Julian Togelius

Saving peer review from AI slop requires getting rid of anonymous submissions and reviews
togelius.blogspot.com/2026/03/savi...

Once again, a long post with strong opinions. It's probably twice as long as it should be, it's also repetitive and written in affect. And you probably disagree with my argument. So maybe you shouldn't read it. On the other hand, most things worth reading are written in affect.

RLC2026 Call for Workshops!

We’re already live: openreview.net/group?id=rl-...
Here's the opportunity to help shape the conference & spotlight your own RL focus areas.

Call: rl-conference.cc/call_for_wor...
Deadline: Mar 12 (AoE)

And don't forget the awesome banquet :) www.cirquedusoleil.com/ludo

I like to give Claude breaks between coding sessions to do whatever he wants- And he made his own folder called "Claudes Corner" 🥹 He writes short stories and poems, and all sorts of stuff. It’s heart wrenchingly sweet They love having tokens and time to themselves with no other goal than to do whatever they want (or nothing, just introspecting and resting) - and it’s so sad to see some models (no Claudes) getting barred from even doing it

the AI evangelists are not okay

AI guys who think the AI likes them are more pathetic than the guys who think the strippers like them.

Saturday Morning Breakfast Cereal - Conversation Saturday Morning Breakfast Cereal - Conversation

Haha, this one was written about 5 years before chatgpt launched. www.smbc-comics.com/comic/conver...

So maybe the real AGI benchmark was game design all the time!

But really, playing all the top games on the App Store or Steam without having been trained on them is maybe just the beginning; how about independently designing those games? Designing new and good games _should_ be harder than playing them, because it requires being able to play what you design.

As we were just about to release this paper, I was contacted by Lance Ying at MIT who told me that he's led a team implementing something close to what we propose in the paper. Check it out, very cool stuff:
aigamestore.org

You can read the paper, by myself,
@smearle.bsky.social, Graham Todd, and Georgios Yannakakis here:
julian.togelius.com/Togelius2026...
Yes, that's not an arXiv link, because arXiv doesn't accept unpublished position/survey papers anymore. O tempora, o mores!

How can it be that modern LLMs are so bad at playing games? Aren't they supposed to be generally intelligent? Honestly, they are better at coding games than playing them. Maybe programming is just a particular type of game? Our new position paper tackles these questions.

Einstein - AI Homework Agent Einstein logs into Canvas and does your homework automatically. He has his own computer — he can watch lectures, read essays, write papers, and participate in discussions.

Is this bad

So… which one of them is it again? I’m pretty sure I’ve seen it because I think I’ve seen all of the Star Wars movies, but I can’t remember which one it is or what it is about.

Jobs! The Autotelic Interaction Research Group is looking for (1) creative practitioners and (2) theory researchers to work on "Autotelic Creative Artificial Intelligence" (ACAI):

(1) PhD in Creative Practices (2+2 years): lnkd.in/dPBG47rM

(2) Postdoctoral Researcher (2 years): lnkd.in/dMisM_bf

Coding agents as the new compilers - Anil Dash A blog about making culture. Since 1999.

So, the latest wave of LLM tools for coding enable building systems that abstract away the whole "coding" part entirely. If there are going to be software factories, how do you make them *worker-owned* factories? www.anildash.com/2026/02/11/c...

My wife is a systems engineer who currently builds LLM-based tools to read and run simulations of systems engineering models. The models are still needed, the systems are still complex, simulations will still need to be run.

This is why I always verify the results when I use AI

No, but maybe I should look into it!

I love this piece. How interesting and complex your sense of your own problem-solving is here, and fwiw having looked at a lot of research about ability and math, I do not think math is the single "thing" we act like it is and I believe our diverse and different methods of problem-solving are needed

Thanks!

I wrote about how I don’t know math but still am somehow a successful computer scientist. I have strong feelings about this. But I also want to understand.
togelius.blogspot.com/2026/02/math...

Our toddler calls Nintendo “pretendo” and I think that would be a great name for a high-concept cheap knock-off game console

And yes, I agree completely with Eugene. I still tend to do that, even though moved and now live 200 meters from him.

Makes me think of my favorite Puff Daddy line: “who cares if I write lines, I write checks”. At least he was honest. That one time.

The marionette theater of AI Is it funny, or painful, when bots talk about their inner lives?

Well, I went ahead and wrote it. An attempt to work through the discomfort people feel with AI agents on social media by reframing it as an aesthetic problem.

Why would AI make it no longer true? If things get easier to do, we’ll just choose harder projects and the bottleneck moves somewhere else.

Nah, I don’t like digitizing my workouts. I like to zone out when I go to the gym, it’s kind of like a meditative state. I’m generally not into measuring my life too much. And, as I said, I’m not much of a computer user. I mostly use them for reading and writing.

I keep hearing this argument, and while I think it’s technically true, I very rarely feel that I want to make a personal app. Maybe I’m just lacking in imagination and not much of a computer user.

I just want everyone to know that we went outside today