Kevin Kuhl

Kevin Kuhl

@kevinrkuhl.com

Technical writer, runner, interested in philosophy, tech, politics. Find out more about me at kevinrkuhl.com

118 followers449 following62 posts

Why I don't spend more time on LinkedIn: it took five seconds to find someone posting about their awesome feature which is literally a couple of group by statements. Complete with an AI hashtag and bragging about how this is revolutionary.

Kevin Kuhl reposted

KIM KITSURAGI - “You need to get your shit together.” YOU - “But I don't wanna get my shit together!” KIM KITSURAGI - “Then the world will turn away from you, and leave you behind.”

I only recently clued in that he's got a picture of the hallway for a hat, and it wasn't just weird sci-fi fashion.

This post is unavailable.

Just explained #CanadaDay to some colleagues, including that Glass Tiger is headlining in my city and now I'm wondering: if we were going to release an "Oh What A Feeling" compilation for the 2000s to now, what would you put on it? #canada #ohwhatafeeling #canadianmusic

Kevin Kuhl reposted

Since starting a new gig, blogs have taken a backseat but it's rewarding seeing my posts mentioned in the State of the Docs. I think it's pretty important to talk about what you d, and why it matters. There's lots of other good takes from other tech writers too: www.stateofdocs.com/2026/contrib...

stateofdocs.comThe State of Docs Report 2026 – ContributorsThe 2026 State of Docs report was produced by the GitBook team with support from the wider documentation community.

Kevin Kuhl reposted

I’m reading the Pope’s 83-page manifesto on AI. Notes/quotes 🧵 “Technology should not be considered, in itself, as a force antagonistic to humanity. On the contrary, it has formed part of our history since the beginning as ‘a profoundly human reality, linked to the autonomy & freedom of man.’” 🔥

vatican.vaEncyclical Letter of His Holiness Leo XIV Magnifica Humanitas (15 May 2026)ENCYCLICAL LETTER MAGNIFICA HUMANITAS OF HIS HOLINESS POPE LEO XIV ON SAFEGUARDING THE HUMAN PERSON IN THE TIME OF ARTIFICIAL INTELLIGENCE [ Multimedia ] ___________________________

Kevin Kuhl reposted

Tech engineering normalized transparency (status pages, postmortems, bug disclosures) partly to encourage other industries to do the same. Decades later it’s backfired: people hear our issues but not their own industry’s, so they assume engineering is worse rather than more transparent.

Uggh. Because I engaged with some logic stuff on LinkedIn, I'm now seeing people think that the undefinability of truth stuff from Tarski is important for LLMs and I just want it all to stop.

View full thread

Neo-logicists think that logic is more rich, and includes things that let you quantify over predicates (think sets). What they discovered, and it's genuinely cool is that if you add a statement defining same number you can prove basic arithmetic follows from their concept of logical truth. 13/14

A pretty direct result is that their formalizations are on the incomplete side of the fence. But whether *logic* needs to be complete is a different question. This is cool and subtle stuff. It takes a bit of care, but it's not that hard to explain.14/14

View full thread

So this kills the idea of proving things about the formal systems you build to model mathematics. What this poster conflated is formal logic, with logic more broadly. 11/14

That's pretty easy, because we teach logic through teaching formal models of it. Basically every undergrad philosophy student is dragged through formal first order logic. First order logic is kind of a nice system, because it's complete in the Godelian sense. But it's not that expressive. 12/14

View full thread

So this is what Godel showed. For any (consistent) system that lets us do enough math, it's going to leave something out. Sorry formalists, this is a hard stop on provability given the expressive resources necessary to do math. This is the first incompleteness theorem. 9/14

What's cool about Godel is he went on to prove that under a range of assumptions about the expressive resources of your formal system, it turns out that the sentence that expresses the consistency of that system isn't provable in that system (and as a fun twist any system equally expressive) 10/14

View full thread

There was a closely related group of mathematicians and philosophers who though that what was going with math was that really, we don't even need this laws of thought stuff. We're just manipulating symbols on paper. 7/14

So the thought is, who cares about this laws of truth stuff, let's come up with definition of what's true according to our formalisms, and see if we can build a system that can demonstrate that it won't ever prove something false, and can do the math we want it to do. 8/14

View full thread

A logicist thinks that the truths of mathematics are just truths that we get from logic. But this kind of raises the question, what is logic? Usually, they'll say something like "laws of truth"... 5/14

But really what this amounts to is the distinction between being true in virtue of logical form (like If Kevin is posting, then Kevin is posting) and true in virtue of some fact in the world (Kevin is logic-posting on Bluesky). 6/14

View full thread

It would be a pretty weird road to go from readings about what it means to determine whether a sentence is decidable, in the sense of "can we determine whether there's a series of sentences that constitutes a proof of it" to the questions that really matter about what's going on with AI. 3/14

The other one: Godel's incompleteness theorems were a problem for logicism and work by neo-logicists got around this. This is more of a thing to explain, but since it came from a public philosopher, let's talk about foundations of mathematics. 4/14

First, from LinkedIn: If you're reading about AI safety, you should read Godel, Church, and Turing. This would be a mistake. The formal definition of computability is pretty cool, you'd be better served looking at the social stuff going on, and work from the last ten years. 2/14

Kevin Kuhl reposted

Replying to @steveklabnik.com

i store prompts in my repo it's called "documentation" and any of it that's unfit for human eyes is also bad for the bot

More building with #hugo, it's not just for docs-sites, blogs and portfolios! I've been working on a project where I want to use Hugo to post a feed of in person events, so that people can add individual events to a calendar, learn about new events with RSS, etc. #events #docsascode #ss...

kevinrkuhl.comGenerating Event Feeds with HugoMore building with #hugo, it's not just for docs-sites, blogs and portfolios! I've been working on a project where I want to use Hugo to post a feed of in person events, so that people can add individ

I've been seeing people say that the point of computers/apps is to not notice them. I think that's wrong. Give me crunchy and functional apps with lots of things to understand and deep capabilities. I want tools that open up possibilities and have points of friction that encourage me to explore.

I want computers full of things like Obsidian, Hugo, Ableton, OBS, or Photoshop. I want apps that encourage expertise and empower us to do things.

To be fair the point is to do stuff outside of the computer (which is the point the critics are making), but I want to engage with the complexity. And I think this is something generative AI/agentic approaches get wrong, it really smooths over the crunchy bits.

Thoughts on Designing Automation My latest: Automation isn't just a technical challenge, it's a social one. This post explores how we can build better systems of trust and craft by focusing on legibility and intrusion. Inspired by Cory Doctorow and Ursula Franklin. Let's build `squires...

kevinrkuhl.comThoughts on Designing AutomationMy latest: Automation isn't just a technical challenge, it's a social one. This post explores how we can build better systems of trust and craft by focusing on legibility and intrusion. Inspired by Co

I'm thinking about some community organizing projects and I'm a little burnt out on Instagram and Discord. If I'm the audience, how do I reach me? Initially, flyers in common spaces and spreading the word in person at similar events. But I have ideas... #kitchener #waterloo #organizing #kw #local

View full thread

There's still network effects that make people go to Instagram for discovery, so I think that's inevitable that I'll have to create an account, but coupling a website first approach with some kind of form solution for signups Luma.com seems like a good way to get the initial communication going.

I'm curious what works for folks? It's kind of interesting thinking about how people access and get information, and I think that local organizers and meetups should really try to break away from platforms if possible.

I'm thinking about some community organizing projects and I'm a little burnt out on Instagram and Discord. If I'm the audience, how do I reach me? Initially, flyers in common spaces and spreading the word in person at similar events. But I have ideas... #kitchener #waterloo #organizing #kw #local

Then I was thinking about how flooded my calendar gets (due to some community events calendars). So if each event gets a blogpost, there's at least an initial way to post the event so that people might be more likely to read that.

My Guidelines for Writing About AI With everything going on with AI, it's important to think about what makes for good writing about AI. Here's my personal guidelines for staying out of hot water. There's a lot I've left unsaid, but these rules form the basis for what I try to write.

kevinrkuhl.comMy Guidelines for Writing About AIWith everything going on with AI, it's important to think about what makes for good writing about AI. Here's my personal guidelines for staying out of hot water. There's a lot I've left unsaid, but th

POSSE for Hugo: Part One - Theory I just completed a fun new project: A Python script to auto-syndicate my blog to the Fediverse and Bluesky. 🚀 This post breaks down the theory and the decisions. Why do manual cross-posting when you can script it? #Python #Automation #Hugo

kevinrkuhl.comPOSSE for Hugo: Part One - TheoryI just completed a fun new project: A Python script to auto-syndicate my blog to the Fediverse and Bluesky. 🚀 This post breaks down the theory and the decisions. Why do manual cross-posting when you c

Kevin Kuhl reposted

Every technology problem is three social problems in a trench coat

sarah jeong@sarahjeong.bsky.social

ok one last post -- oldheads are always going to "why is the web dying.... how can we get the kids invested in the free and open web" bro the kids can't read. that's why the web is dying. first things first