classic. this has been a problem since the language was new.
the problem is that Rust is designed to not have this as a use case, or rather has machinery to make this marginally safer (mutexes etc) or make it unstable (panic) or just as unsafe as C (unsafe). if this is a pattern you want to maintain, Rust is going to throw friction at you by design, because tbh this pattern is the source of a ton of bugs.
it’s conceptually more complicated, but game engine devs have been moving to some form of Entity Component System[1] architecture, because it gives you a way to access memory that is both cache efficient and safe.
all that said i’m not a game dev. i think Bevy[2] was the blessed solution for a while?
yeah, i mean it’s all bit abstract, but that’s essentially what it is.
a concrete example would be Nvidia. they’re wildly profitable, and they don’t care how overleveraged you are as long as their barrels full of cash keep coming in.
i agree with you that there is a concerning amount of debt and leverage going on, but it’s undeniable there’s money changing hands.
but also, speaking of abstract, all this shit is made up and most of the money “moving” is just flipping some bits in a database somewhere. i also agree with you that it’s all a bit of a facade for people with power to play with power while the true “value” of society such as cost of living and quality of life take a low priority cuz “AI has a $T valuation”. but if you live in this system, this is what it is.
yeah, i know the argument you’re trying to make, but that doesn’t change the actual capital and labor being used to create these things. whether it’s profitable once the investors run out of money is another story.
i mean, people are making money. sure there are debts owed, but Mark isn’t going home to a 1br rental even after being so “humiliated”. the bag holders are the investors.
i kinda feel sorry for Yann LeCun and his team at FAI. he’s been saying from the beginning that LLMs are not the way forward for AI research as a field, and that has kept Meta from competing in this gold rush effectively. world models, multi-modality, etc are fine and good research but aren’t here to compete with Claude Code or even as effective at being good enough for stuff like search.
not to be too reductionist, but it’s a tough spot when your head of AI is insisting that this tech has no future when there is a literal $T market for it.
the irony is that these “frontier labs” like OpenAI and Anthropic have basically dropped every greenfield research project that doesn’t directly support their “agentic AI” infrastructure (ie endless Markdown prompts + a while loop) while the labs working on the true frontier are seen as failures because they can’t make a bazillion bucks burning megawatts of energy to replace their front end devs.
this has been “the plan” since 2005. when i was in high school trying to figure out what the heck “cloud computing” was, this is what they were talking about: anything requiring more compute than secure authentication and pixel drawing would be rendered in the cloud and delivered to dumb terminals. this is what netbooks, Chromebooks, and smartphones have been a step towards if not an implementation of.
is there a material difference between taking profits and “getting very nervous”? the reason to “take profits” is that you’re “very nervous” that the price won’t keep going up; that’s the game right?
the key is just an offsite machine with a stable IP plus a VPN. that way you can route all public traffic to that stable IP and send it where it needs to go regardless of its physical destination.
i have a VPS offsite to act as a gateway. it’s just a small piece of a machine somewhere in my region that routes requests to my home network via Tailscale. this has a few benefits:
i don’t have to worry about my ISP changing my IP. my VPS has more stable IPs.
i don’t have to expose ports directly to the internet. Tailscale authenticates the connection. plus i have Caddy routing the whole system. i use subdomains like foundry.chrash.net, jellyfin.chrash.net, etc.
another benefit of Tailscale to point out is that you don’t need local IPs to be static either; Tailscale will allow you to access your machines by hostname or another static IP. this helps to decouple your local topology from your service network.
i use Nushell for this! works with JSON, YAML, TOML, markdown, Polars Dataframes, SQLite, and a bunch of others including builtin parsing tools for whatever formats and a plugin ecosystem. i use it at work and for personal projects as my main shell, and it’s super handy for exploring, unpacking, sorting, and visualizing all sorts of data. i use it to:
find specific parts of YAML cloud configs
visualize JSON logs, including a parser that restructures journalctl logs.
_re_structure data from CLIs to work with them as structured: git logs, Unix coreutils, etc
script my environment: common kubectl queries, specific web API helpers, building and running and testing applications, etc
it is a slight learning curve, and technically you could do all of that with bash or zsh and jq or jc, but i appreciate the modern take on your base shell terminal env.
i dunno if “realism” is an argument here. you’re talking about a specific market segment targeting a specific hardware configuration and distribution medium. developers still have the choice to target Nintendo or Sony hardware, to sell physical copies or codes through Walmart, Amazon, Target, Gamestop, your local game store, etc, to sell via mobile platforms like iOS or Android, etc etc.
honestly, if i sat here and listed them all out it would be an enormous comment.
i do see how Valve has a hegemony over a big part of the market, but they haven’t been anticompetitive or tried to push anyone out or buy up competition. at least that’s not what’s being claimed, as far as i can tell. Epic’s lawsuits against Apple and Google don’t even apply cuz you can install friggin Windows on their hardware if you had some sort of mental illness.
don’t just wrap std::io::Error like that. same with request::Error. your custom errors are there to tell you about your domain. “there was an IO error” doesn’t mean a lot to your users. you could have wrapped that like AppError::ConfigLoadError or whatever; this is a big reason for having domain errors (or custom errors if you like) in the first place.
thiserror is a great package with great docs that does a lot of this for you.
don’t be afraid to nest your internal module errors. we use ConfigError all way down to something like OurBeskpokeFormatError, for example. then use impl From<ConfigError> for AppError. this of course depends on the scale of your application, but a 1000 line AppError enum is a mess.
great advice generally! i’ve seen apps start off with anyhow and get stuck there with badly structured errors wedged in everywhere
the problem is that MCP isn’t really solving a problem that can’t be solved with existing technology. it’s actually infuriating that access to resources is hidden behind this silly new protocol for agents only as if just giving me API access, which is already well-understood and supported, is insecure or traffic heavy.
plus, the gold rush at my company of adding an MCP server to every resource has turned into a graveyard of wasted effort and abandoned projects.
as someone who has a similar job, i don’t think it’s so obvious. there’s a lot of middle ground between an AI slop PR and artisanal, hand-crafted code. if i use a library or algorithm or pattern suggested by ChatGPT or use Copilot to autocomplete a simple function or have Claude generate test cases, that’s all “AI assisted”.
i always hated the web as a programmer. i couldn’t put my finger on it as a junior Android dev. now i’m a cloud engineer doing federated GraphQL at a huge corporation, and i know exactly why i hate it. HTTP sucks. HTML sucks. CSS sucks. JavaScript sucks. the web is jank and designed to leak information about you. you have to be pretty obtuse to say that any website operates as smoothly and seamlessly as any decent native app, unless it’s the “This is a Fucking Website” design with no modern features or JavaScript. sure, there’s a mountain of RFCs around these protocols, but not even Chromium follows them all to the letter. documentation has always been a desperate way to enforce protocols. i’ve always been interested in alternative protocols, but this has finally inspired me to give Gemini a go in my off time.
but i’ve been more interested in data passing than application distribution (ie HTML/CSS/JS). i’m curious to see what these protocols offer beyond yet another markup format. maybe a better object format than friggin JSON? 🙏
my point is that we’ve just internalized this as normal. that it seems like such an insurmountable chore to even go on a bike ride or hike or maintain a garden or building project is sad to me as well. i grew up in a culture that absolutely hated any form of exercise, and i watched a lot of people live absolutely miserable, short lives because of it. i’ve put a ton of effort into rejecting that lifestyle, and i recognize that as a privilege.
i don’t think it’s so natural for beings to optimize their effort to zero. dogs will chew through cages or self harm. and i’ve seen humans who, without stimulus, will act similarly.
it’s a sad state of affairs that people bemoan 2 hours of activity a day. i personally struggle to meet that so i get it, but it’s insane that we’ve created a society that wails in fear over not being able to be (ETA: or rather having to deal with the consequences of being) completely sedentary for a vast majority of their life.
classic. this has been a problem since the language was new.
the problem is that Rust is designed to not have this as a use case, or rather has machinery to make this marginally safer (mutexes etc) or make it unstable (panic) or just as unsafe as C (
unsafe). if this is a pattern you want to maintain, Rust is going to throw friction at you by design, because tbh this pattern is the source of a ton of bugs.it’s conceptually more complicated, but game engine devs have been moving to some form of Entity Component System[1] architecture, because it gives you a way to access memory that is both cache efficient and safe.
all that said i’m not a game dev. i think Bevy[2] was the blessed solution for a while?
1: https://en.wikipedia.org/wiki/Entity_component_system?wprov=sfti1 2: https://bevy.org/