Profile avatar
mikecaulfield.bsky.social
Author: Verified: How to Think Straight, Get Duped Less, and Make Better Decisions about What to Believe Online (University of Chicago Press). Researcher, infolit/misinfo/rhetoric/civic reasoning. Currently researching AI as tool for critical thinking.
7,506 posts 32,578 followers 2,246 following
Regular Contributor
Active Commenter
Posts 26 Comments 24

In a world that took science seriously you'd pull this cop's gun and badge until he had a serious psych eval. It is not safe for people this susceptible to panic-based delusion to be carrying guns.

submitted 15 minutes ago • 1 comment

This is why the people who insist that Trump is going to try to round up every protester who went out for No Kings day so you need perfect protest opsec make me laugh. They're going broke failing to hit their current quotas, they can't just magically round up every grandma with a sign.

submitted 31 minutes ago • 4 comments

In a more serious country, unilaterally withdrawing from the JCPOA would have immediately ended Trump's first presidency & destroyed the GOP's reputation on national security policy for a generation 🧐

submitted 11 hours ago • 4 comments

we've been shouting about this since gamergate btw

submitted 3 hours ago • 4 comments

When I die I’d like the @nytimes.com editorial board to lower me into my grave so I can be let down one last time.

submitted 13 hours ago • 52 comments

Unless I am missing something The University of London report on academic integrity seems to have been generated (poorly) by AI? (H/t Mark Bassett) www.london.ac.uk/sites/defaul...

submitted 12 hours ago • 3 comments

This is great news. We tried @mikecaulfield.bsky.social's SIFT framework last week. It was amazing - verifying and cross-referencing multiple claims using reputable sources. Using Claude Deep Research and MCPs for browser & structured thinking we got it to list all claims then validate each one.

submitted 18 hours ago • 1 comment

People have been asking me for this for 3 months, and until openAI allowed model choice in the GPT architecture it wasn't possible. That changed last week. And then the 8,000 character instruction limit made it near impossible. So it's still odd to me that after all this time it's just... done.

submitted 19 hours ago • 2 comments

I whittled down the best AI based fact-checkers small enough to fit in a GPT and now it's free to everyone. It represents months of thinking how to build LLM fact-checkers and a painful night of figuring out what to rip out of it. And somehow it still works! open.substack.com/pub/mikecaul...

submitted 19 hours ago • 3 comments

Kinda wild that we all get to pay taxes to fund a civil war against us

submitted 23 hours ago • 400 comments

We got him. Providing an update on the manhunt momentarily. m.youtube.com/watch?v=RxT-...

submitted 22 hours ago • 768 comments

I'm going to get ahead of some uninformed outrage. The second-degree murder charges against Vance Boelter are all that can be brought against him *right now.* Prosecutors can charge him with first-degree premeditated murder, but that requires a grand jury indictment, not just a criminal complaint.

submitted 22 hours ago • 25 comments

I think a good rule of thumb is if your preferred source of news referred to yesterday's protest as "tens of thousands" and not "millions", you need to find a different news source.

submitted 1 day ago • 33 comments

This is serious, rigorous analysis and it should change the way the media has been covering these protests. These were, in the aggregate, *massive.*

submitted 1 day ago • 434 comments

Honestly this song is a good celebration of soldiers but a weird choice for leadership

submitted 1 day ago • 1 comment

What people mocking flag discourse refused to get through their head because it was too wine mom-coded is that those arguing that protestors should bring flags were very much aware the flag isn't a shield from police brutality, and understood the optics of exactly what's happening in this photo.

submitted 1 day ago • 4 comments

Rick Steves finishing his speech in Edmonds. This is about halfway back in the crowd and to the side. The patriotism of this crowd was something else. And Steves' point is right -- change is going to come from an engaged center that stands in *solidarity* with the most vulnerable.

submitted 2 days ago • 5 comments

Rick Steves finishing his speech in Edmonds. This is about halfway back in the crowd and to the side. The patriotism of this crowd was something else. And Steves' point is right -- change is going to come from an engaged center that stands in *solidarity* with the most vulnerable.

submitted 2 days ago • 5 comments

The guy with this sign has my favorite kind of old guy energy

submitted 2 days ago • 0 comments

So I don't know who sent out the memo but I got here with a couple extra American flags to hand out and it turns out there are American flags everywhere

submitted 2 days ago • 2 comments

Trace claims quotes and media to the source. And if you can't, and if there is no high quality verification, just wait a bit, or find something more solid to boost.

submitted 2 days ago • 0 comments

i hope someone is ready to film tomorrow’s rainy military parade in B&W like a czechoslovak documentary

submitted 2 days ago • 1 comment

PRECISELY. People should be getting this by now? This is the same reason fact-checking and pushback against lies matter. Your Trump loving uncle may not care but this is irrelevant. A polluted information space impacts the institutions we rely on, and that's why all this matters.

submitted 3 days ago • 1 comment

I know a lot of my followers are pretty AI skeptical. Which is fine. But for people trying to find a middle way in-between ridiculous waves of pro-AI and anti-AI propaganda you may find this relevant: open.substack.com/pub/mikecaul...

submitted 3 days ago • 4 comments

The last three bullets there are worth more than a whole conference

submitted 3 days ago • 1 comment

My thoughts on the "get people hooked on chatting with the chatbot as their new friend" model of (some) LLMs

submitted 3 days ago • 1 comment