realme 6

[...] It sees the general pattern in an API, then makes up an endpoint that sounds good, follows the pattern, but isn’t actually implemented. [...] I look at the readme for what return values I should get, go to use them, and get an error. [...] The LLM just through they sounded good. [...]

[...] build btproxy from source and write the two units by hand - no script and no binary from me! [...] Also fixed a few things the review shook loose, including a bug where the port restriction I added could be bypassed with a /0 CIDR while still reporting itself as locked down. [...] Pinning the hash and making the build reproducible isn't me trying to convince you that I'm trustworthy. [...]

[...] The last screenshot in the readme has the snippet I was reading that day. [...] if it’s over 18pt then 1/3 of padding is too much, but I think it works well for 12-14pt So if I’m sharing my screen with a large font I change it to mono.

Hi HN, The invoicing SaaS I was using to bill my clients raised its credit card processing fee to 6.9% for free accounts (4% on top of Stripe's 2.9% + 30c fee), so I decided to build my own to fit my minimal needs, hence I called it Minvoice. [...] Along the way of building this, I ran into an issue with env variables being inadvertently set with Cloudflare "1 click deploy" functionality, which bypassed the login, so I documented that here in case it would prevent anyone else from making the same mistake: To try it out, you can use the Cloudflare 1 click deploy button on the repo README or run "npm create minvoice". [...]

[...] That said, on a bunch of test pieces - it obviously isn't gpt-5.5, it definitely lags behind k2.6/glm/ds4-pro, but it absolutely is usable. [...] Performance as described in the tables in the readme here: ...with a bit less than half that at "low power" (30w). [...]

[...] This gives rise to lots of frustrating false-positive "bugs". [...]

For me it was gradual, then sudden. I liked using the early models to do autocompletion. It could do a leetcode style thing, pretty nice, but only useful for small things. Then I sought out Cursor because that seemed to be able to do multi-document edits. Not bad, but models at the time (2024) still got stuck pretty often. So, cross-document autocomplete. Useful, but definitely within the realm of "nice shortcuts to have". Then a friend (who works in AI) told me to try Claude last year. I was on holiday at the time, but I spun up my work repo and looked at the backlog. It chewed through the entire 6-9 months of estimated work in a two-week period while I was watching that Lord of the Rings series with a friend (we watched an episode or two in the evenings). I just chatted with him about the series while checking the progress every few minutes. It was a huge amount of refactoring, and it didn't get everything right the first time, but it made enough progress that it could be directed the right way. Since then I have hardly coded any manual lines. I just tell Claude what to do, with very little harness (skills, MCPs, instruction files), and I get what I want.

[...] The dense 27b Qwen models I found to be refreshingly good in comparison. Still not beasts of intelligence, but in my simple tests, but they’d reliably follow instructions and were honestly not that much slower than 35b-a3b! [...] Further fine-tuning for agentic work certainly sounds like a good thing, but I wonder what they are sacrificing for it. [...] That ought to tank performance. [...] The performance numbers also seem really off - on my Mac, Qwen 3.6 27b is much faster than what they claim (by multiples).

Been kicking around a similar idea in the back of my mind from the first moment "functional core / imperative shell" [0] and "sans-IO" [1] infected my brain. I've been chasing that high ever since. Unfortunately, whenever I try to apply this pattern 100%, I hit all kinds of walls: language isn't expressive enough to support what I want; the amount of wiring/glue to support it becomes a burden; the resulting code is spaghetti because the "declaration of intent" lives too far from "implementation of the intent"; "oops I invented my Nth leaky DSL"; and so on and so on. Part of the problem is certainly my own capabilities as a developer as well. I can't help but fantasize about the platonic ideal of a "perfect" system where all that nasty evil I/O is banished to the Shadow Realm and I can frolic in the Fields of Idempotency and Reproduciblity -- one of these days I'll bite the bullet and try Haskell. Nowadays I aim for 80% "perfection", and only in the areas where it matters. In addition, instead of effects I rely more on "reduce complexity as much as possible", which is (frustratingly) much harder to put into practice than "use X library/pattern to solve all problems". Though if I can model the system as a state machine and proptest it [2], that usually gets me where I want to be. Though my soul feels like I just woke up from a dream where I was perfectly content, and now I'm back in the real world with all of its imperfections [3]. --- As for your specific project, it heavily reminds me of the Crux [4] model, which is itself inspired by Elm [5]. Also Flawless [6]. I wish you the best of luck with it. [0]: [1]: [2]: [3]: [4]: [5]: [6]:

[...] I had been using it for about 6 years at that point. [...]

My phone, a Realme 6, has put in a solid 5 and a bit years of graft, but alas the battery is increasingly shagged and everything's slowing down and the storage is filling up So, new phone Quite nice, this

C^4 is a statically typed, systems programming language fixated on clean code and giving power to the user. [...] It still has crashes in certain error cases, but to my knowledge no correct code compiles with errors. [...]

Ok, done: My main focus is systems with very low ram. On my M1 with 8Gb or RAM llama.cpp cannot run gemma4 26b, but this implementation works fine with 5-6 tok/s. I also replaced my crap implementation of the oai server following yours, I hope it's all right. I did add a mention to your project and this fact in the README. It is a lot more barebones than what you have, but I have to admit that it works really nicely for me

you can tell pandoc to use typst as pdf-engine. I use this command to create pdf from my md file. Downloads/pandoc-3.9/bin/pandoc \ README.md \ -o "my-output.pdf" \ --pdf-engine=typst \ -V papersize=a4 \ -V fontsize=10pt \ -V margin-left=2.5cm \ -V margin-right=2.5cm \ -V margin-top=2.5cm \ -V margin-bottom=2.5cm \ -V title="My title" \ -V lang=de \ --include-in-header=typst-header.typ \ --toc --toc-depth=3 2&1 That's the typst-header.typ set figure(placement: none) show figure: set block(breakable: true) set table( inset: 5pt, stroke: 0.5pt + rgb("cccccc"), ) show table: set text(size: 8pt) show table.cell: it = { set text(hyphenate: true) show raw: set text(size: 6.5pt) it } set par(justify: true) show heading: it = block(sticky: true, it)