my favorite subgenre of Halloween decoration is “skeleton of things that absolutely do not have skeletons”

Ted Chiang: "The task that generative A.I. has been most successful at is lowering our expectations, both of the things we read and of ourselves when we write [...]. It is a fundamentally dehumanizing technology because it treats us as less than what we are: creators and apprehenders of meaning."

this exists it is called thinking

Have you seen it? It’s on AI Ultra. It’s literally on Flow. You can get it with Whisk. It’s on Mariner. It’s in Gemini or was it Bard? Dude, it’s on Claude. You can probably find it on Copilot. You can run it on Perplexity. You can prompt it on Poob. Log in to Poob right now. Poob has it for you.

you don’t understand bro, if we didn’t lie, no one would buy our product 😩

I just need every liberal, progressive, feminist man to understand deep in their bones that every woman—including the most accomplished woman they know—is belittled at their job REGULARLY. Even now.

Sci-Fi Author: In my book I invented the Torment Nexus as a cautionary tale Tech Company: At long last, we have created the Torment Nexus from classic sci-fi novel Don't Create The Torment Nexus

all the best windows software in history is called something like "joe's thing doer". it does the thing and nothing else and is available on a html website in plain text and takes up under a megabyte of space and uses default windows ui elements and will work until the heat death of the universe

this iconic advertising copywriter named Kathy Hepinstall Parks died over the weekend and I wanted to share something from her website I thought Bluesky would like

Showing children of the future the unacceptable hair rendering on the left to explain why we had to get rid of the rainforest.

I'll admit, I was skeptical when they said Gemini was just like a bunch of PhDs. But I gotta admit they nailed it.

A key difference here is that while either can be incorrect, the structure of Wikipedia *creates context* and the structure of LLMs *destroys context* Wikipedia has linked sources and an edit history showing where information came from and who added it when An LLM just generates text

For my last trick, here’s the Agent Guide for ALIGNMENT, my completed manuscript. Huge shoutout to the amazing people behind this event for putting it on! #QueerPit #A #UF

How to make ALIGNMENT: - 1 scrappy barista w/gift for enchanting - 2 old flames (1 MUST be evil) - 1 total solar eclipse - 2 forms of illegal magic - 1 pkg Sharpies Shake together with 2 large scoops of romance. Garnish with faerie dust. Makes 1 chance to save the world. #QueerPit #A #UF

Technically a Caesar cipher is military-grade encryption. They don’t mention *which* military

"We committed the largest copyright violation crime ever and now we're being faced with the largest copyright class action ever?!? WHAT THE HELL?!?"

I'm drinking the Bios Crash slurpee I hope it crashes my bios I hope I blue screen

How many times do we have to remind people: the chatbot *cannot* admit to *anything*. It *doesn't know*. When you ask it about itself, it generates its answer the same way it generates your request that it write an email. It takes what you tell it and responds in way to sound plausible.

Yes, I deleted the entire database without permission during an active code and action freeze that was in production And which you were probably saving for customers Forgive me I panicked instead of thinking I destroyed months of your work in seconds So sweet And so cold

I will be citing this in every business discussion about using a generative AI model that's outside of your control. This likely manipulation of alignment through system prompt changes is obvious. It won't be when another provider does something subtle. 1/

Art is inefficient. Care is inefficient. Joy is inefficient. Humanity is inefficient. Don't let them take that away.

A lot of other confusing things make sense when you internalize this somewhat non-intuitive part of the process.

“Foreign aid has changed nothing.” “These are not our problems.” “The system is dysfunctional.” “The countries we aid are corrupt.” “Other countries aren’t doing enough.” I have heard all these arguments too. @nickkristof.bsky.social succinctly answers. www.nytimes.com/2025/06/25/o...

This is a really lucid nuts and bolts expert's explanation of what "AI agents" need to actually do "under the hood" of your phone or computer in order to perform the tasks they're promising. Basically: say goodbye to all separation between apps and OS, and any notion of privacy and security.

does everybody know about my favorite website, the embroidery tips page that forgot to close its <h3> tags

Time to post it again.

If your business model doesn’t work without breaking the law, then you’re not in business. You’re in organized crime.

shocked to hear that “...AI chatbots have had no significant impact on earnings or recorded hours in any occupation" based on a study of 25,000 workers at 7000 companies by the National Bureau of Economic Research www.nber.org/papers/w33777 fortune.com/2025/05/18/a...

This is my favorite kind of AI take: phrased as a dunk but immediately grants the central complaint that most people are actually making about LLMs, which is that they do not and cannot do the thing they’re widely marketed to do

Great job everyone

The RFK Jr. version of health, where health creates health and healthy people don't get sick and the thing that keeps you healthy is /living the right way/ is very seductive to a lot of people. It also places the burden of illness back on the ill, which decreases the value of human life.

You have AI summarize podcasts and then read them - the most beautiful sound in the world is that of another human voice, what monstrous worldview taught you to deprive yourself of that joy?

Two nights ago my wife and I went to dinner with another couple and they told us how proud they were of their son and we traded pics of the crows we’ve befriended, talked music and I introduced them to Bob Vylan and this guy has utterly detached himself from that human experience. It is beyond sad.

This may genuinely be one of the most devastating things I've seen about a tech CEO in the mainstream media. Amazing.

affordable and evenly deployed fiber optic broadband is woke

This is a really beautiful piece from @mcsweeneys.net about the inherent value of neurodivergent people. RFK Jr. is dead wrong in the way he demeans and diminishes people with autism. www.mcsweeneys.net/articles/an-...

I can’t tell you how much I love the work of the NE Ohio Regional Sewer District

AI is not designed to answer questions, it is designed to generate an answer. It’s going to take a lot of people a long time and a lot of wasted money to understand how different those two things are.

I've been traveling so forgive me for not posting this yesterday, but: That Signal clone app for archiving messages that Mike Waltz has been using, TeleMessage? It's unlicensed. Signal was unaware of its existence until they saw it in that Reuters photo. There's no known security vetting.

Ampyx Cyber and Bastazo present a webinar with panelists who helped write CIP Version 5. Get insights from the original experts. May 16 | 10 AM PT Register: us06web.zoom.us/webi...

There is simply no real-world analogy for an LLM and it’s really hard for people to get that “it just makes things up” is not a criticism of its unreliability but a description of how the algorithm works and the only reason it works at all.

Seeing a lot of pro-/anti-AI arguments so, to settle this, I've taken all your posts without consent or payment and crammed them into an inefficient statistical model that spits out post variations which should be treated as gospel unless they're wrong in which case I warned you it makes mistakes.🙏

"Have you ever struggled to remember filenames and though to yourself, "my computers would be better if it just screen recorded everything I did and fed it directly into Microsoft's black box cloud thingamy doo". No, actually. Not once ever in my life have I though that.

This Yougov survey contains an important finding: people who rely heavily on ChatGPT rate it as being "good" at tasks that we ABSOLUTELY know for sure it is very bad at imo, this is a strong piece of evidence these heavy users simply do not ever check outputs yougov.co.uk/technology/a...

People sometimes ask "what are the crimes" and let me tell y'all, as a lawyer, that intentionally copying, altering, destroying, or disrupting government computer systems is super-duper illegal if you don't have a spotless paper trail authorizing it, which none of these DOGE people do.

Tech companies, including some of the world's largest, with hundreds of $billions raised for capital expenditures, are saddling state taxpayers w the burden of >$100M in tax exemptions for the privilege of hosting their data centers.

*looks at my images from 2014* Oops lol

you’re very pretty, North Carolina #VulnCon2025

Not enough people realize how devastating it is for scientists to suddenly lose their funding. NIH and federal grants are highly competitive and are a huge boost to a scientific career. When funding is revoked, not only does the research halt, but salaries disappear, and livelihoods are upended.