The industry has experienced slowed growth and is restructuring which is scary for many people right now. There have been continuous layoffs and studio closures. That sucks (though is not all caused by AI).
But a few quick thoughts:
Video games have always been about cutting edge technology. Because they are interactive, they are best positioned to leverage AI tech. (unlike static media)
Prototyping tends to be the most important but most neglected process for finding the fun. AI is a catalyst for rapid prototyping such that a studio can more quickly build and assess game loops and de-risk the rest of their dev cycle before staffing up or pulling team members off other projects.
AI may long-term create more leisure time macro-economically for everyone, meaning more consumer time that may be consumed playing games. (Owen Mahoney thinks the industry will soon triple in size)
>Because they are interactive, they are best positioned to leverage AI tech.
You have not supported this argument.
On the contrary, I think LLMs are not a big help.
We've already been here. Procedural generation was the magical solution that was supposed to help a single dev make giant worlds. And it did. It helped you make giant, utterly empty and soulless worlds.
Great for minecraft. Not so great for No Mans Sky, as there were significant limitations. Useless for anything that depends on a story or immersion in that story or characters.
This idea that you can wire an in game character up to an LLM is misguided and doesn't seem to understand what players want a character to be.
In Mass Effect 2, a fan favorite character infamously had little content for most of the game. Garrus was a very loved character, including being the love interest for a lot of people who played the game, but during 2 he just sits in a part of your ship and says "Sorry, can't talk, running some calibrations" for almost the entire game.
Put the entire game's script into an LLM as context and have it pretend to be Garrus and try and talk to it. Will the LLM take a strong stance that Garrus would take? Will it correctly figure out that Garrus would rather kill a bad guy than let him get away, and then have him make convincing arguments about that in reference to the previous mission, and then in the next mission put you in a situation where you have to decide whether Garrus is right and whether you should dome that "Bad Guy" instead of letting the police fail to aprehend him? Will that Garrus make you think about this world and your place in it and whether your personal or chosen morals are right?
Probably not.
People don't want to chat about the weather with random NPCs. People want characters that have character notes and integrate into the story and make you feel.
So far LLMs can't really write that well, and certainly are not cohesive and multidisciplinary enough to be able to build that into a game in a convincing way.
There's a famous game called "Facade" which plays up this big fanfare about how it uses "AI" and Natural Language Processing to bring two characters to life for you to talk with and navigate a crisis with, but it's almost entirely lies. The actual logic of how it works is almost identical to old fashioned text adventure parsers in their heyday. It's a heavily scripted sequence, with fairly few actual paths it can take, and the script is not that big. It got so much press for what was basically done in the 80s. I think some people have tried to hook LLMs into it, but it just doesn't feel good. The problems that AI dungeon adventure game always had still exist, just under more paint.
One EGG studio is wiring LLM’s to npc’s in a hilarious way. We’ll be curious to see everyone’s feedback when they make a playable available.
However, I’ll say that as a result of their AI integration they are also doing way more human writing in the form of prompts and other procedural elements than if they just used old fashioned dialog trees.
I think AI can only be used as an enhancement in certain specific and controlled ways.
One mistake I see a lot right now is the assumption that you can delegate design and creative direction to AI. I think that generally yields slop. In fact I think the Creative Director/design role at a game studio may be the hardest digital job for an AI to replace. I had the opportunity to express that idea to Sam Altman once and he did refute it.
1. "More leisure for everyone" is utter bollocks. We know it doesn't work like that. If it did we would all be doing nothing but leisure because of how much leisure we gained by switching to email.
2. Games are just code that's fun. How does this sounds as a process for making something fun: "Start by de-risking." Hmmm OK, yes, this tracks with my experience of private equity companies being the most innovative and successful creators of games.
More leisure will only occur if it becomes true that the marginal return of doing additional work falls almost to zero. So to say people will have more leisure time actually suggests more that they will lack opportunities to do things of value than it does that they will choose to have more leisure time. Which is depressing.
If everyone could live a life of leisure, the logical end result would be roving gangs engaging in street warfare for funsies. If money is no object for anyone, what else is there to do but to seek fame, even infamy?
Each gang member could be running their AI value-miners at home, but of course since they're the only kind of value in the new AI-communist society they'll be the obvious target for the other leisure gangs. So after enough rounds of violence each leisure gang will run a fortified, paramilitary "intelligence mining" operation, and oh by the way indie software dev is punishable by death in these territories.
Is this scenario, like, 9 kinds of insane? Sure! But so is the idea that we'll all be at the beach doing Idunno what, fucking? All this to say yeah, I'm with you that anyone who describes that AI will make a future defined by a lack of productive work is describing a depressing future...
It bases this entire chain of assumptions on "AI is really big and everyone agrees".
But I don't see the games industry as all that vulnerable to AI at all. Game engines already drive constantly-improving dev efficiency through improved abstractions.
None of this is necessarily zero-sum. I'm skeptical AI is going to be a meaningful tailwind for games, but if I'm wrong it could absolutely benefit customers, studios, and labs through boosting productivity.
Making a game that will sell well on Steam is typically much harder than finding a bunch of boring business leaders and pitching them a SaaS or consulting package. On the surface it might seem simpler to do a game, but once you get into the mechanics of building, testing and publishing something for the masses, the fear of cold calling or emailing total strangers begins to evaporate quickly.
About 99% of the work you do on a game will wind up in the trashcan. Doesn't matter what kind of work it is. Code, audio, textures, models, map layouts, multiplayer balancing work, etc. are all susceptible in the same way. No one is safe from the chaos. It takes a lot of human energy and persistence to produce sufficient 1% content to fill up a player experience.
I'd estimate for a B2B SaaS product, the ratio is approximately the same, however you don't need such a broad range of talent to proceed. One developer with a desire to do the hard things constantly can be all you need to make it to profitability. Going from one employee to N employees in a creative venture is where things go bananas. If you absolutely must do an indie game and you need it to succeed or your internet will get cut off, you will want to strongly consider doing it by yourself. Figuring out how to split revenue and IP with other humans when you can't get the customer on the phone is a nightmare.
Many people agree that being a video game founder is harder than being a tech founder. Staying small and being as resourceful as you can is a good way to mitigate risk.
As a YC founder turned gamedev, I can tell you that failure is the norm in both pursuits. As with YC, the question is: can we create more outsized outcomes when we succeed?
I don't think that's possible and frankly I'd much prefer capital not even try. To the best of my knowledge, the only games which generate the really outsized outcomes you'd need for a VC portfolio do really gross, anti-player shit (gacha, lootboxes, whale fishing, etc.) to get it. Or they become distribution monopolies like Valve, which is fine-ish when Valve is private but would be a ongoing catastrophe if it had been VC-funded. I'd rather not encourage that.
I agree. I'm not the moral police, but video games ultimately have to walk a line where they serve up entertainment that is engaging/addictive without being all-consuming and abusive, and do so for an amount of money for which there is general consensus is "reasonable".
Trying to ride that to the moon is a very different proposition from a B2B play where you sell some service that concretelt delivers $X/mo recurring value to each customer for a $Y/mo price tag, and X > Y, but Y - your costs still turns a healthy profit. If you do that right, everyone is winning and the economy as a whole grows, not at all the same as the zero-sum game that is soaking a few whales and ruining their lives.
I appreciate this perspective.. but I think there might be a false dichotomy here. Some of the biggest gaming success stories didn't rely on exploitative mechanics - Minecraft, Among Us, even Fortnite's initial success was based on solid gameplay before the monetization kicked in. The question is whether you can build sustainable platforms that create genuine value rather than just extracting it. Steam takes 30% but provides real distribution value. Maybe the trick is focusing on companies that help other developers succeed rather than trying to create the next Genshin Impact
> Fortnite's initial success was based on solid gameplay before the monetization kicked in.
Fortnite is a bit of weird backwards example because the early PvE iteration had paid lootboxes, but they were scrapped in the Battle Royale spinoff which actually got popular, and eventually removed altogether. They still do things like engineering FOMO to drive sales but ironically the games monetization was the most exploitative when nobody was playing it.
Agreed. It is possible to make money in video games while treating customers with respect. This is the way! (and what EGG looks for because it builds stronger IP and longterm retention and good will)
I helped Indie Fund Hollow Knight back in the day and look a them now!
I think it is possible to be successful in games without being predatory. Humble Bundle was a demonstration of that.
I don’t believe that ultra-predatory mechanics are long-term sustainable. They usually yield a “ring of fire” effect that creates a growing ring of users for a while but really you’re burning out all your core users and will implode. This is how many describe the original Zynga model.
Supercell (founded around the same time) has cultivated longterm ecosystems and IP by respecting their players.
>I don’t believe that ultra-predatory mechanics are long-term sustainable
I don't think they're trying to be. I think they're whale hunting, find a few high spenders, and milk them for all they're worth. Then they spin up a new IP (or license one out), rinse and repeat
Yes they are. The implementation tends to differ from how eastern-developed gacha games work, but they're making billions from virtual slot machines nonetheless.
Yeah, I still indulge in video games, and understand that on the surface CSGO skins feel different than Genshin summons, but from 100 feet up it's all the same crap, imo
I do kinda get what you mean, though. Gacha mechanics feel expected in anything western, while 'loot boxes' are still a 'feature' of some games in the east. Eastern studios have definitely noticed, though, and are running the same playbook.
I see. You can remove 'Overwatch' from my original comment and the point still stands, but I do appreciate the fact check. I know Blizzard from HearthStone and Diablo....not great experiences with gacha there haha (Diablo Immortal, atleast), but those are far from the most popular Eastern games
You're right though. Industrial software/hardware in general always has money in all times. But gaming is essentially entertainment and people only spend on entertainment last. So gaming industry has a lot of failure but even if you're successful in a huge way, you won't earn huge money. There's a big cap there.
VC math follows a power law and expects almost all investments to flop, and the one winner to pay for it all. The question here is not about the flops it’s: are the winners big enough?
Isn't this the exact logic behind the stagnation in media right now? The risk aversion/stakes is so high that real risks aren't taken and we get bland sequel/remake slop.
Ironically when media culture is at is at its healthiest is when winners are diverse and common, and more importantly smaller shows that try out new things can still break even, with periodic flops being generally tolerable. That low risk culture for attempting new ideas is precisely what creates legendary franchises later when a few of these hit everything right.
And now that they're eliminating diversity in their investments are you still certain they will pick the next generation's winners? All they're investing in are AI companies...
Once upon a time someone like me for whom engineering competence is a core aspect of my identity would have never considered turning my back on YC. But now I'm just embarrassed by them. The things they now think are the only things worth investing in mostly make me want to vomit, like the vibe coding casino-IDE startup. As someone who still espouses their old values rather than their new ones, I'd rather succeed on my own.
Both Mark Andreessen and Andrej Karpathy say AI is unique as a new technology in that small teams and individuals seem to be the earliest and most cutting edge adopters. (unlike computers and the internet which were used by government and then large companies)
YC just so happens to invest super early in small teams.
So the overlap of YC and AI is inevitable. AI is not an investment genre per se but it can be used to accelerate or improve any ecosystem if used carefully and cleverly.
Since my Humble Bundle days, I’ve always been partial to small companies and small dev studios. Not all EGG companies use AI but they are all keeping tabs on the technology. Mitch Lasky has said that AI may have opened a window in which small studios may have their best shot at outsized success in recent history. Eventually the big dogs will catch up and adopt the new tech themselves but right now David has a shot at Goliath.
It's funny, because in my industry it's the slavish attention to AI by goliath-scale companies like Microsoft that is leading them to set fire to quality, innovation, and consumer trust, when then gives me and my tiny startup the opportunity to jump in and eat their lunch.
Over the next 10 years, no, I think the market will course correct within that time frame. AI is the sauce that's being slathered on everything right now and demand for it is driving record valuations, particularly for AI startups and their founders. That demand is all investor-driven though: investors are falling over themselves to make AI investments, while consumers are not actually especially eager to have all human contact progressively stripped from their lives.
> That demand is all investor-driven though: investors are falling over themselves to make AI investments
Largely true.
> while consumers are not actually especially eager to have all human contact progressively stripped from their lives
Hmm… I agree with this sentiment, but I think it’s mostly a straw man. There are many things that AI can do well that people will end up embracing directly or indirectly.
Medical scans is one big one, imho.
Mundane but important legal services is another.
Skillful mediation of scutwork is definitely embraced.
Good and fast simple customer service via phone or text will end up being very welcome (at least in some contexts). I realize that most people will prefer superlative human customer service, but that’s currently not a widespread available reality, especially for simple tasks.
All sorts of learning (great and essentially free tutors).
All sorts of practice (e.g., language, speeches, debates, presentations, etc.).
All of the above (and more) are things that people are using AI for right now, and they seem to be loving it.
I realize that some folks use AI tools in regressive and sometimes dehumanizing ways, but that’s not the fault of the tool, imho.
I dunno, I see problems with every one of those things.
You could make a customer service AI that was an advocate for the consumer, but it would likely spend the company's money liberally. So instead you'll end up with AI agents incentivized to be stingy and standoffish about admitting the company could improve, just like the humans are.
You can tutor with AI, but there's no knowing what it will teach you. It will sound as convinced of itself when it teaches you why the earth is flat as it does teaching you why the earth is round. The one thing it will certainly do is reinforce your existing biases.
You can practice with AI, but you'd learn more by posing yourself the questions.
A doctor can have AI look at medical scans, but they can't defer to AI judgement and just tell the patient "AI says you have cancer, but I don't really know or care one way or the other". So again, the skill in reading results needs to be in the doctor.
On some level it's the role of the publisher to pick winners and guide them over the finish line. See for example the hit machine that is Devolver: https://www.devolverdigital.com/games
I'm not an indie dev, but if I was I would happily give up a chunk of my potential profit to be listed on there, knowing the size of the market that says "oh yeah... a Devolver title, I would blind-buy this, it's probably pretty good."
I suspect games are like movies: for every 100 movies, around 20 of them make enough money to cover the losses of the other 80. But predicting which movies those will be is extremely difficult (Goldman's Law: Nobody knows nothin'). Any studio / label / distributor which can do better than the average is probably headed for greatness.
The thing is when a game flops it can be popular in the future. It is still a game that does game things. When $dumbapp flops though, it is a stronger signal that whatever it is trying to do may have no market at all. Like giving a fish a skateboard.
The game industry needs no such thing. You can make profitable games very cheaply, and business models are simple and well understood. It’s a matter of just making a good game and getting it good exposure to customers.
Video games are a subset of entertainment which is capped in TAM by the population the game reaches, the amount of money they're willing to spend per hour on average, and average number of hours they can devote to entertainment.
In other words, every dollar you make off a game is a dollar that wasn't spent on another game, or trip to the movies, or vacation. And every hour someone plays your game is an hour they didn't spend working, studying, sleeping, eating, or doing anything else in the attention economy.
What makes this different from other markets is that there is no value creation or new market you can create from the aether to generate 10x/100x/1000x growth. And there's no rising tide to lift your boat and your competitors - if you fall behind, you sink.
The only way to grow entertainment businesses by significant multiples is by increasing discretionary income, decreasing working hours, or growing population with discretionary time and money. But those are societal-level problems that take governments and policy, and certainly not venture capital.
Interestingly, Steam’s first hardware product was a Steam Box: a little computer brick that could boot Steam on linux and let you play all your games on your TV (with a Steam controller or game controller of your choice). The cycle is now complete.
The history is a bit more complicated than that. Valve themselves never released a “Steam Box” that could run games on Linux. They partnered with a few different companies (Alienware, Gigabyte, etc.), who released co-branded “Steam Machines” which were just those companies’s normal hardware design, but with a common set of specs ideal for running SteamOS. See https://en.wikipedia.org/wiki/Steam_Machine_%28computer%29
You might instead be thinking of the Steam Link, which *was* produced by Valve, and *was* a tiny little brick that let you play games on your TV. But the Link wasn’t running the games itself, it was streaming them from a dedicated PC (which may itself have been a third-party Steam Machine) elsewhere in your home. https://en.wikipedia.org/wiki/Steam_Link
> "which were just those companies’s normal hardware design"
For Alienware (not sure about the others, but AW was Valve's lead partner on it anyway) you're right in that it was a computer designed by Alienware not Valve, however it was a) very different to other Alienware PCs, and b) Valve were genuinely part of the development process, they didn't just say "hey make a small computer". They also shipped with the first gen of Steam controllers, which were created by Valve themselves. (Unfortunately, due to delays with SteamOS, the first version of the AW "Steam Machine" actually launched running Windows only, but with the Steam Controller, because Alienware weren't willing to delay their launch further and instead developed their own controller-based UI for Windows in a rush job...)
(Source: me, I was in the loop on those goings on at the time.)
To this day, I think the Alienware Alpha (as the Windows version got called) was one of the nicest machines Dell ever made and one of the best small PCs I've ever seen.
> first gen of Steam controllers, which were created by Valve themselves
The best goddamn controllers ever made, too, I still have one in a box somewhere around here and I won't use it because it's so awesome I don't want it to break. Pretty dumb eh? The two touchpads were the absolute best, I've never had control like that in an FPS and to this day I can't play any FPS with a stick because I was ruined on the Steam Controller.
It's the absolute worst controller I've ever used (Joy-Cons are a close second).
I didn't use it because I hated it so much, put it straight back in the box after a few days of trying to use it and eventually I sold it for as much as I paid, and I was glad to see the back of it.
The size/shape of the controller as a whole was fantastic but I just really hated only having touchpads instead of sticks, and that made it unusable.
I struggled with keeping my thumbs on the pads properly, probably needed sensitivity adjustment too but I just couldn't get myself to _want_ to try harder with it.
Sticks work really well for me in terms of controllers, I just wish we weren't being shafted by stick drift all the time when manufacturers could be using hall-effect sticks; I've got 2 PS5 controllers and 3 PAIRS of Joy-Cons with drift, while I also have my original 2 x PS1 controllers from my childhood, neither of which has drift.
100% agree. I still can't believe it didn't become a long running product that everyone uses.
I've never been a console/controller gamer, but I remember the first time I saw it - it was an early fake, that looked and felt very similar to the end product but inside it just had a metal weight and no actual electronics, and it seemed like such an exciting product, it was genuinely hard to keep that secret...
(And come to think of it, I think I misremembered when writing my last comment - I might be wrong, but I think the Alienware Alpha that launched with Windows actually shipped with an Xbox controller, and the steam controller was only available once the proper Steam Machine version was out. Not 100% sure, maybe we just needed the Xbox controllers for the press sessions before launch...)
Close in a lot of ways, better in a few ways, like haptics. Steam Controller haptics are not great, and the physical click is loud and echoes within the controller. Deck haptics are fantastic, but there's something about the large circular trackpads that feels better. Maybe it's just the larger touch area.
Perhaps I am unfairly lumping Alienware in with some of the other Steam Machines, which very much did look almost identical to their manufacturer’s other PCs at the time.
As someone who at the time was VERY into buying the “console-like” PC gaming experience that Valve was seemingly selling, I remember being pretty disappointed not just by the SteamOS delays, but also how much most of the Steam Machines still basically looked, to me as an uninformed buyer, at least, to basically just be a different SKU of their regular lines rather than the true “Steam experience” that I was hoping for (and which the Steam Deck eventually delivered).
I still use my steam link all the time. I have it fiber back hauled to the computer that it runs off of. I'm thinking of buying a couple more. Give one to my kid and one put on a projector so I don't have to keep moving it back and forth.
Also, I think the device you're looking for is a deck because you can plug that into a television and use a wireless remote with it.
The steam link is the best remote display device I've ever used. No frame drops or artifacting, even on scenes that make the 3090 chug. It forwards controllers to the PC.
Now, the software, "big picture mode" and otherwise using a controller for PC input aren't the greatest, but you gotta figure it's me and like 2 other people still using this.
BTW airscreen/miracaat/screen mirroring/"wireless display" all suck. If your TV has smart bullt in that supports miracast, that in my limited experience is the second lowest latency, then firetv devices, and then roku and everything else. Roku only usable for presentation or digital signage, unless first party built in.
I have used an old sony bravia tv to cast COD from an android phone. Every time I connected, latency varied from 100 - 200 ms to 10 seconds. I had to reconnect several times until the latency was satisfactory.
Conclusion: it has been technically possible to cast to a tv for some years.
yes, miracast/screencast whatever was a thing prior to the Steam Link being released in November of 2015 (9 years and some change ago). some of the current devices can actually do sub-100ms of input latency, but you can't be in the same room as the source device or you'll go crazy. The roku stand-alone have the worst network and input latency, they're unusable for anything other than presentations.
firestick was <100ms network and barely noticeable input latency (on the order of ~20ms so interframe lag at 60fps). steam link is link latency + some small constant - whatever the "frameserver" processing takes, call it 3ms but definitely <10ms - and that's both network and input.
when i said network i meant both the network and the actual refresh of the screen. watching a movie is one thing, but pushing "Y" and your character jumping should be "as instant as practicable" and steam link is the only one that is that that i've used, so far.
I guess Proton/Wine/Linux gaming wasn't mature enough back then. Also a handheld wasn't really an option because there weren't any powerful enough yet energy efficient and cheap x86 chips available either.
Proton didn't exist yet, IIRC. The Steam Boxen relied on devs/studios/publishers being able and willing to port their games to Linux natively. The result was a handful of AAA and indie games that put proper effort into ports that ran well, a modest but larger selection of AAA games sloppily ported (such that they often can't run on current distros without containerization or extensive library preloading shenanigans), and a deluge of indie shovelware / forever-in-early-access vaporware produced by clicking the "gib me Linux" button in Unity and calling it a day.
Unfortunately, while it was certainly a boom in the number of games Linux users could play (easily enough for me to ditch Windows entirely and game exclusively on Linux and consoles), it wasn't quite the critical mass needed for Steam Boxen to be a commercial success. Proton was the missing piece.
There were many years between the Steam Controller (2015) and the Deck(2022). I’m not sure how much of those learnings could still be applied, considering changing technology as well as staff changes.
Maybe Valve has such excellent staff retention that this would be a non-issue. At my previous employers, two years would have been enough to have to start over from 0.
PG does say he actually looks for “naughty” founders as one of the key filtering traits. I link this essay in my post.
But you still have to be ethical and do what you say or it won’t be possible to grow your business long term.
The gamer in me wonders if the “ideal founder” can be described as “chaotic good”?