Roguelike celebration 2026 just happened and I wanted to share some of my favorite talks. Let’s start with an awesome talk about Street Photography (?!)

The City, The Street, The Lens: Perspectives on Building and Appreciating Generative Systems

Cities are beautiful, complex systems that have been built over hundreds, often thousands, of years. Many hands go into shaping a city’s character and heartbeat, and most of the work is done completely unconsciously. Street photography is the art of walking into this temporal soup and finding moments of beauty, sadness, surprise and chance, and figuring out how to capture them to share them with others. In this talk we’ll take a brief walk through street photography, and talk a little bit about why I love it (and what other, better street photographers have said about it). We’ll take some of these ideas and think of cities as procedural generators – and then procedural generators as cities – and talk about how that can help us think about generative and emergent systems in our games in different ways. It’ll give you some new language and structure to talk about procedural generators, perhaps give you a different angle to view your own work from, and – I hope – give you the motivation to pick up a camera and head off into your nearest city, to sample its procedural delights.

What does Street Photography have to do with game development? Mike Cook (AI researcher, founder of PROCJAM, and more) shared how they approach their other passion, and how some of that approach could be translated into procedural generators and game design.

Cities are constantly colliding within themselves – Mike feels a photo should have more than one reason to exist for them to consider it street photography; it’s the collisions of many different elements into a single framing what makes it a rich story worth sharing.

Roguelike Games and their procedural generators are often complex systems with identities of their own that are hard to put together in a sentence; when discussing what they are it’s often easier to think of them as the “streets” that make up the city, as these smaller venues that hopefully give players opportunities to build stories through their own lenses.

I take then that good roguelike design should aim to create opportunities for the game’s sub-systems to collide and thus give chances to players to create unique moments within the game.

There’s a lot more to their talk so I recommend giving it a look, currently it’s up at the roguelike celebration stream, but will at some point be uploaded as an individual video. I’ve been rewatching it and I find something new every time.

I also loved their responses to the audience’s questions, especially this one:

[…] Permadeath vs. No-Permadeath mode, […] do you have any thoughts on that? […] It reminds me how some photos I take could only be taken for a fraction of a second [like] they’re gone forever; but there are other places where I keep coming back to them because I know they always set up a good shot. They kind of have one of a half reasons in them already and then I’m kind of waiting for the other half of a reason. And I think when you have non-permadeath you have a chance to actually really get to know a location or a particular sector for procedural world. And that can be nice because as you return to a place, fewer things are changing and so maybe the things that are changing you are paying more attention to. [..] It’s interesting, I haven’t thought before on the question, I like that.

As I mentioned, during their talk Mike mentioned a good street photo should have at least two reasons to exist; so maybe having more control over the half of the reason can make it easier to create inspiring and memorable moments?

Also, in this context, isn’t lack of permadeath just a chance to consume procedurally generated content more than once? In which case I’d argue that a similar criteria applies for the question of how much procedurally generated content vs. how much authored content you want to put into your game, balancing the need to give players a narrative and spatial context with which they can form a stronger longer term relationship, while at the same time providing them a fresh experience every time.

In my latest project I’m exploring just that, with a world that uses procedural generation for its detailed areas (the Streets?) but “The City”, as an overall structure, remains the same on every playthrough. It’s still a long way to be showable, but it goes.

The other question was also interesting: Can excessive journalism “kill the immersion” [of a game]? They discussed how journalism of games has changed so much over the years, especially how Let’s Plays used to be a much more carefully crafted product where you had to tell a story with a limited number of screenshots and a text weaving the narrative, closer to the ideas discussed on Street Photography, where the text, and the screenshots only told a partial story, whereas nowadays you would normally just watch full playthroughs of a streamer that tell you everything you need to know. The question went unanswered explicitly but I get that player lose a chance to live the experience themselves.

I couldn’t help but remember the Let’s Play kind of things I used to do back on around 2007. After a little bit of digging I was able to find one of them surviving in backups. It’s fun how they were a mix of review, fan-fiction and imagination. roguelikes were great for that.

In any case, I really enjoyed this talk and I’ve been listening it several times already, so it’s really recommended. Share what you thought about it if you feel like it!