All models are reductions, but if yours fits in a 2x2 quadrant, you might have reduced too much.

Nobody complains when you make assumptions and they turn out right

It's exactly how it started for me and I still can't believe I did it twice. Wonderful conference for speakers, even with less experience, very supportive team and nice people attend, so now you have no excuses :P

If you're new my work on models: verraes.net/2022/09/crit...

It's not that I don't agree, it's more like it's treated somewhat like a mystic text (we all nod solemnly at its profundity) and that it seems like we have little else to say about models. (The essay and talk you did with Rebecca is a rare exception.)

It is misused or misunderstood (I think) to create a narrative of absolute relativity of models.

There is also perhaps a certain facileness in how we talk about models which rubs me the wrong way, which has its roots in the multitude of models and their uses, and also that we seem to be naive about models in general.

It's no deeper than the observation that a linear model is not useful for non-linear systems. But an excessive faith in the relativity of models seem to get in the way of that understanding.

The wrong model is much wronger than the right model.

It's also that people seem to read weird things into it, like all models are equally wrong, or all models are equally useful, or all models are useful on the same parameters, or that models are uniformly wrong or useful irrespective of context. I could go on.

To evaluate usefulness, we can't do analysis, we have to put the model to use and iterate. That's the antidote to perfection and paralysis: exist the model to daylight and see if people use it in productive ways.

Well, yes, the quote implies that we don't need to seek perfection, and we don't need to represent all elements from reality accurately in the model. But the second half of the quote tells us to evaluate not perfection (elegance, symmetry, realism...) but usefulness.

Interesting, I always read it differently - to remember that there's something "as good enough for my situation/context" to avoid analysis paralysis and looking for perfection :D

All of the sudden feeling an urge to go to Øredev

The quote "All models are wrong but some are useful" should not be read as an excuse to stick with your model. It's a call to actively search for more useful models.

That ceasefire lasted about as long as it takes to reload the canons.

Related to the paper we discussed at the most recent Papers in Systems meetup, "Critically Engaging With Models" by @rebeccawb.bsky.social and @mathiasverraes.bsky.social wirfs-brock.com/rebecca/blog...

What if an advertisement was your girlfriend?

I think the thing that is different with chatbots is that they have been shown to have the capability of influencing people's beliefs and behavior. Obviously this isn't a linear deterministic thing, but still, not great.

Imagine being fooled into entering a hologrammatic prison filled with make-believe personalities that know everything about you and can gaslight you into doing anything from buying products to making up a spiritual delusion to killing yourself

I added three new tips to my article on "How to Read More Books". verraes.net/2012/12/2012... I've been growing that list since 2012!

Bully logic: "I will stop beating you, unless I feel you are a threat to me". That's not a peace plan, it's an excuse for the next genocide.

Vibe codin' plus Ballmer peak equals what?

Today is a day when arts degrees are worthless, but the product of those degrees is so valuable it would kill an entire industry if they were made to pay for it.

Sometimes the constraint is Mb/s, sometimes it's WTF/s

@thunberggreta.bsky.social has done more to prevent and punish genocide than all 193 States in the UN, especially the 153 signatories of the genocide convention. Let that sink in

The only surprising thing to me, is that so many people still believe that Apple is somehow different from all the other billionaire-owned capitalist tech monopolies. These people always side with power.

Google spokesperson Castaneda: "allegations concerning YouTube are "simply not true," in part because it's a streaming service & not a social network" audit results: "YouTube’s algorithm violates the platform’s very own policies, recommending misinformation, violent content, hate speech & scams"

To be honest, given time and access to the people who vibe coded it, I quite enjoy cleaning up codebases that are a mess. Maybe I'm weird. But usually, having to deal with people who don't understand the value of the cleanup are what make a job like that shit.

I was told today that "we don't need to experiment, we don't have a problem". 🧵

The “et al” is such an amazing community! For those who weren’t able to join us: the paper is a valuable read and the discussion was so vibrant and engaging, we were a minute over before I even looked at the clock :D Thank you @rebeccawb.bsky.social , @mathiasverraes.bsky.social and everyone there

This was fun, thanks for having us @ruthmalan.bsky.social et al 😀

It's in 5 minutes, but I wouldn't miss it: "Papers in systems: critically engaging with models", with @rebeccawb.bsky.social and @mathiasverraes.bsky.social 😍 ti.to/bredemeyer/p...

I have so much respect for Greta Thunberg because she easily could have coasted on speaking gigs and Davos appearances for the rest of her life

Discussion will be led by @rebeccawb.bsky.social and @mathiasverraes.bsky.social Also, for those who can stay on after the hour, there is a further 30 minutes of “hallway chat.” All of which makes for a place to have a focused discussion, and some time to range further, and connect.

“Every model makes a choice about what is important, what categories we classify things in, what we see, what’s invisible, what’s valued, or even what’s valid.“ — @rebeccawb.bsky.social and @mathiasverraes.bsky.social from the paper we’re discussing on Monday, Oct 6 👇

ICYMI: I wrote about how your domains and bounded contexts don't map 1 on 1. #DDDesign verraes.net/2025/08/doma...

Slow is smooth, and smooth is fast

Ah Romeu, with you it's always "confusion of correlation and causation causes stupidity" and never "stupidity causes confusion of correlation and causation"

Get that thinking cap on, Jones!

If you can't imagine a universe without God, that's a failure of imagination, not a failure of the universe.

Next in the #PapersInSystems Discussion Series: Critically Engaging with Models, by @rebeccawb.bsky.social and @mathiasverraes.bsky.social When: October 6th, 2025, 1–2pm Eastern Time (US/Canada) Paper: wirfs-brock.com/rebecca/blog... More info/Sign-up: ti.to/bredemeyer/p...

"Computer science isn't political!", they said. "I'm just going to develop AI products for startup cash and to draw cute girls", they said. "You're over reacting and making everything political", they said. We need ethics training in engineering.

Can't fine the post to quote but something something AI is mediocre at best for things I'm expert in, but amazing for things I know little about, and I won't critically examine that fact at all...

No, Your Domains and Bounded Contexts Don’t Map 1 on 1. It sounds clean and logical, but it’s wrong. Read about it in my latest article: verraes.net/2025/08/doma...

This is the fundamental asymmetry of skin in the game. The AI will not eat the mushroom it says is safe for you. Only one of us is serious.

No, Your Domains and Bounded Contexts Don’t Map 1 on 1. It sounds clean and logical, but it’s wrong. Read about it in my latest article: verraes.net/2025/08/doma...

If you use genAI to write your thoughts, your comedy punchlines, your hiphop rhymes, your video game art, your PhD, who are you cheating really? Creating and honing your craft, that's the fun bit. Teaching and convincing others by the quality your of thinking and writing is a live well lived.

'A NSFW ban won't negatively affect me, I don't draw or post porn!' You sure? Here's how it went down when tumblr did it:

And this is another useful reflection on which protests are publicized internationally.