It's okay till it's not. Everyone I know who had Celery in production was looking for a substitution (custom or third-party) on a regular basis. Too many moving pieces and nuances (config × logic × backend), too many unresolved problems deep in its core (we've seen some ghosts you can't debug), too much of a codebase to understand or hack. At some point we were able to stabilize it (a bunch of magic tricks and patches) and froze every related piece; it worked well under pressure (thanks, RabbitMQ).
How is the WASM target looking nowadays? I tried it in 2023 and early 2024, and it was far from complete (JS interop, poor documentation scattered across GitHub issues, and so on). I still can't find a single trustworthy source of documentation on how to proceed. C# would look great at the edge (Cloudflare Workers).
A lot of people use LLM subscriptions, they don't use API keys. Why devtool devs continue to ignore this? I mean, ok, you have your "pricing" reselling an access to LLM models, that's the only, probably, simple way to monetize it, but it's a bad way in 2025, seriously. Allow me to bring my oauth with claude max, and/or offload whatever you do to my local claude code (or if there's some magic behind orchestration use SDK). And, oops, there's nothing to monetize and the entire thing can be recreated with vibe coding within a few weeks (it's not so much there, let's be honest).
I apologize for the writing style, don't take it personally, just every devtool product I see on HN nowadays fails for this particular reason.
So from my vantage point we need some way to create revenue. We have tried to make as much of the tool free as we can. We do a 10% markup on tokens issued by us. That's better than cursor who does 25%. We support BYOK so you can use your own claude key, and vertex key. If you do that you are basically just paying us a flat troll tole of $16/month to use our entire frontend but you are free to be unbounded by a markup on your tokens. We actually prefer this because it's better unit economics for us. So please bring your own key!
But here's the thing: most of your real audience doesn't have API keys (except a few enterprise-ish folks or startups who got credits). They already pay subscription(s) (and will continue to pay the maximum, which will keep growing). The entire token resale model creates a weird economy and interdependency that shouldn't exist in the first place. In the end, all the deals with top-tier labs will be changing, most middlemen will start manipulating the token exchange rate at some point, and there's no transparency or single source of trust. What's the Endspiel here?
Well, I'm saying I'm happy to not be involved in the resale token trade in general. I don't think enterprises are the only folks with API keys. There are a lot of people who might not be savvy or motivated enough to setup an API key with Anthropic, OpenAI or Vertex and in those cases we want them to be able to use our key to reduce user friction.
It's not in our interest to resell tokens, it actually diminishes our margin but it's a must if you want to be accessible to folks who don't want to go sign up for an API key. We choose to delay launching by a couple of days so that people could bring their own keys because we don't want to be middle men if you don't want us to be.
If you want to just pay me $16 a month to be a good IDE and inspector, I love that. That's the value I think we provide to you. We cannot control whatever manipulation of token prices occurs with other providers. All we can do is give you enough context for your prompts that you don't need the latest bleeding edge models to make edits to your react code with confidence. We aren't an AI company, we are a devtool that helps with prompting.
I don't think I'm really the best person to defend the practice of recurring revenue but I would rather be transparent that we collect a monthly fee for some features than vanity launch as a one time fee product to hacker news only to pull the rug on my customers a year later.
To be clear, we are not charging you for up to 10 prompts a day with flash or any of our IDE or inspector stuff. We respect your offline privacy and never upload anything that you don't explicitly ask us to use. We open sourced the dev server so that other folks can build JSX inspectors to keep us in check.
> What happened to installable products where you pay for a version and then you use that version.
We're a two person company with a lot of bugs and important IDE features that still need to be built out. You want auto-version updates right now. When we are in a more stable place development-wise I would love to put out a hard cut that folks can pin to.
AFAIK, every tool out there that lets you do oauth with your Claude Max plan is doing so with the same copy-pasted client id/secret that are extracted from Claude Code. It's not clear at all that this is above board, and when I asked our Anthropic rep, they encouraged us not to do this.
If there's an official way for 3rd party devs to piggyback on Anthropic plans, someone please tell me how!
fyi sometimes its because anthropic may have disallowed it. oauthing with your plan to bring the anthropic subsidy to non anthropic products is understandably a tricky deal unless you're scaled enough to make it make sense for both sides.
Yes, it's mature, but you (and your potential audience) basically need to learn a new language, a lot of quirks and "weird" (I'd even say counter-intuitive) nuances, and it's also significantly less readable in comparison with strict and typed Python. Even its modern syntax doesn't click immediately (also performance wise the new syntax somehow is a bit slower in my tests)
I am by no means a Cython fanboy but I think you are exaggerating the syntactic differences and readability differences.
Apart from type annotation they are very minor, well worth the speed benefits and type-error benefits. Given that we are discussing it in the context of SPy, SPy is not fully compatible with Python either, which is quite understandable and in my opinion a Good trade-off.
The benchmarking features are great, interactivity with C libraries is great.
One annoyance I have with Cython though is debuggability. But it's an annoyance, not a show-stopper.
It does seems very much like "worst of both worlds" in that these nuances make it harder to write than a language like Go while still being drastically less performant
It does mean. The switch from writing “applicable” software to creating cutting edge AI is almost impossible. The parent comment makes great examples, we can add to that list JetBrains (amazing IDEs, zero ability to catch up with ML), for example. It’s a very different fast-paced scientific driven domain.
Yes I'd expect a dict-based approach to do better here. That's probably how it should be done. But --long is compelling for me because using it requires almost no effort, it's still very fast, and yet it can dramatically improve compression ratio.
From what I've read (although I haven't tested and I can't find my source from when I read it), dictionaries aren't very useful when dataset is big, and just by using '--long' you can cover that improvement.
I don’t think the size of content matters, it’s all about patterns (and their repetitiveness) within, and FASTA is a great target, if I understand the format correctly
Size of content does matter. Because the start of your content effectively builds up a dictionary. Once you have built a custom dictionary from the content you are compressing the initial dictionary is no longer relevant. So a dictionary effectively only helps at the start of the content. Exactly what "start" means will depend on your data and the algorithm but as the size of the data you compress grows the relative benefit of the dictionary drops.
Or another way to look at this is that the total bytes saved of the dictionary will plateau. So your dictionary may save 50% of the first MB, 10% of the next MB and 5% of the rest of the first 10MB. It matters a lot if you are compressing 2MB of data (7% savings!) but not so much if you are compressing 1GB (<1%).
I tried building a zstd dictionary for something(compressing many instances of the same Mono(.net) binary serialized class, mostly identical), and in this case it provided no real advantage. Honestly, I didn't dig into it too much, but will give --long a try shortly.
PS: what the author implicitly suggests cannot be replaced with zstd tweaks. It'll be interesting to look at the file in imhex- especially if I can find an existing Pattern File.
I have a small/insane project of mine, I wrote a compiler for Python (strict and static subset only) to WebAssembly (bc-to-bc approach, 1:1 CPython compat due to walking internals), than I do wasm2c to sandbox it + pledge and compiling with cosmopolitan into a miniature standalone thing (fast as hell). Just because you have zero dependencies and it's a pure Python and properly typed, lemme try next weekend as PoC. No promises, but this message clicked in my heart
