I linked it in a top level comment, but if you’re someone who likes watching long, thoughtful video essays, this is an excellent one on the precise topic of your comment, from one of my favourite science communicators

Ooh, an opportunity to share a video from one of my favourite science communicators, the inimitable Dr Fatima (an ex-astrophysicist): “Einstein Was a Socialist; Should We Care?” (Duration: 1h16)

You’re right, it was the .se one I used and I misremembered when writing my comment. Though this mistake highlights why adding sites like this to your bookmarks is important. Opportunistic scams are always common when a big domain goes down, and mistakes are easy to make

They already had a big target on their back. Sure, scraping Spotify made that larger, but given that their aims are explicitly political, I think there’s a lot of value in not being overly cautious and using takedowns as a way to bolster the resiliency of your service.

Domain takedowns are inevitable, even if they’d just continued with their existing archival efforts. That’s why, when I found I wasn’t able to access the .org domain over the weekend, I just went to https://annas-archive.se/ instead.

Edit: wrote one of the old domains (.gs) because I just use what’s bookmarked and got mixed up. It was .se that I used

Makes such a useful piece of software, and is also wise enough to set boundaries to protect himself from the toxic pressure of open source development.

What a G.

I get what point you’re making in distinguishing between pedophile and ephebophile, but personally I don’t find the distinction particularly relevant. As an adult, the level of grossed out I feel at the prospect of sexual interactions with a young teenager Vs a literal child is approximately equal, because it’s not their physical attributes that cause ick, but rather the exploitation and power dynamics involved.

Edit: I guess what I’m arguing is that in practice, we see the term “pedophilia” used as an umbrella term that encompasses pedophilia, hebephilia and ephebophilia, and I think that is a reasonable use of the term. It does muddy the waters a tad, given that pedophilia does still have its more specific use of referring to sexual attraction to pre-pubescent children, but I don’t think that an issue in the majority of contexts. When it comes to the law, an adult having sex with a child is equally illegal as an adult having sex with a 15 year old. Sure, we can split this hair and distinguish between the terms, but we don’t need to

I use Fedora, but I frequent the Arch wiki often enough that I feel like an honorary Arch user.

I’d rather hone my skills at writing better, more intelligible code than spend that same time learning how to make LLMs output slightly less shit code.

Whenever we don’t actively use and train our skills, they will inevitably atrophy. Something I think about quite often on this topic is Plato’s argument against writing. His view is that writing things down is “a recipe not for memory, but for reminder”, leading to a reduction in one’s capacity for recall and thinking. I don’t disagree with this, but where I differ is that I find it a worthwhile tradeoff when accounting for all the ways that writing increases my mental capacities.

For me, weighing the tradeoff is the most important gauge of whether a given tool is worthwhile or not. And personally, using an LLM for coding is not worth it when considering what I gain Vs lose from prioritising that over growing my existing skills and knowledge

I want this on a t-shirt

I see your point, but as you say, there would still be the tradeoff of missing more recent stuff. That might only involve missing a couple of years’ worth of stuff now, but AI isn’t going away any time soon, so it would mean that there’d be an increasing amount of human made music not being archived; One of the things I like about Anna’s archive is that they seem to look at this problem as a long term, informational infrastructure kind of way, so I imagine they wouldn’t be keen on stopping the archive at 2023.

It seems they’ve opted for a different tradeoff instead: lower popularity songs are archived at a lower bitrate, and even the higher popularity stuff has some compression. Some archives go for quality, and thus prioritise high quality FLACs, so Anna’s archive are aiming to fulfill a different niche. I can respect that.

I agree with the ethical standpoint of banning Generative AI on the grounds that it’s trained on stolen artist data, but I’m not sure how tenable “trained on stolen artist data” is as a technical definition of what is not acceptable.

For example, if a model were trained exclusively on licensed works and data, would this be permissible? Intuitively, I’d still consider that to be Generative AI (though this might be a moot point, because the one thing I agree with the tech giants on is that it’s impractical to train Generative AI systems on licensed data because of the gargantuan amounts of training data required)

Perhaps it’s foolish of me to even attempt to pin down definitions in this way, but given how tech oligarchs often use terms in slippery and misleading ways, I’ve found it useful to try pin terms down where possible

Yeah, that was my understanding of things too. What I’m curious about is how the Indie Game awards define it. Because if games that use ((Procedural Generation) AND NOT (Generative AI)) are permitted, then that would surely require a way of cleanly delineating between Generative AI and the rest of procedural generation that exists beyond generative AI

I’m not so much talking about machine learning being implemented in the final game, but rather used in the development process.

For example, if I were to attempt a naive implementation of procedurally generated terrains, I imagine I’d use noise functions to create variety (which I wouldn’t consider to be machine learning). However, I would expect that this would end up producing predictable results, so to avoid that, I could try chucking in a bunch of real world terrain data, and that starts getting into machine learning.

A different, less specific example I can imagine a workflow for is reinforcement learning. Like if the developer writes code that effectively says "give me terrain that is [a variety of different parameters], then when the system produces that for them, they go “hmm, not quite. Needs more [thing]”. This iterative process could, of course, be done without any machine learning, if the dev was tuning the parameters themselves at each stage, but it seems plausible to me that it could use machine learning (which would involve tuning model hyperparameters rather than parameters).

You make a good point about procedural generation at runtime, and I agree that this seems unlikely to be viable. However, I’d be surprised if it wasn’t used in the development process though in at least some cases. I’ll give a couple of hypothetical examples using real games, though I emphasise that I do not have grounds to believe that either of these games used machine learning during development, and that this is just a hypothetical pondering.

For instance, in Valheim, maps are procedurally generated. In the meadows biome, you can find raspberry bushes. Another feature of the meadows biome is that it occasionally has large clearings that are devoid of trees, and around the edges of these clearings, there is usually a higher rate of raspberry bushes. When I played, I wondered why this was the case — was it a deliberate design decision, or just an artifact of how the procedural generation works? Through machine learning, it could in theory, be both of these things — the devs could tune the hyperparameters a particular way, and then notice that the output results in raspberry bushes being more likely to occur in clusters on the edge of clearings, which they like. This kind of process would require any machine learning to be running at runtime

Another example game is Deep Rock Galactic. I really like the level generation it uses. The biomes are diverse and interesting, and despite having hundreds of hours in the game, there are very few instances that I can remember seeing the level generation being broken in some way — the vast majority of environments appear plausible and natural, which is impressive given the large number of game objects and terrain. The level generation code that runs each time a new map is generated has a heckton of different parameters and constraints that enable these varied and non-broken levels, and there’s certainly no machine learning being used at runtime here, but I can plausibly imagine machine learning being useful in the development process, for figuring out which parameters and constraints were the most important ones (especially because too many will cause excessive load times for players, so reducing that down would be useful).

Machine learning certainly wouldn’t be necessary in either of these examples, but it could be something that could make certain parts of development easier.

I don’t remember where I heard it, but a phrase I quite like is “AI is the death drive of capitalism”

Might also be a context switching thing

Like, when I have a dedicated space to go for work, then I find that really helps me to get into the right headspace. My productivity has always been shit when I’ve lived somewhere that doesn’t have enough space to do this.

Maybe what’s happening is that the different language forces you to be in a different headspace, which for some reason, helps you to focus better.

This theory is weakened somewhat by the fact that your mother tongue is Portuguese, and you don’t find your focus to be improved by English.

It does feel intuitively plausible to me that there is some underlying linguistic thing going on here. There might be some research studying the link between different languages and ADHD experiences, because it does seem like there’s something interesting there. If there isn’t currently any such research, I have no doubt that it’s just because it hasn’t been done yet (the wide domain of “academic research on autism and/or ADHD that respects the personhood of the people being studied” is unfortunately, a relatively recent development, but I have been pleased to see that it has been growing rapidly in recent years). If I find anything, I’ll report back (which may be in many weeks or months)

How to Be Composed and Focussed: ADHD edition

• Step 1: throw away self help books that are aimed at neurotypicals. The advice in them is probably not helpful for us, and will just exacerbate internalised ableism. Not only will it take different strategies to get there, but “composed and focussed” will look different for you than it will for neurotypical people.

For example, a friend I had found that she was only able to complete her university essays when she engaged in an odd sort of task circuit-training, where she had multiple different tasks that she could cycle between as soon as she found herself losing focus. To an external, neurotypical observer, this looked like absurd chaos, but that was how she found her focus.

• Step 2: try your best to work against the aforementioned internalised ableism. This, unfortunately, is an ongoing task, because even once we throw away unhelpful frameworks, we can’t escape from the unreasonable expectations that the world places upon us. That is not your fault, and you are not broken just because you can’t fit into the pre built mould that society offers you. It is possible to build new frameworks that will comfortably fit and support you, but we’re going to have to do a lot of that work ourselves. This is a task that will be an ongoing one, so proceed to step 3 whenever you feel ready.

• Step 3: find neurodivergent community. This is the most important step, because it can do wonders for helping with step 2; it gets tiring to have to constantly remind ourselves that we’re not broken, so it’s helpful to have other people help remind us of this sometimes. Plus ADHD folk often find it’s easier to care for other people than for themselves, so you might find it easier to affirm other people than yourself. That can be a good starting point for learning how to extend that same grace to yourself.

It doesn’t matter whether it’s online or irl, a space specifically dedicated to discussing ADHD/autism or just a hobby community with lots of neurospicy folk — just find your people. It’s daunting to feel like you have to build an entire mode of living from scratch, but you’re not doing it alone. Ask people what strategies they have found useful for coping, and if you find anything, share that with others too. We’re not a monolith, so not everything will work for every person, but having these conversations about what works and what doesn’t is super useful.

• Step 4: Remember that there is no silver bullet here, no single strategy that will fix everything. I’m sorry to have to emphasise this, but the best tool is the one you use. Try not to fixate on the next shiny thing, because that’s a false comfort. I know that actually using the tools and strategies is the hard part, but that’s why we need to keep working at it. You will struggle with this, but that’s not failure, it’s part of the process. Refer back to Step 2 if you need to.

Step 5: Remember the big picture. What we’re building here is social and informational infrastructure. My own experience has been improved by having access to resources and communities online that are made by and for neurodivergent people; if I were born 100 years ago, I might’ve ended up in an asylum. It often doesn’t feel like it, but things are getting better. It’s overwhelming and scary to be building something new on the margins of society, but we have the ability to improve things both for ourselves, and the people who come after us.

We’re trying to do something radical here, and that will take time and a lot of work. Most of us were only taught how to be successful neurotypicals, which is something that we can never be. We are having to learn from scratch how to be successful neurodivergent people, but there isn’t a simple guidebook for that. We have to muddle along as best we can and write that guidebook ourselves. In this way, learning how to live as ourselves is a powerful form of political praxis^[1] (which may be a helpful thing to remember if you tend to beat yourself up about being too burnt out to engage in as much activism as you’d like).

[1] : Praxis can be generally defined as the process of putting theory or ideas into practice. In this case, we can say “we deserve better than to live believing that we are no more than failed neurotypicals”, but then there’s the tricky question of how do we put that ideal into practice? That’s the ongoing quest. Praxis in this context also draws from how it’s used in Marxist thought, which is that praxis is about actions that are oriented towards changing society.

Yeah, I’ve been seeing an increasing number of artists who are pro piracy, who basically say “steal our music, save your money, and if you want to support us, come to a gig and buy some merch”.

I’ve also seen more and more artists staying off Spotify entirely. One such artist is the wonderful folk artist Lucy & Hazel . This was the first time I actually bought music in years, and a big part of that was because I wanted to support their active choice to stay off Spotify.

An unexpected side effect of this is that because I’m aware these guys are situated less optimally for algorithmic discoverability, I find myself actively recommending them to people. It feels nice compared to the more passive mode of algorithmic music discovery

I’m not sure how they would go about doing that at scale without also getting some false positives and removing human music too

Can someone help me to understand the difference between Generative AI and procedural generation (which isn’t something that’s relevant for Expedition 33, but I’m talking about in general).

Like, I tend to use the term “machine learning” for the legit stuff that has existed for years in various forms, and “AI” for the hype propelled slop machines. Most of the time, the distinction between these two terms is pretty clean, but this area seems to be a bit blurry.

I might be wrong, because I’ve only worked with machine learning in a biochemistry context, but it seems likely that modern procedural generation in games is probably going to use some amount of machine learning? In which case, would a developer need to declare usage of that? That feels to me like it’s not what the spirit of the rule is calling for, but I’m not sure

They’ve released torrents of the metadata, and they plan to release the music files, but they haven’t yet. They intend to start by offering the downloads as bulk torrents, but they’re open to considering implementing the ability to download single songs in the future.

So in short, yes, but you can’t download them yet