widedub's recent activity

  1. Comment on Thoughts on AI in game development [from the CTO of PH3] in ~tech

    widedub
    Link
    Im reproducing the text here after being frustrated that I couldnt read it on my phone where I dont login and refuse to use new reddit. Apologies if I missed anything, I recreated the markup by...

    Im reproducing the text here after being frustrated that I couldnt read it on my phone where I dont login and refuse to use new reddit. Apologies if I missed anything, I recreated the markup by hand


     
     
    Given that there are news posted and widely discussed almost every day that have some relationship to AI (including but not limited to insane C-suite statements, interesting mod and reverse engineering projects, layoffs, cloning accusations, opinion pieces, fears, and fiction), and many of them cause a lot of discussion, I thought some people might be interested in this.

    Before getting into it, some context / disclaimers:

    • I'm an enthusiast gamer and modder and have been for over 20 years
    • I'm the CTO of PH3 -- we've been doing generally highly regarded porting and optimisation work in games for almost 10 years now
    • I have a PhD in computer science, and do research in HPC (including a bit about LLMs in that field)
    • None of this is written by AI or even touched by it in any way -- I enjoy writing and I don't enjoy reading typical LLM prose
    • By no means do I want this to be taken to even remotely represent "the" industry voice. To make clear just how ridiculous that would be: I don't even represent PH3, because while we only have ~5 people in games that still means that there are diverging perspectives on this topic
    • I'm including all the above here, despite really not being a fan of lengthy self-introductions, because everyone's background shapes their biases and with a topic as divisive as AI I think it's important to be transparent

    Baseline, Actual Capabilities & Shortcomings

    First of all, I want to go on the record that, at this point in time, ideas such as "GenAI doesn't actually increase productivity" are silly. I don't know if they are still espoused because people genuinely believe them, or merely because they want them to be true. If the latter, then I certainly understand and sympathise, but we have to deal with reality rather than wishful thinking. If the former, then they are simply not sufficiently informed about the capabilities of agent systems with frontier LLMs in 2026.

    At the same time, people on the other end of the spectrum seem to overestimate the power of these systems. I have a very recent real-world example here. Earlier this week, I tasked a coding agent with implementing the detection of a specific pattern of actions for recognising a new achievement in a game we're currently working on. I'd say this was actually a pretty good and suitable task for an agent: the idea arising from playtesting was well-documented, it's quite self-contained, and doesn't require complex decision making or design. Implementing it does, however, require finding the right points to hook and some relatively routine housekeeping, so it would require quite a bit of manual work.

    As one would expect, the agent correctly implemented the functional requirements for this, and tested them. However, when reviewing the changeset, I noticed that it introduced a full iteration over up to ~11000 game entities every simulation frame, to check the achievement conditions. The same outcome could be accomplished with slightly smarter integration by a small extension of an existing processing step, and an additional temporary storage of, at most, 5 elements. This isn't something that would appear noticeable in any playtest, because while I'd estimate that it would consume at least 100x more CPU time than required, it's still not enough to, on its own, substantially slow down the game to a noticeable degree. But imagine the same thing repeated across a few dozen features and implementation choices, and it will.

    To be clear, this wasn't some "weak" model -- depending on which benchmark you choose to look at it places #1 to #4 in terms of coding capabilities across everything that's currently out there. In my experience -- and here I can actually really say our experience across the whole company, using different state-of-the-art tools -- the type of higher-order reflection on outcomes required to recognise something like this pro-actively simply does not currently happen unprompted. "Unprompted" is important in this context, since of course the agent was perfectly capable of implementing the much more efficient variant when I wrote ~2 sentences explaining the issue and roughly laying out the better approach. I still didn't have to manually find the right source locations and entry points or describe implementation details.

    This is just one anecdote, but I do think it reflects the overall state of these systems pretty well. In short, they are a tremendous productivity booster, but relying on them completely without review / experience can have very sub-optimal outcomes, even in isolated tasks. (And that's before going into maintainability -- I went with this performance example since I think the issue is perhaps a bit more directly understandable for non-programmers)

    Specialities of Games Compared to Other Fields

    One thing that I think few might realise is that games are, at least for now, actually less directly affected by AI than other software development tasks. Obviously the focus here is on discussing those effects on games, and as a consequence it really might not seem like this is the case. But it's really not too complicated. The primary reason is that agent systems thrive when there is (1) a fast feedback loop, and (2) a clear objective function to measure.

    In many fields in software development, those are actually easier to establish than in games. Regarding (1), especially real-time games really are not something current agent systems and harnesses are particularly well-equipped to deal with. It can of course be done, but at tremendous cost and low speed compared to e.g. observing the outcome of a web query, measuring the efficiency of a simulation, or verifying a business process on a database. And regarding (2), that is even worse, since e.g. something like "good-feeling gameplay", which is often also a function of code, is really hard to quantify.

    This also means that some types of development in the gaming space -- e.g. ports, clones, and reverse engineering -- are inherently more suitable for current agent systems and will make faster progress at the same cost. And I think we clearly see this in practice right now.

    As a side note, I think this feedback loop / objective function issue is actually why, despite being large language models, I personally think they are actually more effective in (particular parts of) math and programming than "simply" writing prose. For both, you can measure and prove at least some objective qualities, and iterate on them. That's not generally the case for writing.

    Ethics / Code vs. Other Components

    A lot of the discussion about AI isn't actually about capabilities, but about ethics. I'm not going to go into much detail on that, because it's not on topic for this subreddit, and because I'm also not an expert. That said, I believe that the biggest societal issues we face due to AI aren't fundamentally AI issues but capitalism issues, magnified. The fact that data centers don't pay for their externalities or that it's even allowed to build them with fresh water evaporative cooling is a policy failure and not an inherent feature. And that most people are and have been required to continue working 40+ hour weeks for essentially stagnant remuneration since the 70s despite massive increases in productivity just so that very few people can become extraordinarily rich and powerful is a distribution problem.

    There's one ethical point I do want to discuss, that is related directly to AI specifically in games. A currently broadly accepted perspective amongst the gaming public (and I'm not saying that this is true for everyone, just that I frequently see it; also, "gaming public" here is of course just the small part of it that cares at all about any of this) appears to be that any level of GenAI utilized for any type of graphics or sound asset should be immediate grounds for a boycott, while all GenAI use in any software development position is basically fine.

    The primary argument for this (and, as I'll discuss, the only one I believe is at least partially valid) is usually that apparently everything related in any way to programming -- even high-level software architecture -- is completely devoid of creativity and a purely mechanical endeavour, while anything that places pixels in an image -- even, say, creating a generic background texture in a photorealistic game with clear constraints -- is artistic expression which must be protected. Now, I don't dispute that one tends more towards art than the other, but I don't believe that the average entry-level artist on an AAAA project involving thousands of people really has more room for expressing themselves than e.g. a lead software architect. Again, I do believe that the average tendency certainly does go in that direction, and I fully expect this to be the most controversial part of my post.

    But the art argument is really only a part of what is usually used in AI criticism regarding games, and all other aspects, in my opinion, cannot truly discriminate:

    • the "intellectual property argument": these systems were trained on vast amounts of data without regard for copyright or compensation --> this applies equally to e.g. images and code.
    • the "social argument": their use clearly displaces workers --> this applies equally to e.g. images and code.
    • the "environmental argument": they require lots of natural resources --> this applies equally to e.g. images and code.
    • finally, the "output quality argument": it makes my games worse --> I think even this applies roughly equally -- if e.g. image generation would be so bad at this point that it invariably creates an obviously terrible result, then people wouldn't (rightfully!) insist on labelling. And as I outlined above, extreme "vibe-coding" can absolutely have negative impacts on the final product.

    Now, this is empathically not a call to boycott all games that use any amount of LLM support in their code. It's also not a call to just accept games with art made via generative AI. But it is a call to maybe be a bit more nuanced in your assessment and less binary about the topic. Maybe an indie developer increasing the variety in their non-essential background assets via GenAI isn't the worst thing that ever happened to the medium, and maybe a highly profitable AAA publisher replacing 80% of an engine development team with Claude or boasting about vibecoding is actually not great.

    The Gamer / Enthusiast / Consumer Perspective

    All of this was really more from an industry perspective. I do have a few more points I want to bring up that look more at what I think this means for consumers.

    Let's celebrate the positive outcomes

    With all the negativity about AI in games, or the wider-reaching impact of AI on hardware prices (including gaming hardware), I do think it's important to at least celebrate some of the happier effects of this. For example:

    And hey, if new gaming hardware is unaffordable, maybe we just need to focus on efficiently using what we already have.

    We really should be expecting more, not less

    Recently, there has been quite a bit of discussion on (presumably, and I think there are good reasons to presume so) LLM-assisted official porting projects with very middling overall quality. I don't think such discussions should be primarily about blaming AI. Think about it like this: as I outlined above, I believe it's fair to say (and maybe even an underestimation) that for many tasks involved particularly in porting, efficiency can now be increased by at least twofold. So for a port to turn out worse, what it really means is that a publisher somewhere decided to spend less than 50% of the effort, and call the result good enough to sell.

    If anything, given these efficiency improvements, the fact that no one really seems to care to prevent AI use in development, and the fact that games really haven't become cheaper, you should be more critical of technical issues or incomplete feature sets than ever before.

    Conclusion

    So, I've been thinking about all of this for a while. I hope at least some of it was interesting. I realize it's long, but I didn't even get to every point I wanted to cover. You can now scream at me :P

    13 votes
  2. Comment on Best way to setup email forwarding for a domain I own? in ~tech

    widedub
    Link Parent
    I've seen a few people say this but can you expand on why? I have a few domains on name cheap and am using their email service but have been thinking about jumping to porkbun with proton handling...

    I've seen a few people say this but can you expand on why? I have a few domains on name cheap and am using their email service but have been thinking about jumping to porkbun with proton handling my custom domain email addresses

    1 vote
  3. Comment on HYPERYuki: Snowboard Syndicate | Demo trailer in ~games

    widedub
    Link
    This may be a bit too flashy for me but it looks fun. That said, I had the pleasure of working with Luis on a separate project pre covid and he was a true professional. I don't closely follow his...

    This may be a bit too flashy for me but it looks fun.

    That said, I had the pleasure of working with Luis on a separate project pre covid and he was a true professional. I don't closely follow his career but am really glad to see him continuing to get work

    4 votes
  4. Comment on Machinarium 2 | Announcement trailer in ~games

    widedub
    Link
    Not much to add to what has already been said but I also loved the original and other Amanita games have been excellent. I posted a little while ago about how a used PS2 and some games broadly got...

    Not much to add to what has already been said but I also loved the original and other Amanita games have been excellent. I posted a little while ago about how a used PS2 and some games broadly got me into gaming. Machinarium is arguably responsible for getting me into PC/Steam gaming . It ignited my love for puzzle games, the art style is memorable and the soundtrack has spent a lot of time as my background coding music. I own the original on multiple platforms and have completed the game several times, including helping friend and family through parts if not the entire game.

    All of this to say that I'm greatly looking forward to the sequel

    3 votes
  5. Comment on What single episode of a show would you recommend as a standalone? in ~tv

    widedub
    Link Parent
    We've watched so much with my nephew that my partner jokes I can name every Bluey episode from the first 2 seconds

    We've watched so much with my nephew that my partner jokes I can name every Bluey episode from the first 2 seconds

    1 vote
  6. Comment on What single episode of a show would you recommend as a standalone? in ~tv

    widedub
    Link Parent
    No worries, I just grabbed the titles from your links. It was super early when I saw your comment and I read it several times trying to figure out what the episodes were. I was pretty sure knew...

    No worries, I just grabbed the titles from your links. It was super early when I saw your comment and I read it several times trying to figure out what the episodes were. I was pretty sure knew the Euphoria one but you left me guessing on the Black Mirror one :)

    1 vote
  7. Comment on What single episode of a show would you recommend as a standalone? in ~tv

    widedub
    Link Parent
    That wasn't the prompt. OP asked for an episode that stands on it's own, even if the viewer never watches another episode in the series. Most of the series focuses on Bojack's personal growth (or...

    if you only watched this episode you would have absolutely no idea what Bojack is even about.

    That wasn't the prompt. OP asked for an episode that stands on it's own, even if the viewer never watches another episode in the series. Most of the series focuses on Bojack's personal growth (or lack thereof) and requires an understanding of the wider plot. Fish Out of Water can be watched and understood without any prior knowledge.

    I considered others like Free Churro here but even that arguably relies on some familiarity with Bojack/Beatrice's relationship. Is there an episode of Bojack that you think fits better?

    5 votes
  8. Comment on What single episode of a show would you recommend as a standalone? in ~tv

    widedub
    Link Parent
    "This is what a cornfield looks like, honey!" "mmmmm" It's so stupid but it's one of my most memorable lines from that era of the Simpsons

    "This is what a cornfield looks like, honey!"
    "mmmmm"

    It's so stupid but it's one of my most memorable lines from that era of the Simpsons

    2 votes
  9. Comment on What single episode of a show would you recommend as a standalone? in ~tv

    widedub
    Link Parent
    Euphoria - Fuck Anyone Who's Not a Sea Blob Black Mirror - San Junipero (also mentioned here) for anyone who don't want to click through for the answer

    Euphoria - Fuck Anyone Who's Not a Sea Blob
    Black Mirror - San Junipero (also mentioned here)

    for anyone who don't want to click through for the answer

    5 votes
  10. Comment on What single episode of a show would you recommend as a standalone? in ~tv

    widedub
    Link Parent
    "Nosedive" (S3E01) is my pick if we're including Black Mirror/anthologies. I love San Junipero as well but the social credit score episode is one that I think people can see themselves in the...

    "Nosedive" (S3E01) is my pick if we're including Black Mirror/anthologies. I love San Junipero as well but the social credit score episode is one that I think people can see themselves in the Bryce Dallas Howards character with minimal effort. I especially love the optimism of the final scene

    2 votes
  11. Comment on What single episode of a show would you recommend as a standalone? in ~tv

    widedub
    Link
    Bojack Horseman - Fish Out of Water (S3E04) This episode probably works a little better when you have the entire context up to this point of season three but it's strong enough to stand on its...

    Bojack Horseman - Fish Out of Water (S3E04)

    This episode probably works a little better when you have the entire context up to this point of season three but it's strong enough to stand on its own. It has a self-contained narrative with plenty of gags and wit but also hints at the stronger emotional themes of the series. The setup gives you enough about the main characters to go in blind and the ending, while perhaps a little predictable, is a fantastic payoff. One of my favorite single episodes of any show

    10 votes
  12. Comment on What is your dream game? in ~games

    widedub
    Link
    I've always wanted an RPG style game set in the modern day. Something like what I imagine Earthbound was like back when it was released. You and your party are kids from a suburban neighborhood...

    I've always wanted an RPG style game set in the modern day. Something like what I imagine Earthbound was like back when it was released. You and your party are kids from a suburban neighborhood and your weapons and stuff are all things you'd find in normal day. No magic in the traditional sense, maybe some basic science or chemistry to craft things that explode, burn, etc. The villains can be regular humans but maybe there's something that's driving them crazy, like a toxic spill. South Park:Stick of Truth sorta scratches the itch but I want a game that takes it a bit more seriously and can't rely on years of established story

    I'm not the biggest gamer so hoping for some suggestions toward existing games :)

    4 votes
  13. Comment on What have you been eating, drinking, and cooking? in ~food

    widedub
    Link
    After years of instant ramen packs, I finaly worked up the courage to try making "fresh" ramen. I went fairly simple: a mixed chicken/beef stock base, white soybean paste, chili paste, aromatics ,...

    After years of instant ramen packs, I finaly worked up the courage to try making "fresh" ramen. I went fairly simple: a mixed chicken/beef stock base, white soybean paste, chili paste, aromatics , etc over noodles from my grocery international section. Pan fried some chicken thighs, deboned and sliced on top

    It was quite good! If I had skipped/pre-prepped the protein, it would have been very simple and I could easily see myself trying this again in various permutations. The chili paste adds a nice full kick without being hot for heat sake. It was definitely a step above the instant stuff, even my favorites like shin black. I'm obviously not a connoisseur it I was impressed with how easy and delicious it was

    2 votes
  14. Comment on What have you been listening to this week? in ~music

    widedub
    Link
    New album from Bonobo - Distance in Static was released today. Just settling in for my first listen now with high hopes

    New album from Bonobo - Distance in Static was released today. Just settling in for my first listen now with high hopes

  15. Comment on Little late but...anyone interested in fantasy football? in ~sports.american_football

    widedub
    Link
    I'm down if we play on sleeper!

    I'm down if we play on sleeper!

    2 votes
  16. Comment on LG smart TVs caught logging audio with screen off and snooping on local devices in ~tech

    widedub
    Link Parent
    I use an old school htpc setup on two different machines An old Mac mini running Linux mint An old minisforum They work great for streaming and the minisforum can handle some light steam gaming. I...

    I use an old school htpc setup on two different machines

    • An old Mac mini running Linux mint
    • An old minisforum

    They work great for streaming and the minisforum can handle some light steam gaming. I use a basic Logitech keyboard with an integrated touch pad for control. We put all of our frequently used apps on the desktop and cranked up the desktop icon size so that they're easily viewable from distance.
    I never made the jump to services like plex or jellyfin, we largely use Kodi for media consumption

    10 votes
  17. Comment on What games have you been playing, and what's your opinion on them? in ~games

    widedub
    Link
    My nephew has recently discovered mods in Stardew Valley and so we're diving back into the game before the coming 1.7 update. We've kept it pretty light to start with an existing save and no new...

    My nephew has recently discovered mods in Stardew Valley and so we're diving back into the game before the coming 1.7 update. We've kept it pretty light to start with an existing save and no new content mods. He's enjoying CBJ Cheat mod, Tractor mod and some of the QOL picks we made together. I think he had as much fun shopping for mods as he did playing the altered game. Our approach is going to be to try out a whole host of mods on his old save game, see which ones he likes, then choose one of the content expansion mods and start a new game in tandem with his selected improvements.

    My only non negotiable point is that our next playthrough must include Jorts and Jean

    1 vote
  18. Comment on What are you reading these days? in ~books

    widedub
    Link
    I missed the last thread or two but I borrowed and finished Klara and the Sun by Kazuo Ishiguro. This was my first exposure to Ishiguro, largely driven by wanting to read the book before the film...

    I missed the last thread or two but I borrowed and finished Klara and the Sun by Kazuo Ishiguro. This was my first exposure to Ishiguro, largely driven by wanting to read the book before the film release that finally escaped development hell and is coming out this October

    This book is about Klara, an Artificial Friend (AF) and her life ranging from a display model in a store to being purchased by a family and all that comes after. This book is very approachable. I devoured a section/chapter per day and never found my desire to turn the page wanting. The sections are well constructed and give you a logical place to pause and reflect on the books events. I would recommend this book

    I loved this book! Klara was enjoyable to follow and, while I wouldn't call her an unreliable narrator exactly, I did catch myself wondering just how much her knowledge or limited knowledge tinged her interpretation of people and events. I was really impressed with how human Ishiguro makes Klara felt despite many reminders that she an android and I think her experience from the store to the final scene is a pretty obvious allegory for a human life. I think one of the questions Ishiguro wants us to ask after reading this book is, what makes a human? Klara may be an AF but seems to run a wide swath through the human experience...she learns, she sacrifices, she ruminates on her experiences. She's not human but where do we draw the line?

    I found making predictions as I read the book

    Some of my predictions while reading (whole book spoilers)
    • Capaldi would be a sexual predator
    • Klaras dad would be a deadbeat
    • Klaras sun-based faith would be shattered
    • Josie will die and Klara will become the new Josie (probably the obvious one)

    and nearly every one was wrong. I'm not sure if this is an Ishiguro calling card but the misdirection was masterfully subtle and I never felt cheated when the book took a turn I didn't expect. Ive caught up on some of the major criticisms of this book, like Klara should have been programmed with wide knowledge before leaving the factory but I never felt like any of the gaps took away from my enjoyment during reading. I found the ending a little sad but also peaceful. It never occurred to me that AF would age out as quickly as Klara did (maybe I missed a key line in the book) and I hated how she was discarded but I found the final encounter in the book to be a fine closing point for the novel in a way that wraps up the human life allegory I mentioned above.

    I'd love to discuss Klara further with anyone else who has read it. It really made an impression on me and I almost certainly will pick up a copy for my bookshelf when I happen across one. I'm fairly interested to try more Ishiguro with the perhaps obvious The Remains of the Day being near the top of my list. Depending on which is available from the library first, next for me is likely Ursula K Le Guin's Left Hand of Darkness or John Williams' Stoner, both of which come highly recommended

    3 votes
  19. Comment on Git at any scale in ~comp

    widedub
    Link Parent
    Not trying to be reductive but this is largely where I landed as well. I enjoyed reading about the challenges of running git at scale and how GitHub engineered around them. I liked hearing about...

    Not trying to be reductive but this is largely where I landed as well. I enjoyed reading about the challenges of running git at scale and how GitHub engineered around them. I liked hearing about how cursor has tried to increment on the pitfalls of 3pc/spokes and engineer a better system. But entrust Elon with my code, both enterprise and personal? No thanks. I'm sure someone will point out how silly that is when MS is the next best option but I'd rather move away from these services and try something like Gitea

    8 votes
  20. Comment on What are you reading these days? in ~books

    widedub
    Link
    I finally made it to the top of the wait list to borrow Klara and the Sun by Kazuo Ishiguro from my local library. I've only just finished part 1 where Part 1 spoilers Klara is chosen by Josie and...

    I finally made it to the top of the wait list to borrow Klara and the Sun by Kazuo Ishiguro from my local library. I've only just finished part 1 where

    Part 1 spoilers Klara is chosen by Josie and purchased by her mother
    I am intrigued so far. This is my introduction to Ishiguro. I've heard that Klara is perhaps not the best introduction to him as an author but I'm hoping to finish it at least once before the move comes out in October

     

    Just before this, I read The Mouse that Roared by Leonard Wimberley. It was an easy read and I very much enjoyed it. I did not love the ending where

    Final act of TMTR spoiler The Q-Bomb was revealed to be a dud. I liked the ambiguity and I don't think the book would have suffered without the final scene
    but it did not really detract from my enjoyment of the book. I'm interested in the Peter Sellers film based off this book and am hoping to check it out some time
    3 votes