Half my feed today is about the heatwave and the climate crisis. And then this. It’s insanity.

We need Broken Windows theory but just for white collar crime.

feels like a real landmark of impunity that major corporations are making federal crimes part of the interview process

It’s because a lot of people’s antifascism is actually fascism anti-fandom

It never stops being amazing just how out of touch tech executives are, not only from society at large, but with the very businesses they are in charge of and yet don’t seem to understand what their products and services are actually for

Safety engineering for the collective human spirit

Very cool to build a system, take credit for anything good you claim the system does, and completely absolve yourself from anything evil done with said system. Strange how that works.

Interesting: AI summaries mean Google is now indisputably a content creator and can be held libel (in Germany, at least) for incorrect AI information. www.dw.com/en/german-co...

Science and technology studies for studying the occult

I wonder if it would be crazy of me to consider doing a sociology PhD

If you can’t choose to not be associated with some person, group, or idea, you don’t actually have freedom of association; and a lot of people loath the idea of freedom of association

jfc seeing tech guys responding to the OpenAI theft of Apple trade secrets by saying "Apple should be thanking them" is such a perfect example of the culture of no-consent in tech

For validation, there is more consistency, but limited scope. In 8 tasks (across 6 papers), authors reported no attempt to validate their measurements. Most commonly, researchers assessed convergent validity (e.g., comparing individual model predictions to a reference), but details varied greatly.

For operationalization, we find reasonable reporting, but with a wide variety of approaches, including some use of few-shot examples, model roles (e.g., "You are a medical expert"), attempts to constrain the output (e.g., "Answer only with a number."), prompt engineering, and so forth.

Overall, we find a strong lack of consistency or clear norms. Surprisingly, researchers often do not define their concept of interest (29/50 tasks). Rather, they might ask a model "Is the following post offensive?", relying on the model to correctly interpret the meaning of "offensive".

Everything in the AI industry is built on theft and cheating.

BREAKING: Apple sued OpenAI for trade secret theft, accusing the artificial intelligence startup and its hardware chief of engaging in a coordinated campaign to steal information about upcoming products

I think I've probably written enough posts about it that I could use as a basis to write a philosophy book on my theory of participatory make-believe, but that would require me to take doing research about it seriously which I am always avoiding doing for reasons.

I mean, that's what the actual magic that actually exists in the world is: someone creating change in the world in accordnace with their will through an appropriate medium; it happens to be the case that beliefs are a highly effective medium, the medium par excellence, for reproducing change.

Not to be a hater but while I think I get what the onion is trying to do with their new info wars, I can’t shake a feeling that there’s something a bit Eco’s Foucault’s Pendulum about it where conspiracist make-believe ends up crossing the line of satire to just doing the actual work of conspiracism

Hey y’all, I track rural policy news for a living, so I wanted to share a quick and easy explainer about what led to the current Cyclospora outbreak. In a nutshell, Trump and the wingnuts he tapped to lead the CDC are to blame: 🧵

An excellent piece from Vogue about Meta's continuing attempt to make their "smart glasses" fashionable and cool that makes all of the right points about the ways in which normalizing ubiquitous surveillance creates a world we don't want to live in. www.vogue.com/article/do-s...

#1 way to make an academic pundit mad at you: tell them the purpose of knowledge production is, in fact, improving the world which is, in fact, activism. We don't all have to have the same view of what "improvement" looks like for that to be true.

It never ceases to amaze me how so many people can spend 5+ years in a PhD program learning about knowledge production and still believe that neutrality exists and apolitical research is possible.

I literally got roasted by a couple of 8-year olds in a doctor's office earlier this week for using a lot of words in my speaking

One thing I keep coming back to is that the reason the Totenkopf didn't sink him, imo, is that the only place in the political universe people cared about the tattoo was Bluesky. Schumer, Gillebrand, and Mills gave no indication that they thought it mattered. Neither did the ADL!

Stop trusting platforms that don't let you own your content and relationships. Stop trusting platforms that don't let you own your content and relationships. Stop trusting platforms that don't let you own your content and relationships.

We really must acknowledge that when you see “AI” making half-assed work it’s because someone looked at it, said “good enough” and shared that half-ass work just to stand by it, point out that it was half-assed, and ask you to blame “AI” for making the thing they asked for and opted to use

Speaking of automated spam being a problem, has anyone else noticed a sharp uptick in email spam lately. I used to get ~30 spam emails in my spam box ever month or two, now I'm seeing 500 per week, and Google's spam filters are also noooot doing their job anymore and I get 20/day in my inbox as well

Interesting to consider that instead of another Great Awakening we're closer to a "schism at large" but it involves multiple denominations vs evangelicals

the main schism points imo are this (if you pay attention closely, you'll notice Jesus mentioned by name a lot less lately even if most of them won't outright say this yet) + a dualism where the devil is either equally powerful as God or outright more powerful and that's why there's gay people

One of the truly insulting things about big-name mags publishing AI-written stories is that there are those of us out here who still anxiously write and report each story we do – and get paid next to nothing to freelance for big-name mags

Some more context and history on Leverage Research and other rationalism-associated organizations, if you're interested:

This is one of the most deranged things I've ever read in my life. www.todayintabs.com/p/over-lever...

Did I mention that we've been writing a book? Fundamentally, security, privacy, trust&safety, AI safety -- all the fields aimed at dealing with dealing with how technical systems hurt people and how to avoid it -- share a foundation. We cover how to think and achieve *truly* effective practice. 1/

Ironically, one of the problems is that misogyny *is* a form of cultural practice, or a culture really, and not simply an individual failing. It is a socially learned way of life with its own beliefs, practices, institutions, media forms, etc that people participate in and enculturate people into.

The cult of line go up

As much as 40 percent of longform posts on LinkedIn and X that users actually see are fully AI-generated, according to new research Particularly notable because we know these platforms are flooded with slop, but this studied what the algorithm actually shows people www.404media.co/linkedin-and...

Someone tell me again how we're going to fix everything at midterms because elections were going to proceed as normal.

"Once these systems have you in their crosshairs, there's pretty much only one way it can go."

Reminder that theft of likeness and duping the identities of experts and creatives is one of the leading AI business models, as conducted by the largest companies operating in the space, and the victims are fully aware of this and how it impacts their livelihoods.

Instagram saw all the nonconsensual imagery generated by Grok earlier this year and still decided to opt in all of its adult users to having their photos "remixed" by AI www.nytimes.com/2026/07/08/t...

Huge amount of "long-form" posts on LinkedIn and exTwitter is AI spam. We are being drowned in slop, and the big platforms are fine with that.

Standard tech corporation internal information hiding behavior?

Losing Google Search is not the loss of a public good - it’s the loss of a rent seeking extractor that enclosed our information & sold the story that it was providing a public good. But rebooting real public goods Google ate will not be easy.

"The billionaire quest for political power and the billionaire quest to defeat death are, I’d argue, the same project wearing two hats. Democracy is a limit on power. Mortality is a limit on power. Both projects are, at bottom, a tantrum against the words 'limit' and 'reality.'" - Gil Duran

been saying this - it is actually an opportunity to make it a cultural point of pride. Technical measures matter too, but make them cultural first. A reason to pick the site based on what it will fight for.

Sincerely, how do you recover from moral injury?

Being in tech is great sometimes