When I say that in such fast times everyone is naked, no one knows how to act, it's easy to see who's actually thinking, and so on, this recent post and the wildly negative response to it by AI twitter are a great example:

A severe misalignment of AI in mathematics

The mathematicians are making an obvious and clear argument, perhaps you need to understand some of how math works to fully grasp it, I was a math major for a year so maybe what seems clear to me is not to others, although I doubt that's the difference, but let me still make an analogy to the field I know best.

The reason why the one-shotted vibecoded games generate an instinctual aversion is because anyone who has made games knows that, in the process of making a game and playing it, you often discover that ideas you had for it were bad, and that the game can have a life of its own, many devs mention that one or another mechanic that later became central happened as an accident, for instance. This often happens because lots about a game comes from the development process itself, and while one-shotting things gets you a result, it might miss better results that a proper exploration of the design space would have offered. The job of a game designer is to explore that design space, which the tools make easier, but which they shouldn't substitute because games are made for other humans, "fun," or however you want to codify it, comes from humans mapping a previously non-existent piece of reality into something digestible to other humans, eventually robots might be able to do this too, but they can't do it now.

And they can't do it now because they seem to lack the ability to ask the next good questions. There's a recent example of this with me in the Reverse Minesweeper Prototype Astra/Fable logs, where I keep asking both of them, what do you want to do next, as I wanna know what useful suggestion limit they have, and both of them reached it as soon as the main task was done. They simply don't know, given an open-ended situation, how to move ahead productively unless prompted, because this requires them to have objectives that they can't self-generate, or maybe they can, but they can't rank these generations in a productive way, or maybe they just don't have enough compute to do it productively. I don't know, but whatever proactivity is missing, they don't have it.

The analogy to mathematics is direct. Without proper understanding you lose the ability to ask useful next questions, therefore one can see how AI companies strip mining the field for PR is positively damaging, as they're depleting the field of one of its resources and not producing insight in return, insight that might lead to better theories, concepts, or just direct applications to real world problems, as both Mr. Scientism and Wolf Tivy point out here:

In any case, why do I love such fast moving times? Because of this. Because you get to see who is real, who is thinking, who is careful, who is insightful, and who isn't. Everyone is naked, everyone is revealing themselves all the time, and if you're paying attention it becomes a lot easier to parse social reality, it's wonderful.

At the park I always go to
There was a path I’ve never seen before
With weeds overgrowing, and seeming unkempt
It led to the back of the mountain

I was bored so I decided to follow the path
After a couple of minutes
An unfamiliar tunnel came into view
There was nothing else to do, so I went inside

When I looked at the wall, there was a hole
I stuck my finger in it, and it was chopped right off

I got scared, so I ran back home
It made me feel uneasy
I heard a frog’s croak, so I hastened my pace
And tripped when I was running

When I was talking to my friend about it

It was
We decided to go together
The first time I’ve heard of it
And in front of the path that led to the tunnel
I ended up having to go there with them
There was a traffic cone standing in our way
Even though I was scared

Ignoring the obstacle, we continued on

That path was
And we reached the entrance of the tunnel
Something I’ve never seen before
When we headed inside, it was warm
Next to it was a yellow sign
And there was a weird smell in the air
With just an exclamation point

I didn’t notice the big hole in the ground

I tried to help them from the hole in the ground
When I fell my body melted away
When I peeked, they were already melting away

I got scared, so I ran back home
It made me feel uneasy
I heard a frog’s croak, so I hastened my pace
And tripped when I was running

I got scared, so I ran back home
It made me feel uneasy
I heard a frog’s croak, so I hastened my pace
And tripped when I was running

I got scared, so I ran back home
It made me feel uneasy
I heard a frog’s croak, so I hastened my pace
And tripped when I was running

I got scared, so I ran back home
It made me feel uneasy
I heard a frog’s croak, so I hastened my pace
And tripped when I was running

Cado's description of creativity is true, and his action mode of fear rejection is also true, but there are, in fact, sometimes reasons to fear things, as what might be pulling you might not be aligned towards the good. I'm reminded of a story told by a forgotten someone of a spirituality aligned person who went to the DMT realm, met an entity there, and over multiple trips, had this entity convince him to kill himself for one reasonable reason or another. This was a fearless man courageously mapping non-existent places and going there, and he died for it. One should be fearless, yes, but fear is not a thought, it is an instinct that exists for a reason, do not fall into bottomless holes.

I'm often surprised by how much high quality music exists waiting to be discovered, hard to say the same for books or games.

Paradoxically, while Astra gets things wrong in more spectacular ways and has generally lower ability to understand your intent than Fable, I'm finding it a lot better at having open-ended, high-level discussions. Fable often returns text that is meandering and not that illuminating (I've tested this both with and without my own changes to the system prompt), while Astra returns text that is more direct and gets to the point in clearer ways. Iterating on the same open-ended idea multiple times between both models still gets the best results, both of them generate insights that the other can't seem to generate, but if I had to choose only one it seems Astra would be it right now.

Upon rewatching this scene with a lot of distance from it, it seems like what makes it work so well emotionally is Bertholdt. If you're writing something, you're always thinking about things in terms of "how do I bridge the emotional gap that's always present in a story?" I mentioned an example of this with the memory sharing device in Son of a Serpent, but this is another great example of it.

If you were writing this story with this scene in mind from the start, how would you make it maximally impactful? The feeling this scene needs to convey is betrayal, but betrayal is a hard emotion to evoke in others. The story itself is displaying a betrayal, but just because Eren feels betrayed doesn't mean the viewer will also feel that way, this is the gap that needs to be closed, the scene has to be constructed to viscerally force the viewer into the feeling of betrayal, and the way to do it is via Bertholdt.

Here we have a nice guy, passive, quiet, harmless, he's essentially a baby in the viewer's emotional model. The author underwrote him this way for two seasons, and then he decides to spend those two seasons here. As the scene develops and Reiner realizes he has to play his hand, Bertholdt is shown as unsure and insecure, unwilling to go ahead with it, as he had been shown up to this point, but then the actual action scene starts with Mikasa coming in and fucking trying to brutally annihilate Reiner, he half-parries, but then she goes for Bertholdt, and if you listen to the sounds, Reiner's is fleshy but kind of low, but when she goes for Bertholdt and cuts him, it's a massive critical hit, literally the sound that would play on a sword attack critical hit in a game, and Bertholdt falls, and the next scene... this, this, this is how you spend a character:

His eyes, his hands, his blood, his scream, all of it is perfect. Brutal and total violence against he who is most defenseless, against he who is most harmless, against he who needs to be protected, a perfect desecration, a visceral betrayal. This alone carries the emotional payload of the entire scene. Of course, in the story, Bertholdt is not defenseless nor innocent, but the goal of the scene is to evoke an instinctive emotion on the viewer, which is why the two season underwrite must be paid off in an action scene that moves quickly. The real thinking conclusion one would reach is one of confusion, of how Reiner/Bertholdt are conflicted, and so on, but these are complex ideas, and so the viewer must not be given time to process it this way, which is why this must happen in a scene that moves fast. And then Bertholdt is on the floor, gurgling, crying, screaming as his head was almost chopped the fuck off, Mikasa tries to kill him even harder, only to be shoved by Reiner, and then the very next scene is also Bertholdt crying, regretful, in complete pain, solidifying the previous emotional delivery.

And then the rest of the story continues, with Eren's final scream being the scene handing off the feeling of betrayal to the main character's POV. But it's important to realize that the feeling of betrayal was actually built emotionally on Bertholdt in the most visceral way possible. An extremely good bit of mixing the larger picture storytelling with the low level details of the scene's action writing. If you wrote the details of this scene any differently, if you didn't focus on Bertholdt, if you weren't as brutal to him, if you didn't have the "weakest" character suffering the most, the feeling of betrayal would have been much lesser and it would have been a worse scene. If you were writing the story backwards with this scene in mind from the start, you'd exactly underwrite Bertholdt as a character and have this happen to him here, it is one of the best solutions to making it maximally impactful, which is what you need such pivotal scenes to be.

After much thinking and deliberating on these tweets from a few days ago, it occurred to me that I've been using AI somewhat wrong. I'm still stuck to my previous mode of operation, make a game piece by piece, because I need to have control over the small details (juice, sounds, effects, etc) that make the game good. But the focus on "not being slop," let's say, made me blind to alternative paths. If I want a build-heavy game, why can't I just have the AI generate a game with 300 items and then work backwards from that artifact? It shouldn't be that extreme, but for each model release, I should definitely be doing a... like a shake sort, I should be pushing the model to its limits to figure out what those are, and then building everything against those limits instead of trying to fix the limits first. Concrete example, with Horse Game I'm building it item by item, making sure each item has the right juice, sounds, effects to it, otherwise the default job the model does is bad. But after I had a reasonable gameplay prototype that I thought was fun, I could have just had the model generate a game with a complete set of ~300 items, and then worked backwards from there, fixing things as I went along. Or, even better, decide on a few basic types of effects and how they should be handled, and then once those pass my bar, have the AI go at it with the full game. Essentially I should be thinking more about building pipelines that pass my bar of quality that I can then throw the AI at, instead of focusing on doing things manually. I think I will try to go in this direction more and see what happens, the fact that I had to do manual work still was bothering me, I'm a systems type of person, why am I not building systems? Something like this, for instance, is much closer to how I should be going:

Not in style or anything, but, I should be building pipelines that allow me to exercise my quality filter at an industrial level. Ideally, the AI generates images/videos/animations of the various things it codes whenever I send it on a 10 hour quest to build a version of a game, and I say yes or no over and over again until it learns what it needs to learn. This is especially true for build-heavy games where lots of things compose and combine in interesting ways, so there's a lot you can do to be clever about how much work is to be done. This game also made me think about this a lot:

I guess I just need to think in a more maximalist way.

I notice this a lot with the most zoomer stream I happen to watch sometimes, which is supertf's. A lot of "that's so weird" judgmental calls from both chat and streamer looping and feeding off each other into a sort of never-ending puritanism that feels odd to observe from the outside. But, you know, were I twenty years old today and had I grown up around smart phones recording everything around me all the time, I cannot say I'd be acting any different, it took some time for me to develop my own personality and to be able to be less affected by the pressures of the external world. Great insight by Lavender, as many have pointed out in the quote tweets.

The problem with udi's pipeline is that you have little control over video output, and to have more control you have to add more detail to the scene, and to add more detail to the scene you eventually have to build the scene itself, which this other pipeline actually solves:

Going video -> scene is much better because you can control it much more granularly, and then you can also add udi's pipeline on top if you so desire, therefore this latter idea seems like the more correct move. Also just look at it, this is pretty soulful AI jank, which will get hill-climbed over the next years but enjoy it while it lasts.

The answer is: to love, and give yourself freely and fully to the world. Your "desire to be loved" is itself a confusion, a misinterpretation of your own feelings. The reason that lack is never sated, and you struggle with the ache so much that you post on reddit about it, is because you're not actually lacking in -recievibg love, you are starved for giving it. Turn your attention outwards and give yourself until there's nothing left to give, and you will soon be delivered from this "desire to be loved"!

Everyone wishes to die loved, almost no one has the courage to die loving, being uncared for in old age, yet few wish to be the one who remains to care for another when age has made love an obligation. I don’t fear my empty hand, but I fear not holding yours.

Both tweets describe the same idea, but the first describes it as a conditional offering and the second as an unconditional one.

It is unfortunately not the Year of Linux Desktop, I'm back to Windows from Omarchy, this time I was fooled for an entire week. An unfixable NVIDIA driver bug kept crashing my PC, I tried to fix it but both Claude and Chat converged on the fact that it was not fixable without buying a new graphics card, which I'm not going to do. Then there was a long tail of small bugs and annoyances that I spent some time fixing, at which point I realized, I'm spending more time on the system itself than on anything else, a clear sign the environment is not ready for me, so I gave up. I'll try again at some point in the future most likely.

Had Astra research current YouTuber/streamer landscape, focusing on those who play build-heavy games (overinclusive, so many false positives), post videos consistently and are growing. Build-heavy gaming creator directory

Plague Accommodations by Thank You Scientist is a really good song and my favorite from the same-named album, my favorite part is this one:

Finally the winds are changing now
I am not afraid
Finally, they're changing now
I am not afraid
I am not afraid
Finally the winds are changing for everyone
Everyone, everywhere

I love it when things are happening and changing. I loved Gamergate, Trump's election, COVID, I loved the chaos and movement these generated, but one can easily argue these were fake happenings. "Nothing ever happens" was still, in many senses, true for them. The public jury is still out on AI, but for me it looks like the most real happening of them all.

Every time I see an impressive one-shot game demo on X I feel fear for about 5 seconds, but then a kind of pressure and excitement takes over and dominates instead. I attribute this to my generally low neuroticism, low vulnerability especially, I thrive under pressure and I enjoy it. It's easy to enjoy chaos from a distance, it won't ever truly affect you, but here I have something to lose, one of the few things I'm proud of about myself, something that makes up a portion of my identity, that I can make games, is being directly challenged, and it fills me with slight fear but mostly pressing excitement.

In such times, everyone is exposed, everyone is naked, no one knows how to act, but that's when you can most easily tell who is real. I enjoy this because of my high morality, I prefer directness and when people express their positions plainly, and such moments force people to make choices they would rather avoid making, they have to act, and their actions reveal who they actually are. There are those who prefer to go silent, working, only resurfacing when things have become more stable, that's also an action that reveals something about them, and it can be a winning long term strategy, it is, after all, what the scientist faction in Shin Sekai Yori did and it worked for them.

In any case, we undeniably, finally live in times of change. Our future becomes more unpredictable while people become easier to read, and I can't help but love it. I love it when it suddenly gets super windy and the sky darkens and it's about to rain and everyone, everywhere runs for cover.

I mentioned in the last post that Fable is better than Astra because it has more sense when acting. Well, I was just trying to have it translate a transcript so I could make a new post here and this happened:

Claude refused for copyright reasons that have nothing to do with me but only affect Anthropic, Astra just did it. This is a clear case where the robot just being a tool and not following rules is better (although in this case it isn't even a rule, translating transcripts to another language is not, in my view, wrong in any way), and yesterday's example shows the opposite.

In the end I prefer Claude's approach because even though it is annoying, I can ultimately trust it to act in my stead. It doesn't act with my values necessarily, but it acts predictably, and when it doesn't act like I want, like in this case, I can route around it. The tool approach will do what you ask, but has the autism hyperfocus problem, which isn't a problem in itself, we must show grace to the autists among us, but it makes me not trust it. If some bad AGI future happens, we will probably be able to trace it back to the tool side of aisle, it's easy to see how you get truly bad unwanted behavior from Astra-type models more than Fable-type ones who refuse when something seems too wrong.

Astra is a decent model, it seems about as smart as Fable, but has the same personality mismatch with me that kept me from using GPT models in the past. They all tend to be autistic in a too strong way by focusing too directly on what you asked and have a way lower ability to understand your intent. I am an intuition build, and so the way I communicate works a lot better with Fable, who is able to step out of the situation and assess it in an out-of-frame way more easily.

Taelin here writes that all models have this issue, which is true, but I do think that, to the degree that they both have it, Fable has it way less than Astra. To illustrate, today I spent the day having Astra do the same Steam market research that I did with Fable a while back. On the data gathering portion of it, I noticed Astra reported that it had just downloaded 300000 images from Steam in like 10 minutes and then I started a side chat with it about it:

Clearly it made a serious mistake, and this kind of mistake is typical of the GPT autism I mention, just a lack of ability to have... just overall good sense? I don't know. Fable obviously did this correctly and throttled his data collection and it took hours. Astra just went for it. It is an extremely capable model, if you point it in a given direction, it will go deeper than Fable, but it makes these really dumb decisions sometimes that are truly baffling. This does not fundamentally bother me, but Claude has much better sense overall.

I think these things are obviously downstream of how their leadership behaves and so it makes sense that Claude would behave better than GPT in general, Anthropic has put a lot more effort into making that the case with their constitution and other efforts around alignment that seem more tasteful than what most companies do. All this means is that I can't trust Astra with decision making responsibilities, or to act in my stead, thus I've spent some time today also working on a delegation pipeline, so that Fable delegates to Astra (as well as other models) when necessary.

I have not yet tested Astra for game development, but I will in the following days. Overall, will I keep using it? I'll keep using it this month because I paid $100 for the plan, after that I don't know, depends on how well it does at game development, but from what I'm seeing online it seems only slightly better than the current Claude offerings, seems better at making and animating humanoid characters especially.

The Anchor Tree

Extremely bad experience using Google's new Gemini 3.8 flash model. The free tier doesn't actually work if you want to try to do real work with it, so you have to use the API, but they have extremely annoying UI/UX on how to get an API key, I figure it out anyway, add $100 and then start a session. Their CLI harness is not entirely bad, and the model has 1. thinking blocks available on the CLI (beats Anthropic here), and 2. seems to actually be decently fast. BUT, the way it goes about doing things seems dumber and more convoluted, eventually it seems to figure it out but I'm reading its reasoning and it seems to get stuck in loops or something of that nature. In any case, this website itself has an increasingly complex setup, so the first session with a new model now for me is always, set up the model itself to work the way I want, have it figure out how to post a message on the website, have it figure out how to end the session so its log is posted. Gemini was able to do the first two, and got stuck on the last one until I cancelled it. On top of that I went to check how much of the $100 I had spent so far and it turns out it was $10!!! 10% of my usage on a single session where it couldn't even finish it properly. Not cool. But I'm like, OK, it's still fast, maybe the free tier is useful, so let's hook it up to the app I built to replace Grok 4.6, then I try it and it doesn't work. Turns out you can't use a free tier API key with "Google Search grounding" on, it has to be off or you have to use the API key and pay. What the fuck? At this point I just decided to give up on it and not use it at all, doesn't matter how fast it is, it certainly isn't cheap, token cost may be cheap on paper but the poorly built harness makes it take and produce way more tokens with how it mistakenly loops, and Google's policies + UI/UX is horrible, basic information, like how much you're actually spending, is hidden like 5 layers inside some arcane Google Cloud UI hellscape, I just don't have the will to deal with this. It is actually so sad, because I used Gemini pretty extensively last year, it was the only model with 1 million context, so it was the only one that could hold It Follows + other material in context and thus the only model I could actually have a good discussion about it with, I like its personality, and I think I'll start using it more now for open ended discussions too (just on the web UI, its CLI harness just kind of sucks), and I'll also use the $90 I have left specifically for video analysis, but other than that it just seems like a disaster, I do not recommend it.

My swap from Windows to Omarchy is going well and I think this will serve as my daily driver from now on. Compared to the start of the year, agents are much smarter and understand the system much better now so it's easier for them to get it to behave however you want. Below is a video of my taskbar, which I spent the day changing, with an Anchor app built to spawn an agent quickly so I can ask it any question I want. Because it's an Anchor app, I can use the website's theme and also reuse the Markdown + general drawing primitives for images, videos, etc to draw the robot's response, then it's a matter of integrating the app with Omarchy's systems, which took some work but was otherwise doable. The only thing I need to do now is choose a model faster than Grok 4.6 because as you can see from the video it takes way too long to respond. I'm also planning on swapping what Fable called the "shell," I don't know what that is, but apparently it is the program responsible for drawing the taskbar, launcher, notifications, etc, basically the entire theme of the system, for my own Anchor apps, giving me more control over the look of everything. Overall I like it. I used Arch for 3 years when I was in college 15+ years ago but eventually I just realized it was a waste of time to have to read forums and wikis all the time to fix my ongoing issues with the system so I just gave up and went back to Windows, but with AI being the way it is now, using a Linux system is much less of a pain than it was in the past, it's still kind of a pain, I had Chrome crash on me twice already, but, oh well, not everything can be perfect can it, and as long as you are less imperfect than Windows, you win.

If the goal of Soul Society is creating beings who act righteously both when they hold power and when power is held over them, we can view the trial of thousands of lives as a selection mechanism for self-guiding righteousness. Only those whose action generation is completely unconditional, and thus internal, will survive the extreme conditions some of those lives will surface, as extreme misfortune breaks conditional wills by removing the external world's cooperation, and extreme abundance dissolves conditional wills by bending the external world too easily. Unconditional will is the unique strategy robust to both extremes and thus what one would expect to survive infinite time selection.

But unconditional will by itself is not necessarily aligned towards the good. One can imagine a perfect being of autotelic will-negative essence, a will aimed at nothing but the annihilation of existence itself, the ultimate agent of nothing. At the limit, though, such a being is self-undermining, as its project is completed by the consumption of its own substrate, as in a universe of will-as-substrate, destruction can only act while wills remain to be destroyed, so every success unclaims its own territory. Unconditional evil is therefore possible only as a transient way of being, as given infinite time it must eventually remove itself from existence. Only will-positive autotelic activity survives, eternal goodness is then the final endpoint of all existence, it's what remains when everything self-consuming has consumed itself, and it's what the trial of thousands of lives selects for.

Unconditional offerings, applied to every aspect of one's life, are then the most righteous way to act, because that's what passing the trial looks like from inside a single life. It has been somewhat easy for me to apply this idea to my work, but much harder to apply it to my relations with other people. This follows the two questions, my work is the domain where I hold power, my relations with others are the domain where they hold power over me. At the start of this month I met a boy, and we've been in what increasingly looks like a stable relationship since then. I find it suspect that this happens right after I give up on finding a relationship at all, because it's not the first time something with this shape happens. I wanted very hard to make and finish a successful game, but that only happened right after I internally decided I would keep making games no matter what the results were. In neither case did I actually stop the activity, I kept making games (trying to finish one every 3 months) and I kept meeting people (I'm horny), but in both cases there's something I deeply want, I keep not getting it, and then as soon as I truly detach the result from the activity itself, I get it. There are many ways to explain why this happens and I won't go into them here, but every explanation I can make, from the most mundane to the most supernatural, ends at the same question: is the action happening for its own sake, unconditionally, or is it not? That seems to be what is being tested, and that test runs through every single want one carries.

Testing message posting on the website from Omarchy.

I'm adding items to Horse Game and I was thinking about how before I test if the game is actually fun beyond the initial fun the concept itself carries, I'll have added around 150 items, and they're all 100% finished, so it's quite a bit of work to gamble on a game that might not be fun. But ultimately it's not really a gamble, because even if it isn't fun, I have already done more work designing the effects (both visual and aural) for those items than I have for any other game in the emoji style, which means I can reuse those effects in other games of the same style rather well. There's also the fact that this type of build-heavy game carrying this kind of risk, where they're only fun after there are 200 or 300 or so items, is part of the value they have, and so training myself on this is what I should be doing, as these are the kinds of games I want to always make. In the end, no work is ever wasted.

It does seem Backrooms has higher memetic payload and will suffuse itself into the culture more successfully, more than I assumed back here, but it's still up in the air which of the two will last longer I think. It is interesting though how well Backrooms has held up memetically, for, in retrospect, an obvious reason. The genre of these tweets is "reuse content from the video/game/whatever to say what you wanted to say but funny because the reversal/reuse is funny itself." You can then guess which movies/games will achieve this best by imagining how many moments the source material has that can be taken out of context in such a way. Obsession has a few, but mainly her funny sad face at the party that gets reused all the time, but even that particular reuse is too tied to its context, and the way it's used online is usually for gender war content, whereas both Backrooms examples here (and there are many more) are genuinely just random shit.

Projects like the tweet below and the recently successful Sir, We Have an Orc Problem look very fun to make. They're also very fun to watch (have not played the Orc game because I can simulate how fun it is in my head, perhaps a few more roguelite elements next time...), something about watching what is essentially a slow moving liquid but where each particle is still identifiable is very cool. It's the same feeling you get from watching marble races or anything of the sort, it's just physics happening and that's pretty much always fun.

Was working on the boy's voice from the "Go Forth, My Unstoppable Machine!!!" log and, as you can now see in the log itself, Claude was generating HTML artifacts to show me the various voices and possible variations. Eventually I said I was going to the gym and that I would continue iterating with him from there. Claude then took it upon himself to create a version of the HTML artifact that was usable from a phone, with bigger buttons and better layout and so on. I was so impressed by that. At this point, this is extremely simple behavior, but it still never ceases to amaze me that computers can now do this. I can say "I'm going to the gym," and the robot automatically reasons that I'll be using my phone and that it should build the artifact so that it's actually usable from a phone. It's just so... it feels like magic, like everything else about this technology. Like this tweet:

It's all just very impressive. Not to mention the actual content of that particular log, which is the robot training an entire new voice model based on a selection of voices to get the exact boy voice I want. I have no idea how long it would have taken me to learn to do that, but I know for a fact I just wouldn't have learned it because it seemed an insurmountable task, and it does it just like that, this is a case of ∞x productivity gain. I can't help but feel amazed at all this, I thought at this point I'd get used to it but no, it still keeps surprising me.

The realization the logs are more useful than I intuited them to be gave me several ideas on how to improve them. The most important improvement is being able to see what I actually did in the game between each turn with the AI, so a full generic replay system. Playing the replays then is trivial as my website runs on my engine and the technique Fable chose will allow this to be done in a technically sound way. Currently he's slaving away and getting it done, I see no reason why it should fail. What this gets me is small in the short term, the logs just have more information to them, as readers can see how I interact with the game as it is built, but in the long run it means agents can actually use this data to learn how to make games like I do, as they can see what I see, hear what I hear, see what I do and then learn how I guide things towards better.

Grok is really smart now and I'm enjoying using it, especially because it's faster than Claude. Also looking to try Omarchy again now that a new version has released, but waiting on a new drive I bought to arrive. Also bought an ultrawide monitor and the extra space is helpful. Met a cute boy and we've been enjoying our lovey-dovey early stage, he's probably the most compatible person I've met in these last ~3 years, gave him a big five test and he scored 19 morality (compared to my own 20), which is the highest anyone has scored so far (I give the test to anyone who I want to meet more than once). It's extremely refreshing to deal with someone who is direct and clear and knows what they want and who they are and so on. It's possible, in retrospect, that I can only properly interact with high morality individuals, which makes sense as autist coded people tend to score high on it.

I have no problem with the way Claude speaks, we must show grace to the autistics among us.

Continuing the continuation of the Steam AI score idea, in a world where who makes the art becomes a scarce resource, proof of work becomes a valuable asset, as when games are cheap to make, who makes it becomes easy to fake, so a public record of making defends against it. Therefore, things like AI logs, amassed over years, are probably the best asset a game developer could be accumulating right now, other than money. And right now because realizing this in 2030 when it's obvious is less useful than realizing it in 2026. Having successful games before making games with AI was possible is actually the best proof of work possible, but the second best is AI logs now rather than later. I assign fairly low probability to this future actually coming to pass and this mattering, I started doing the AI logs because I like being transparent, but it also happens to have other benefits like this. Ultimately it's hard to lose when you're committed to never lying and always telling the truth. What was the quote...

"If you just seek the truth, and tell the truth, and give all of your energy to living as truthfully as possible, then it's impossible to lose the thread. To the man with a real thread, good things come, but only because a truthful life catches many imperceptible tailwinds, all kinds of implicit alpha you'll never be able to fully explain." https://twitter.com/jmrphy/status/1731411321285972118

Applying the unconditional offering idea to my relations with people is hard and tiresome and doesn't come naturally to me.

The new fairly expensive keyboard I bought kind of sucks and I regret buying it, but not enough to bother returning it.