fighting misinformation, one post at a time

This series of photographs is a piece titled From Window, which is a series of 32 silver gelatin photographs depicting the above. This was Wanibe's departure to her gallery job in Tokyo captured from their fourth floor apartment.

And there are so many articles and videos and discussions. Like, the scientific community is buzzing about this.

So much research will have to be redone because the data was absolutely compromised, off by orders of magnitude, by using standard lab gloves.

The world is probably not horrifically contaminated by microplastics. Sterile laboratories, however, are contaminated by latex and nitrile gloves.

Sources beyond dude just trust me, for the skeptics.

https://www.technologynetworks.com/applied-sciences/news/scientists-lab-gloves-may-be-causing-an-overestimation-of-microplastics-411138

Nitrile and latex gloves may cause overestimation of microplastics - Phys.org (it’s a pdf)

Yes, the animal in the picture is indeed a rough-skinned newt in a defensive position. It appears to be an uncredited repost of this photo from 2009 taken by wildlife biologist Brome McCreary, who describes it as such. Photo was apparently taken in Oregon, which is part of their natural habitat range. It's also not difficult to find other photos of rough-skinned newts that resemble both the coloration and posture of the one in this photo.

Yes, rough-skinned newts are incredibly toxic! They're part of the Taricha genus (which also includes the commonly confused California newt), which all have tetrodotoxin, the same toxin pufferfish have. The only sources I could find that specifically mentions this being 10,000 times deadlier than cyanide were these two, everything else (including this commonly-cited paper) generally cites a range between 1,000-1,200 times. (I'm assuming this inflated number is some combination of typos, old information, and exaggeration. It's a bit silly, but not really worth ripping into in the mod's opinion.) Tetrodotoxin's oral median lethal dose for mice is 334 micrograms per kilogram versus the 8,500 micrograms per kilograms for potassium cyanide. When injected, it goes down to a median as little as 8 micrograms per kilogram in mice. Which means it doesn't take much of it to kill things.

Yes, rough-skinned newts still have reason to do a defensive position despite this! Their only known natural predator, the common garter snake, has developed a resistance to tetrodotoxin that constantly pits the newts' toxicity against the garter snakes' resistance. Even if they're not a safe snack for most potential predators, they're still at risk of getting put on the menu.

This was a pain to look into because a lot of the sources I could immediately find were tabloids. Fantastic sources, as we all know.

Of course it is not good that people are using AI unquestioningly and accepting whatever answers it gives them.

It is also very fucking bad that people are not double checking ragebait news headlines, are lamenting "trends" that are a single deranged twitter user on a rant, or are seeing something third-hand in a post that drastically misinterpreted the source material and using that as the foundation for their reactionary viewpoints.

As somebody who is critical of AI it is crazy-making to see so many people talking about how AI kills your critical thinking abilities while citing points from a tumblr post of a screenshot of a news article that overasserted a claim made in a paper they didn't read that circulated and picked up tens of thousands of notes from people being confidently wrong about water use and power consumption (based on numbers from 2021-2022) IN ADDITION to using language that the authors of the paper explicitly begged people not to use when discussing their research (which, of course, the rebloggers never saw because they read a reblog of a screenshot of a news article that overstated a claim, not the paper that made the claim).

For instance, in the paper that everyone loves to talk about when they say that AI is making people stupider (the cognitive offloading one), nobody ever brings up the fact that the group that used their brain only for the first few instances of the task had higher overall performance than the AI-using group and the search engine using group and then had higher performance over their own previous metrics when using AI in an optional instance at the end of the study.

Their doctorate is real and from an accredited institution

Kent Hovind is not a doctor. He's a creationist conman. His "doctorate" comes from the unaccredited Christian diploma mill "Patriot University". The "propaganda in science textbooks" he's talking about is the theory of evolution.

Example 2:

From the SAIT LibGuide "Evaluating Sources: RADAR"

Relevance – How relevant is the information to your assignment?

Ask yourself if the information relates to your topic or answers your research question.

Consider what specific need this information is fulfilling. Do you need more sources on this specific aspect of your topic?

Authority – Who/what is the source (author, publisher) of the information?

Check the credentials or qualifications of the author. Does the author have credibility through education or experience to be writing on the topic?

If there is no author listed, evaluate the authority of the company/business/organization that published the information.

Consider the affiliation of the author or publisher.

Date – When was the information first published or last updated?

Consider if you need the most recent information on your topic.

A general rule of thumb is to use content created in the last 5 years, but older content may be appropriate depending on your topic.

Historical sources may need to be used to provide context.

Appearance – What does the source of information look like?

Different source types have different elements that may affect the quality or level of information provided. Consider where the information was published. For example, is it published as a webpage, in an academic journal, in a trade publication, etc.?

Consider the way the information is presented: articles supported by evidence and citations are more credible, while editorials, opinion pieces, and blog posts may be more subjective.

Check who/what the author references and evaluate the quality of the information they use to support their argument.

Check for advertisements - inclusion of ads isn't necessarily a red flag, but it may be depending on the content of the ads (are they related to the content in the source?)

Reason (for creation) – Why was the information published in the first place?

Consider if the information was created to inform, teach, entertain, persuade, sell something, or for some other reason.

Are there any clear biases in the information? To what degree is the information objective and impartial?

RADAR is not a yes/no test, or a be-all and end-all guide. Use RADAR to consider the relative quality of information as you are searching.

Biased, opinionated or even false information can be included in a research project to effectively highlight dissenting opinions or identify commonly held errors. The reliability of such information, however, should always be clearly identified and placed in context with more balanced sources.

-----------------

How do you expect to transfer calories from, say, the deer in a forest to the fish in the river, without mosquito as the middle step? How do you expect to transfer calories from rats to bats, without mosquitoes?

I don't know if it has a name but the fallacy in question is applying an unconditional probability to a conditional situation.

The examples above are comical but this fallacy does have adverse effects in real life.

A famous example is Sally Clark, who was tried and convicted of murdering her infant sons. The prosecution claimed that the chance of Clark having two children die of Sudden Infant Death Syndrome (SIDS) was vanishingly small (1 in 73,000,000), making it beyond a reasonable doubt.

They arrived at that probability by multiplying the probability of a single SIDS death in affluent families (1 in 8500) by itself. But this calculation assumed that the two probabilities are independent when they're not.