fuck this, I'm off to read fanfiction!

It's okay if it takes a little longer than you thought.

people with no common sense: je11yf1sh, je11¥fi5h, j*llyf*sh, je//ÿf!sh, j3ï||yf¡sh, gel lee fisk

result: cannot account for the sheer amount of possible ways to alter the word jellyfish

conclusion: i have to see jellyfish now.

“Walktopus” Octopus Candle Holder by Scott Musgrove

There's lots of things in your body that are like that, because evolution works more towards efficiency than peak performance so the 'default' state is soft and slow and weak because that doesn't require as much food to maintain, but if you live the kind of life where you need to be able to lift heavy loads or run for hours on end or sometimes fight a tiger with a pointy stick then evolution is willing to let you burn more energy and nutrients in order to make that happen.

These Shaolin motherfuckers, they spend years just... hitting things. Over and over. Like, they start out punching sand, or padded wood or something, then they work up to bare wood, then stone and metal, over the course of literal years of just waking up and punching things every day. And that doesn't just make them strong, the way lifting weights or whatever would, it makes their bones ridiculously dense. Like, just carrying around their skeletons is wasting energy from an evolutionary standpoint, making them need more food than normal people just to walk around. But because their bones are ridiculously dense, because they've been almost shattered enough times, over and over for fucking years, they can hit things that would absolutely shatter the bones of any vaguely normal mammal, and make the rock break first.

WHY SRIRACHA WAS IMPOSSIBLE TO FIND AND NOW TASTES BAD

If you like hot sauces and the like you probably have been a big fan of sriracha... specifically the sriracha made by Huy Fong Foods. But, you may have noticed that since 2020 there have been noticeable times where it was sold out for months and months. Even worse - now that it is back on shelves it tastes like crap.

I did some digging.

All of the peppers for Huy Fong Foods Sriracha were grown on Underwood Ranches in Ventura County, CA. Family-owned farm (since the 1800s) that grew along with Huy Fong Foods. Starting at 400 acres and growing to almost 4000 to support the popularity of sriracha. This was when it tasted good. In the late 20-teens Huy Fong decided to demand money back from the farm for... no one really fully understands why. They then severed the contract leaving Underwood with 4000 acres of hot peppers and no one to buy them.

Meanwhile Huy Fong approached a number of other farms scattered across southern California and had them quickly spin up pepper concerns. This put a massive dip in their supply and they lost a year of sauce-making basically. Then bad weather knocked out a lot of these farms and they lost another season or two. Also the quality and flavor across multiple farms was inconsistent.

Meanwhile Underwood sued Huy Fong, won, received 23 million dollars, hired back their workers, and got back to growing. Additionally they were able to mitigate a lot of the weather issues the last few years through better technique and had bumper crops.

So they made their own sauce - Underwood Dragon Sriracha...

and lord strike me down if it doesn't taste much more like the older sriracha than whatever Huy Fong Foods is making now. Anyways, they don't seem to sell to stores but you can buy directly from their website. I did and I've been putting it on everything.

I wasn't paid to write this... I just like doing exhaustive research about things I enjoy.

(EDIT to adjust Underwood Farms to Underwood Ranches, and change location from Ventura to Ventura County)

And then there’s bananas, from Papua New Guinea:

If you don't publish the experiments that didn't yield interesting results, people are going to waste time repeating them because when they look in the database and see nothing published on them, they're going to assume the experiment hasn't been done???

And then my professor told me that scientists just "assume that if the experiment would be relatively easy to set up, but there's nothing published on it, that means someone has already done it with no significant results." WHAT? That means we could be assuming that important experiments have been done just because they seem to be too easy???

Not to mention that this also encourages people to fudge results towards significance for career advancement because no one wants to spend YEARS on a project they don't get any credit for.

The most dangerous part of this sort of publication, though, isn't academic dishonesty and exaggeration (which is ferretted out by other scientists after a few years, who tend to keep an eye on each other and know all the tricks) or missing something in the unpublished data (which is always going to happen). It's the opposite -- accidentally inventing trends that don't really exist.

In science, there is always the possibility that your result is random chance. I could give ten tomato plants fertiliser A and ten fertiliser B and mix them all up and it's possible that all ten fertiliser A plants will do great and all ten fertiliser B plants will die by completely random chance, even if fertiliser B is better, or even if both fertilisers are exactly the same. Most non-medical sciences have a confidence interval floor of p=.05 -- that is, the chance that your result was just coincidence has to be one in twenty (5%), or lower. (In applied medicine they're usually stricter).

So, as an example, let's say that 40 people do experiments that involve MRIs. One person might be checking the effect of sugar on the pleasure centres of the brain, one might be looking at some new medication, one might be pinpointing how imagining stuff translates into brain activity; it doesn't matter. Beforehand they do some baseline tests, and one of them finds, completely unexpectedly, a major difference int he baseline brain activity of their male and female subjects (p=.025)! P=.025 is pretty damn confident! They have to publish that, of course they do! So they do!

Running forty experiments, you expect this result if there is no difference, with a confidence interval like that. In context, this result means absolutely nothing. There's no difference. But the other 39 people aren't going to publish papers about a difference they didn't find while looking for something else, are they? They might mention having done preliminary tests on their subjects in the unrelated paper they do publish, and leave it at that. So what the scientific community sees is a paper finding a difference, and no papers contesting it.

Now, what you're supposed to do in a situation like this, is consider this preliminary and run another experiment. But this lab is studying something else with the brain. They threw this paper out because it's an interesting thing they found, it's not their focus. It's not anyone's focus. And a few years later, another lab does the same thing. Until you end up with a handful of pretty solid-looking papers about this difference between male and female brains, and nothing contesting it.

"Oh, but surely someone will eventually run the experiment properly and -- " and what, write one single opposing paper? Maybe someone will, if for some reason there turns out to be enough attention tot he topic to make it worth it. Maybe a few people will. Maybe someone will write a meta-analysis, comparing all the papers on the topic and diving into a few that weren't on the topic to try to find what the standard results are, trying to find the truth. It's unrewarding work, but people do it. In the meantime, ten years of popular science books have been written casually citing the "fact" that male and female brains think differently and it's been proven with MRI, and second and third generation tests have been built on this utterly failing premise to say things like "MRIs show that women are better at social stuff and men are better at physical stuff because [horrible misreading of the data from some such paper]" and "trans women's brains behave like female brains and not male brains in an MRI" (yes, it's still bullshit even if it's used in support of something good instead of something bad; bullshit science is bullshit science on both sides of the fence). And while the MRI stuff is being debunked, is the stuff about diet that help autistic kids being debunked? Or the differences in metagenomic analyses of the microbial life in different parts of the world? Or the effect of boats on dolphin populations, or whether the colour blue is calming for infants, or what types of music different fish can distinguish, or data on lion pride social structures, or whether or not almonds help with high blood pressure?

Let’s say you have three therapies/treatments (Tx) for, say, dementia. Something objectively awful for both patient and caregiver. Something time-sensitive, where effective, timely Tx can have an enormous effect on quality of life. So you, a researcher who cares deeply about this cause (why else would you have spent a decade of your life reading and talking and writing about it 12 hours a day, 360 days a year), test Tx A. A sister lab tests Tx B, and a third lab tests Tx C.

You find that Tx A doesn’t have a significant treatment effect. Tx B and Tx C, on the other hand, do have a significant effect.

Here’s where the problem comes in: you, because your treatment didn’t work, will almost certainly not be able to publish. What you have is a “negative result.” It’s interesting. It’s extremely interesting. Other researchers and clinicians absolutely should know that Tx A doesn’t work for the circumstances/population you tested.

Except, because you got a negative result, nobody will ever know.

What this means in terms of outcomes is that (1) people keep doing Tx A for dementia because there’s no contrary evidence; (2) you as a researcher have nothing to show for your time/grant money (which threatens your “publish or perish” career / any future funding); and therefore (3) researchers, not being willing to gamble their careers, tend to reach for the lowest hanging fruit lest they end up like you, with nothing to publish for their efforts, or (4) other labs will waste their time doing the same study of Tx A and getting the same results you got, not knowing it’s strengthening the case against Tx A and not being able to publish their findings, either.

(This is not hypothetical, btw, my own research was tanked by taking on a topic that had negative results too often, too early in my career.)

You also tend to see fewer repeat studies because everyone wants to be cited as groundbreaking, but repeat studies are where the strength of science comes from.

(Well, I was going to include Journal of Pharmaceutical Negative Results but apparently they discontinued it in 2022. :/ )

(Again, was going to include the All Results Journal but they have also stopped publishing since I left academia, which is a shame: http://arjournals.com/)

Because the sort of apparent difference we’re talking about is a random chance result, it will happen sometimes no matter how you divide up your participants and compare them. The choice of comparison determines the nature of the false positive result - so always choosing to group and divide participants by gender will result in more false positive gender differences. This is an example of how pre-existing biases can affect scientific results without anyone fudging the actual numbers.

I have bigger fish to fight

We’ll burn that bridge when we get to it

You can lead a horse to water, but if you drown it you have to walk home

Opening a can of whales

You made your bed, now shit in it