Dan Simpson

As a child I was obsessed with umbrellas. Unrelated, I also used to use a broom to process down the side of the house like I was an altar boy holding a crucifix. Later I was an altar boy.

I knew I was evil before I knew I was fruity. come to think of it the two are totally unrelated. well I've made a fine mess of this prompt. kindergarten, though. I knew I was host to something ancient and horrible during arts & crafts

In my immediate life I’ve only seen one example of a man who drives and I appreciate the freak.

bsky.app/profile/danp...

Autocorrect really wants me to call him Meat Load

And this is why, in 2003, Meat Loaf re-wrote the words to Tear Me Down (from Hedwig and the Angry Inch) to be about his life.

Benjamin De Kraker O @BenjaminDEKR Follow OpenClaw is interesting, but will also drain your wallet if you aren't careful. Last night around midnight I loaded my Anthropic API account with $20, then went to bed. When I woke up, my Anthropic balance was $O. Opus was checking "is it daytime yet?" every 30 minutes, paying $0.75 each time to conclude "no, it's still night." Doing literally nothing, OpenClaw spent the entire balance. How?

The "Heartbeat" cron job, even though literally the only thing I had going was one silly reminder, ("remind me tomorrow to get milk") 1. Sent ~120,000 tokens of context to Opus 4.5 2. Opus read HEARTBEAT md, thought about reminders 3. Replied "HEARTBEAT_OK" 4. Cost: ~$0.75 per heartbeat (cache writes) The damage: - Overnight = ~25+ heartbeats - 25 × $0.75 = ~$18.75 just from heartbeats alone - Plus regular conversation = ~$20 total The absurdity: Opus was essentially checking "is it daytime yet?" every 30 minutes, paying $0.75 each time to conclude "no, it's still night."

I’m starting to think the people who are excited about “AI agents” have literally never used a computer in their lives

Back when I was doing academia it was so distressing that there were no senior people who were gay without you having to be let in on the secret. But now I’m in tech maybe I judged too soon.

The gay-I c-suite dickheads are proof.

Every single person who was involved in not calling this “PEGging For Fun And Profit” hates fags more than god does.

computers talk to each other all the time, and have done for decades. having them use natural language is just a step back in terms of efficiency and security

The thing that bothers me so much about Carney's joke about greenlighting the Heated Rivalry show is that it furthers the narrative that it's a needle in a haystack. Anyone who reads Canadian books and watches Canadian film knows that we could have dozens more HRs with enough funding.

Are you murdering the BeeGees tomorrow? I feel like time has mostly got that covered.

There is never a wrong time to watch a Jarman movie.

God-honoring image of Saint Sebastian from Derek Jarman's Sebastiane

Watch “Sebastiane” (1976), honoring the patron saint of twinks getting tied up and penetrated, here:

www.youtube.com/watch?v=SuIW...

Why precisely did I think a Gaspar Noé film was going to relax me after a tough week?

YouTube video by NDR Moira Rose Schitts Creek these are dark times John but not that dark

I MEAN:

www.youtube.com/shorts/OnVxn...

This is infuriating. We've known that these datasets are rife with CSAM since at least 2023, and those are the ones that are public.

stacks.stanford.edu/file/druid:k...

Let. Them. Eat. Gate.

These are the types of moment when you really appreciate an expressive type system

Oh my god so many thoughts about the json schema spec! But what we are doing is basically building guided decoding on top of LLMs to enforce structure. JSON is one example (you want your LLM to extract free text int I structured data) but there are a lot of other things we can do as well

If anyone wonders what I do for a living these days, it’s shit like reading the JSON schema spec cover to cover and making sure when we write parsers for it they are actually correct. And also much cooler things. But we aren’t talking about the cooler things yet blog.dottxt.ai/dotjson-has-...

It’s almost comical how bad a whole class of agents people have actually deployed are.

The Greening of Religion Hypothesis posits that faith communities are in a process of transformation to become environmentally responsible. This project introduces a data science-driven framework to systematically measure such action across thousands of places of worship. $100,00 DATA SCIENCES INSTITUTE (DSI) CATALYST GRANT Meredith Franklin. Tanhum Yoreh. Vianey Leos Barajas A DATA SCIENCE FRAMEWORK FOR QUANTIFYING ENVIRONMENTAL ACTION IN FAITH COMMUNITIES

The most unexpected collaboration thus far in my career but it’s fun. We just got a grant to work on:

“A data science framework for quantifying environmental action in faith communities”

Words fail me.

A good cheese doubles as a fetish

"Machine learning research is not serious research & therefore hallucinated references are not necessarily a big deal, agrees prestigious group of ML researchers"

I ❤️ when the titles write themselves

But seriously, I don't think there's much to debate statmodeling.stat.columbia.edu/2026/01/26/m...

Considerably spicier riff: for a lot of predictive modelling, where your interest is slightly out of sample prediction, formally specifying a treatment to get an ATE is kinda like specifying a bullshit null hypothesis: mechanically satisfying but fucking useless.