End of an Era(erasmatazz.com) |
End of an Era(erasmatazz.com) |
But then you play the games, and they're all so opaque. You have no idea what's going on, and the responses to your actions are so hard to grasp. But if you do figure it out, the model usually collapses into a linear, repeatable strategy and the illusion of depth disappears. You can see this happening from the start, with Gossip. Instead of noticing that his game didn't communicate and looking for points of accessibility, he plunged further forward into computer modelling. The failure is one of verisimilitude: The model is similar to a grounded truth on paper, but it's uninteresting to behold because it doesn't lead to a coherent whole. It just reflects the designer's thoughts on "this is how the world should work", which is something that can be found in any comments section.
Often, when Crawford lectured, he would go into evo-psych theories to build his claims: that is, he was confident that the answers he already accepted about the world and society were the correct ones, and the games were a matter of illustration. He was likewise confident that a shooting game would be less thoughtful than a turn-based strategy game because the moment-to-moment decisions were less complex, and the goal should be to portray completeness in the details.
I think he's aware of some of this, but he's a stubborn guy.
Time and time again I would send my friend Dave Walker an email declaring that Javascript (or something else) was utterly broken, incapable of executing the simplest program without errors. Dave would ask to see the source code and I would present it to him with detailed notes proving that my code was perfect and Javascript was broken. He’d call me, we’d discuss it, and eventually he’d say something like, “Where did you terminate the loop beginning at line 563?” There would be a long silence, followed by the tiniest “Oh” from me. I’d thank him for his help and hang up. A week later, I’d be fuming again about another fundamental flaw in Javascript.
Many of us are stubborn and will work hard and long, without much positive external feedback, under the assumption that our vision is correct and the audience, if one even exists, is wrong. Much fundamental progress has been made this way: Faraday, Einstein, Jobs, etc. But of course many times one simply is wrong and refusing to see it means throwing years away, and whatever else with it (money, relationships, etc.). It's a hard balance, especially for the monomaniacal without much interest in balance. Finding out how to make solid (public, peer-reviewed, evidence-based, whatever) incremental progress towards the paradigm shift seems to be the way if one can manage.
It reminds me of when I was around 10 years old or so, maybe slightly older, and playing around with Turbo C (or maybe Turbo C++) on DOS. I must have gotten something very basic about pointers (which were new to me at the time) wrong, probably having declared a char* pointer but not actually allocated any memory, leaving it entirely uninitialized, and my string manipulation failed in weird and interesting ways (since this was on DOS without memory protection, you wouldn't get a program crash like a segmentation fault very easily, instead you'd often see "more interesting" corruption).
Hilariously, at the time I concluded that the string functions of Turbo C(++) must be broken and moved away "string.h" so I wouldn't use it. But even then I shortly after realized how insane I was: Borland could never sell Turbo C(++) if the functions behind the string.h API were actually broken, and it became clear that my code must be buggy instead. And remember, I was 10 years old or so, otherwise I don't think I would have come to that weird conclusion in the first place.
Nowadays, I do live in this very tiny niche where I actually encounter not only compiler bugs, but actual hardware/CPU bugs, but even then I need a lot of experiments and evidence for myself that that's what I'm actually hitting...
I’m sure we’ve all done it at one time or another, but repeated as habit without learning seems to speak of a certain kind of personality.
Sounds like a classic example of Moravec’s paradox:
https://en.wikipedia.org/wiki/Moravec%27s_paradox
It’s not that a shooting (or action, or heck talk to competitive fighting game players) game has less decisions for the player to make, it’s that the decisions being made are all subconscious decisions about movements and difficult to put into words.
What I mean by that is that there are game designers like Jonathan Blow who have their own theories on what is a great game and are extremely critical of the industry not following those theories, and then have released games that succeed at demonstrating those theories. In Jonathan Blow’s case, you can disagree with the man, but you can’t disagree with the fact that The Witness is a wildly original, successful game (1M+ copies sold) that has a cult following.
That does not seem to be the case for Crawford’s work. Lots of theories, lots of indictment for the industry doing it wrong, but no actual demonstration of what “doing it right” would mean.
Saying that no one gets it and civilization won’t be ready for many centuries (as the article I linked above does) feels like kind of a cheap rhetorical cop out.
For what it’s worth, I disagree with his indictment of the video game landscape as being narratively poor. Lots of video games with great interactive narratives out there, and there are many players who have been deeply moved by such games (of course, which games that might be varies from person to person).
I think a good antidote when one finds themselves in those thinking patterns is to listen to what others have to say, and not dismiss them as not getting it because they don’t follow your particular (unproven) theories.
I played it as a cold war kid and was fascinated by it. Mid 80's, post War-Games, this game blew my mind. It simulated the world.
The lesson I remember was that conflict in the Cold War was not zero-sum. One side would win and one side would lose. There were (in this game) no win-win outcomes. But - and this is the key point - the value of each win or loss was unequally felt. For the US to back down in Indonesia was disappointing. To back down in West Germany was fatal.
Oh - and also the notion of graduated escalation & de-escalation. Playing the game well requried using escalation wisely. Sometimes you escalate (a bit) to see how they respond & judge the value of a conflict to your opponent. Sometimes you escalate (a lot) to signal to your opponent that a given conflict is very serious to you.
I don't know if I ever had _fun_ playing the game - but of the hundreds of games I played as a kid this one stuck with me.
All this with something like 64k of memory - brilliant!
Maybe I'm misunderstanding, but it's not clear to me how this describes something obviously non-zero sum. Independent losses can have different values in a zero-sum hand game; what matters is whether each win is proportional to the corresponding loss. If the USSR winning in West Germany was only a small win, that would demonstrate it was non-zero sum due to the size of the loss there for the US, but I don't think the magnitude of the outcome in Indonesia would relate to that at all.
I first attended CGDC in 1994. It was two years after his Dragon Speech (which I knew nothing about), but the highlight of the conference was his talk about the challenges of story-based games. The part I remember is how he modeled stories as branching trees with fan-out, foldback, tree-of-death, etc (he covers this in the "Architecture" chapter of this book "The Art of Interactive Design"). I didn't really follow Crawford's work on Erasmatron or later, but by the late 90's it sounded like his story model had changed from a tree to a graph network structure, like a finite state machine. While it was an improvement, I was a bit skeptical that this model was enough. Nevertheless, I spent a lot of time thinking about the problem. You see, he'd already infected me with his dream of interactive storytelling.
By the time I moved to California and took a job at EA/Maxis on The Sims 2.0 team, I had decided that true interactive storytelling (as I saw it) was not possible until game AI was sophisticated enough to enable autonomous NPC chatbots. So I put that dream aside while I pursued a career in software development.
Here we are, over a quarter of a century later, and that AI technology is here now. For those of us who have been waiting for this moment, it is almost miraculous. It might be the end of an era for Chris Crawford, but it is just the beginning of the AI-based interactive storytelling era.
I'm "only" in my 50s now, but I am in my 50s now. For a couple of decades I've had this grand vision about how to build the perfect algorithmic trading system. The practical expression of this has come in fits and starts.
During the Pandemic, I finally devoted some time to building a bespoke server and hosting it co-located facility near a high-quality market data provider. I chose to do this before the software was ready in order to "light a fire under my a$$" and help motivate me to move things along. By the end of 2021 I had written, tested, and deployed a set of specialized clients written in rust to consume the market data and perform basic parsing of the real-time information. I stored the transformed output of these processes in log files as a temporary data sink that could represent the data "someday" moving downstream in a custom stream processing platform that would perform trades via an online broker.
And then, it stalled. There always seemed to be good reasons why I couldn't commit to one implementation strategy or another. Would I pass the data downstreeam via pipes? With Kafka? Via shared memory? Would the parsed input be represented via protobuf records? A custom binary format? Apache Arrow/Parquet format? Would the algorithms be expressed in custom rust code? Using Pandas(numpy)/Polars? Would I focus on my imagined "insights" into how prices move or make the transition to Machine Learning by trying to produce "special sauce" custom features that could be fed into models of various types? In the meantime, I have only collected a high volume of daily log files.
There has been some progress, of course. Last year, with LLMs as my pair programmer, I added vector processing to the ingest components, which made them run even more efficiently. I certainly enjoy the sporadic work I dedicate to working on the system. But there always seems to be a fog of possibilities ahead of me obscuring my view towards a broker API getting called.
My family and closest friends look upon this as the "hot rod in his garage" which I sometimes work on and which will probably never get finished.
Someone who clearly wanted to make a difference, but mostly seems to have not just made games.
He made game tools, but then didn't actually use them to make games. And then he blamed everyone else for not being ready for what he was making.
Giving up after only one released work just seems like such a shame.
I have watched the dragon speech multiple times, and it makes me feel like you should put everything in what you are doing, as everything is your magnum opus. I have huge respect for artists and artisans that work like that. It is really difficult, and often not economically viable, but there are people that just `see` something and they want it to exist.
I often think about the dragon speech, when I am doing the same 'pick up 10 things' quest in World of Warcraft that I have been doing since 2005. Thousands of times.. I have picked up hundreds of thousands of things :) and I wonder, what game would WoW be had he succeeded.
Some of his reflections on growing old, remembering his first crush, and even just noodling about home improvement are incredibly beautiful too.
Those are all asides, but what I mean to say is that his other posts are worth reading.
I explored tabletop games after that, modding, expanding, and even building prototypes using 3DP.
When AI exploded 2 years ago I was curious how much of a game it could build and I had recently learned about the existence of Pygame. Ended up building a zero-player 2D space sim. It's stupid, but I appreciate it and there is some art in it.
https://github.com/derekburgess/simcraft
Then recently, a coworker asked if I had tried Codex yet. I hadn't, and while riding in a car had the idea of using Codex on my phone to make a game that I could then deploy on my phone and play. Wanted to explore something meta, so I made an AI dungeon master in the style of Zork. It got out of hand and you cant deploy it on a phone- but I did deploy it remotely on my home workstation and then played over ssh from my phone ;P
https://github.com/derekburgess/dungen
I appreciate games for their randomness and have always leaned into games that incorporate generative sandbox experiences.
[0] https://www.youtube.com/watch?v=nMg2_teHVTs
[1] https://www.youtube.com/watch?v=ajz_1TqccYA
I was not familiar with Chris Crawford before this, though I think I'll look into him more. Reading his idea of People Games [1] I wish he was a younger man with a bit more time to revisit these ideas with new technology. There are new interactive mediums to discover with LLMs, and it's mediums that he's clearly been trying to create all this time...
Quoting his excerpt:
"I dreamed of the day when computer games would be a viable medium of artistic expression — an art form. I dreamed of computer games encompassing the broad range of human experience and emotion: computer games about tragedies, or self-sacrifice; games about duty and honor, patriotism; a satirical game about politics, or games about human folly; games about men's relationship to God or to Nature; games about the passionate love between a boy and girl, or the serene and mature love between husband and wife of decades; games about family relationships or death, mortality, games about a boy becoming a man, and a man realizing that he is no longer young; games about a man facing truth at high noon on a dusty main street, or a boy and his dog, and a prostitute with a heart of gold. All of these things and more were part of this dream, but by themselves they amounted to nothing, because all of these things have already been done by other art forms. There was no advantage, no purchase, nothing superior about this dream, it's just an old rehash. All we are doing with the computer, if all we do is just reinvent the wheel with poor grade materials, well, we don't have a dream worth pursuing. But there was a second part of this dream that catapulted it into the stratosphere. The second part is what made this dream important and worthy: that is interactivity.
"Let me explain to you why interactivity is so overwhelmingly important. Let me talk about the human brain. You know, our minds are not passive receptacles, they are active agents. It’s not as if we have a button on the side of our heads and they come along and push the button and the top of our heads flips open and then they take a pitcher full of knowledge and pour it into our skull, and then they close the top of our head, shake well and say, «Congratulations, you’re educated now!». [...] All the higher mammals learn by playing, by doing, by interacting [...]
"The interactive conversation is effective, but the expository lecture is efficient. That’s the trade-off we make. And over the centuries, we humans have learned that the gains in efficiency outweigh the losses in effectiveness. And therefore we choose expository methods. But the sacrifice remains real! We haven’t ever solved that problem. It’s been with us since the beginning of history. Every single artist has faced this, every communicator, every teacher, every novelist, every sculptor, every singer, every musician, every painter, every single artist through all of human history has been forced to sacrifice effectiveness for efficiency… until now. Because now we have a technology that changes everything. [...] That is the revolutionary nature of the computer. It allows us to automate interactivity to achieve both effectiveness and efficiency. That was the most important part of my dream."
[1] https://en.wikipedia.org/wiki/Chris_Crawford_(game_designer)...
I think he would say they are narratively poor by his defintion that the narrative must be generated by the game/player combo and not just pre-programmed. People love "The Last of Us" for it's narrative but that narritive is something that can arguably be conveyed via book or movie. Crawford wanted something where the narrative itself was generated.
And no, he wouldn't count the choices players make in the average game. Whether to get go west or east. Whether to get the a sword first or the arrow. He wanted the story and character dialog to change. Few if any games do that. Of course today with LLMs it's likely some games will soon / have already done it to some degree and will do better in the future.
Going back to his older work, you'd need to feed a context to the LLM about each characters motivations and then update that context based on player actions so that as the game progresses the way each NPC interacts with the player, and other NPCs, changes in a way that's consistent with each character's intrisict motivations and their interactions with others.
In the single player realm, there are games like Dwarf Fortress, Caves of Qud, Kingdom Come: Deliverance, etc.
Point is, the landscape of what "narrative" means in video games today is broad and deep. If none of those are even remotely like what Crawford thinks is "right" - and he's not able to design a game that meets his standards himself - I'd argue his definition of "right" might just not be workable in the first place.
There's a kind of people who want video games to have all of the possibility, depth, and meaning of real life. A game where you could do anything, be anyone, but still have consequences matter and be far reaching (like "Roy: A Life Well Lived" in Rick & Morty). Well, that exists, it's called real life, but you're not going to recreate it on a computer screen.
The broker API I plan to use features the ability to "paper trade" in a simulated account. So I'm not so worried about switching the thing on one day and immediately blowing my account.
I do trade manually in a broker account but as fun and instructive as that is, it doesn't represent the kinds of high-speed analysis and execution my system would make once it was completed.
Nevertheless, it is quite sad - however, it's difficult for me to relate to the experiences of someone who lived through that first wave of personal computing and played a notable part in it - perhaps through that lens, anything was possible.
[0] https://en.wikipedia.org/wiki/Chris_Crawford_(game_designer)
A whole person -- flattened into little bits gleaned from some text, glued together with assumptions and world-building -- dismissed as "blaming" and "giving up" "after one game"
The YouTube link in the other post has a top comment of "The best speech in all of gaming history delivered by what must be considered the Socrates of gaming.", to give you a sense of there may be more depth to this person than "giving up after 1 game".
If nothing else, it indicates the crowd perceives more depth, which will be enough to make you ponder if you missed something.
I suggest re-reading the article with a different set of assumptions -- when faced with a contradiction, first, check your premises -- it's likely the guy worried about declining programming skills and pointing out the ease at which he was dismissing JavaScript due to simple errors, is being self-aware and sarcastic.
Once you're freed via engaging with your own thinking, instead of rushing to do public judgement, it is a quite beautiful meditation of working on something that fails to get the mindshare you hoped for, and a all-too-familiar to all of us reminder of the cognitive dissonance required to be okay with that, even when you'll never be okay.
I appreciate your advice.
That sounds no different from comming up with narratives on a playground. That's not a designed narrative, that's people making up a narrative where none exists. Hey, the jungle gym in our secret base! The swings are space ship! The ground is lava!
I think we can all imagine what he wanted to build even if he failed to come close to building it. He wanted to make a story machine where you could play the game and converse with the characters in a free flowing way yet still have the game provide a setting and conflict. Imagine talking to characters in the holodeck on Star Trek. Ideally where, over the course of the game, the dialog and interactions are designed in real time within the constraints of the setting. And, the way you treat characters influences how they react. Be a dick to the bartender, all his connections are harder to get consessions from. Be nice to one romantic partner, get snubbed by another, etc... And not just by canned responses. Tell one character a piece of info and it gets leaked to their closest contacts who then change their behavior/dialog based on this new knowledge.
The difference is that Dwarf Fortress is not fully opaque to the player. You see a large part of the world so can see the consequences of your actions. And that is what makes it fun.
An opaque world simulator is frustrating and tedious. Eventually someone just min-maxes a way to break it which makes it just a tedious cookie clicker. That's why D&D has a game master that modifies the world's design to align with the players so that everyone has a good time. That is the key difference. An illusion of rigid structure, which requires some actual structure, with opaque flexibility. A simulator is not a fun game. A world where god cheats to make you have more fun time is a fun game. A pre-defined narrative or narrative tress is one such cheat, but not the only one.
I've been in too many conversations where this topic comes up, and it's very disheartening to me. Gamers insist there are plenty of great narrative games, and every example they give is basically a branching story with bunch of flags that gate which branches can be taken. If I give the Holodeck as a counter-example, well that's just too pie-in-the-sky.
These conversations remind me a lot of Paul Graham's Blub Paradox: "Blub is good enough for him, because he thinks in Blub." Current SotA narrative games are good enough for most gamers, because all they've played are branching story games.
If the argument is “let’s build games that tell character stories as complex as Last of Us with the world building techniques of Dwarf Fortress”, it might be worth considering why 1) it wouldn’t be feasible and/or 2) it wouldn’t lead to a game that’s as fun/compelling as a more narratively linear counterpart.
The narrative in Last of Us is compelling because it’s tight and focused and authored by the authors to be told in a very specific way.
The “emergent narrative out of raw world simulation” argument sounds like filming thousands of hours of security camera footage and expecting Citizen Kane to come out of it.
I mean again, I would be pretty receptive to a “this game had the right idea, we should push it further” argument, but the “I’m a visionary game designer who hasn’t shipped a game that embodies my vision AND the thousands of other game designers are clueless, but everyone will see how clever I was in 3 centuries” argument rings pretty hollow.
For what it’s worth, Chris Crawford’s tool (http://storytron.com/) is a branching story editor, and one that seems much less powerful than what is used in the industry (eg articy).
Crawford's latest software that he developed to illustrate his latest approach for interactive storytelling are called Slubber99 (https://www.erasmatazz.com/library/Slubberdegullion/Slubber9...) and Gullion99 (https://www.erasmatazz.com/library/Slubberdegullion/Gullion9...).
Storytron 1 & 2 history can be found at this link - https://www.ifwiki.org/Storytron
Obviously he's not serious, he's playing the part of the out of touch old man.
Tell us again when you're 74.
I'm still nearly 2 decades from it, but I am a profoundly different human to the one I was 20 years ago, or 20 years before that.
If you've ever been unfortunate enough to have to wrangle a VB script routine, it was (less bad) like that. If not, I would go find some assembly code and teach it yourself, and then imagine that instead of side effects in registers there were random effects on your code/visual state.
And like assembly code, you could now imagine that the same code might behave wildly different on different machines in different browsers.
So a bit of "old man"isms, but also I imagine his JavaScript was tainted by the early days. It's better in some ways now, worse in different ways, I don't mean to say that is the worst or the best, just to offer perspective on where it came from.
But apparently I didn’t get the hyperbole.
Only to then realize (after reading the walkthrough) that there was indeed a way.
I think it's human nature to find (rather search) blame not only in yourself but everywhere else... anyhow, since the author is reflective we should be forgiving as well.