HNHacker News
TopNewBestAskShowJobs

forgotpwd16

3,569 karma · joined August 4, 2018

email: [username]@proton.me
submissionscomments
forgotpwd16··on The Sovereign Tech Fund invests in Scala
True to both. My brain not braining. Was thinking Europe-based/driven. Python started in CWI but PSF is USA-based.
forgotpwd16··on Show HN: Kandle – A WebGPU-based ML library written from scratch in JavaScript
How it compares to jax-js? Besides API preference that is.
forgotpwd16··on The Sovereign Tech Fund invests in Scala
Scala may have fallen out of favor but was quite popular few years ago. And perhaps still is the most popular EU-designed language (developed by EPFL).
forgotpwd16··on Ask HN: Why not make English computational, just like LaTeX, to ensure lock-in?
>Imagine a version of English extended

You mean restrained. More specifically what you're proposing can formally be referred to as controlled natural language with executable semantics. Some attempts similar to this have been Attempto Controlled English and ClearTalk. (And Logos that someone showed here recently.)

>text would not just require reading; it would require executing algorithms embedded in the language itself

Arguably mathematics is just that.

>just like LaTeX for scientists

Future doesn't look very bright for LaTeX with Typst getting traction.

forgotpwd16··on Show HN: I made a new compression algorithm
This is better presentation than README, which currently is marketing-heavy and technically weak. Project as an experiment is acceptable and interesting but certainly isn't "next-generation" when has (assuming benchmarks are valid) <0.2% ratio improvement to an outdated algorithm, at expense (assuming description is valid) of much worse compression/decompression speed. Note such slowdown isn't implementation detail but expected by design; neighbor graph, Levenshtein distance, edit scripts, etc, kill speed. In the end compression is trade-off between ratio and speed, and methods benchmark to both rather one.

As overall note, AIs when you prompt "apply concept X in Y" (or anything really) will tell you what a great idea and then output something that without domain knowledge you've no idea if it's correct or if even makes sense at all. If don't want to do a literature research/study, recommend at least throwing the design back to the machine and asking for critique.

forgotpwd16··on Show HN: I made a new compression algorithm
You or Claude? Have you verified/reason the claims made in README? For starters ZIP doesn't use LZW. Initially used an LZW-derived method with reset mechanism but v2 (early 90s) introduced DEFLATE, combining dict-based LZ77 & Huffman coding (which has become the de-facto ZIP compression). And even this, superior to LZW method, is not considered state-of-the-art nowadays.
forgotpwd16··on Mermaid ASCII: Render Mermaid diagrams in your terminal
Meanwhile they could've just bundled mermaid-ascii and run it on wasm. But then again probably took Claude less than an hour to port this to TS. The main algorithms are 1:1 conversion.

https://github.com/AlexanderGrooff/mermaid-ascii/blob/master...

https://github.com/lukilabs/beautiful-mermaid/blob/main/src/...

forgotpwd16··on Package management is a wicked problem
Nix is not hard coupled to systemd. There's nixpkgs support for Darwin afterall.
forgotpwd16··on Ask HN: Has Show HN become LLM-prompt-centric?
Related: https://news.ycombinator.com/item?id=46702099

But seems, as others have also noted, although the mean quality decreases, the quality of those reaching front page increases.

That said port/clones have always been trendy. What has joined their ranks is the significant amount of what are essentially minimal prompt wrappers.

forgotpwd16··on Show HN: I built a small browser engine from scratch in C++
You seem to have gotten used to perfect written commit messages by Claude. Many people, especially those starting out with, or not heavily using, git, do write messages like that. (edit: To be honest sometimes I too make temporary "wtf" commits; I just later squash them and no one knows.) That can actually be seen in this very repo by the refactor ones. And, the commits themselves are actually where I'm basing it. Small changes (vibe coders tend to forget a vcs exists; why we see 20k LOC first commits), commented-out code parts (have you ever seen AI commenting out code?), plain code comments (again compare to what refactor commits introduce).
forgotpwd16··on Show HN: I'm building an AI-proof writing tool. How would you defeat it?
Or write a script that does all that and bind it to shortcut. Eventually https://files.catbox.moe/zargud.png. (The window/tab switches were 1~2 when ended; reached 9 after few attempts to screenshot it.)
forgotpwd16··on Show HN: I built a small browser engine from scratch in C++
Excluding recent refactor commits code seems to be handwritten, maybe lightly AI-assisted.
forgotpwd16··on Show HN: Pam-db – A hybrid TUI <-> CLI tool for SQL databases
That's such cool tool. The parametrized query management/run alone is great idea, and then you've that sweet hybrid cli/tui on top. Then open repo, and was sold even more with that intro.
forgotpwd16··on Show HN: One Human + One Agent = One Browser From Scratch in 20K LOC
>I'll upload the session data probably tomorrow so you could see exactly what was done.

That'll be dope. The tokens used (input,output,total) are actually saved within codex's jsonl files.

forgotpwd16··on Show HN: One Human + One Agent = One Browser From Scratch in 20K LOC
Someone has already done this: https://github.com/viralcode/vib-OS

Also, someone made a similar comment not too long ago. So people surely are curious if this is possible. Kinda surprised this project's submission didn't got popular.

forgotpwd16··on Show HN: One Human + One Agent = One Browser From Scratch in 20K LOC
Impressive. Very nice. (Let's see Paul Allen's browser. /s) Can say is Brooks's law in action. What one human and one agent can do in 3d, one human and hundreds of agents can do in few weeks. A modern retake of the old joke.

>without using any 3rd party libraries

Seems to be an easier for coding agents to implement from scratch over using libraries.

forgotpwd16··on Show HN: Semantic-diff – understanding intent, risk and impact behind Git diffs
>the tool can't answer, but surfacing the question for the reviewer has value

That's true. Could even expand on this and make it have a --check-msg mode that returns 0/1 (good/bad) if commit messages lack the reasoning for the changes they're doing. Although, linting commit messages is last thing I want to do.

forgotpwd16··on Show HN: AnalysisXYZ – Browser-based CSV/Excel analyzer (privacy focused)
>Client-Side Processing: Your data never leaves your device.

Meanwhile

  curl -X POST -F "file=@file.csv" https://www.analysisxyz.dev/upload
  # responses with report because processing is server-side afterall
Vibe coding for real. Claiming client-side and not even checking whether indeed this holds. (Though an API-driven report-as-a-service is kinda cool; few years ago will've even been a commercially-viable startup.)
forgotpwd16··on Show HN: Semantic-diff – understanding intent, risk and impact behind Git diffs
That's commit review not semantic diff, and "why" some changes happened should be in the commit body (some will argue should be in code comments too) rather try figure it out after the fact, which won't even work. E.g.

https://github.com/tkenaz/semantic_diff/blob/main/semantic_d...

The Intent in this report just paraphrases the well-written commit message. But what about things not written? Funnily the tool catches up the drifts:

  2. Why was 196 lines of CLAUDE_NOTES.md removed? Was this intentional cleanup or accidental deletion of important project context?
  The diff shows massive content reduction in project documentation with no explanation in the commit message. This could represent loss of architectural decisions, bug tracking, or development history.
So... why? No answer is given because not possible.

  6. Why skip dependabot PRs specifically? Shouldn't dependabot updates also be semantically analyzed to catch breaking changes in dependencies?
  Dependabot PRs could introduce security vulnerabilities or breaking API changes that semantic analysis would catch. Skipping them entirely might miss important issues.
Good question. But, again, no answer.

Also maybe have an option on analysis extend, sections to see, and perhaps prompt to lessen the redundant information. E.g. in previous report, the "CLAUDE_NOTES.md massive content reduction" appears in Impact Map, Risk Assessment, Review Questions. Plus the entire sentence "Massive reduction in file size (212 lines removed, 16 added) suggests project documentation/status was significantly refactored or cleaned up" simply states what is obvious from the deleted lines in Files Changed. Some may like this, other (incl. me) won't. Personally will've preferred overall report be shorter.

The indirect impacts section seems interesting. There're cases that one may want to see modules impacted by a change, e.g. a function changing the result, making another function fail. But this is probably doable heuristically rather requiring an LLM to read the codebase.

forgotpwd16··on Kimi Released Kimi K2.5, Open-Source Visual SOTA-Agentic Model
>Kimi Code CLI is not only a coding agent, but also a shell.

That's cool. It also has a zsh hook, allowing you to switch to agent mode wherever you're.

forgotpwd16··on The Rise and Fall of Scala
(2016)

Interestingly article (this and, more importantly, Kranc's) is posted late-16 but Scala's peak was early-17. So basically Kranc's article correctly asserted the then current state. Same cannot be said for HN. In discussion of the article (https://news.ycombinator.com/item?id=12662895), everyone was calling it clickbait.

>I hope it will be replaced by languages like TypeScript, but again, I doubt it.

Well, somehow it did happened. Most web dev is TS-first now.

forgotpwd16··on The Death of Hype: What's Next for Scala
(2020)*

Scala is an excellent language. At the same time, its competitors are good enough and already far more popular. Worse, its biggest one (Java) has adopted over last decade many features, dismissing Scala's strongest suit, being a better JVM language (and with Kotlin in the picture, backed by Android, that advantage is nearly gone). Scala.js/Native though may allow a nice new goal, one language everywhere.

* Discussion at the time: https://news.ycombinator.com/item?id=22830779. Article predates Scale 3 release (2021), a transition worse than Python v2->v3 (though not Perl/Raku level), as years later adoption remains low and some people actually have disdain towards it.

forgotpwd16··on Porting 100k lines from TypeScript to Rust using Claude Code in a month
https://github.com/viralcode/vib-OS
forgotpwd16··on Show HN: The "Structural Formula" of LLMs – No more black boxes
Did this started with a prompt ~"apply physics to LLMs" on Gemini? Frankly everything reads like word salad, i.e. slop. (If it's start-to-end AI-made, not driven with domain knowledge, then can safely assume it is.) Repo is bit messy. You got a structure in README but first link got 5 files with no description which should one start with. Also, unsure what the examples supposedly convey and what role "Reformulation of the Riemann Hypothesis through Information Dynamics" has either. (Seeing N-S, RH, and a novel framework in same repo certainly raises some flags.) Attempt to actual mathematics or philosophical essay?
forgotpwd16··on Got into an argument on Discord about how inefficient CBR/CBZ is
>best girl Matsuri profile avatar

>manga-related project

No more info required to know it's good.

Beyond joking, there was/is indeed unexplored space in this niche. Quite amusing though how they went from an argument on a (yuri) manga discord to actually implementing & delivering a (technically competent and well presented) solution.

Also, got some comments. First will state two info points in regards to ZIP: (i) doesn't use solid compression (each file is compressed individually), (ii) can be used in archive-only (no compression) mode aka "stored" (compression method "0").

>Random Access

ZIP supports random access utilizing an (towards) EOF central directory and due to (i).

>If one file is corrupt, the whole thing won't open.

ZIP lacks corruption resistance but it isn't totally fragile. Due to (i) corruptions don't cascade across files; if one file is corrupted, others can be extracted; if no compression used (files "stored") is possible to recover files by headers.

>Metadata isn't native to CBZ,

CBZ (CBR,CB7,etc too) isn't really a well-defined format. It's just pictures within a zip file. (And due to this, project competes CBZ usage-wise but actual competitive tech is zip/rar/etc.) The ComicInfo file is conventional. This information can (and some readers support this) instead be saved in zip header as comment.

>BBF's content deduplication

Wonder how this works. Assume (due the comment '[...]the same "Credits.jpg", "ScanlationGroup.png"[...]') is done across files so maybe something like linked mkvs, that some anime mini-encode groups use? Then how zero-copy goal is achieved? In any case, no, ZIP cannot do this. But if files "stored", a deduplicating filesystem can handle it.

>Images inside are bit-exact copies. No re-encoding.

ZIP can do this due to (ii).

TL;DR BBF competes vs "stored" ZIP plus (conventional) header-saved metadata plus external tooling if my de-duplication assumption is correct. Issue though, and perhaps why an entirely new format is needed, CBZ is conventional and many things BBF targets to do won't hold if not design-bound.

forgotpwd16··on Competitive Pure Functional Languages
>ported a small fluid dynamics simulation program from C++20 to Haskell, Lean 4 [...]

Interesting seeing L4 be used for general programming. Others worth mentioning are:

1. Idris. General-purpose language but also proof assistant (~L4). It has few back-ends incl. C, JS, and JVM, so quite a large span of usable domains.

2. Glean. Rust-like syntax/UX, runs on Erlang VM, simple tooling (all-in-one compiler, build tool, fmt, pkg manager). Very modern-looking overall.

forgotpwd16··on Nushell: The Shell That Treats Everything as Data
What? Got lazy and had AI post your comment rather do it yourself so ended up in wrong thread? (edit: Assume was for your Show HN (devto is lowrank/banned domain; vouched but still dead*). Amusing post/MCP application. Surprised seeing Rails nowadays. Why not used Ruby for MCP part too? As extension can compare how well LLMs play compared to heuristic/algorithmic solvers.)

*If comment wasn't auto-posted, maybe a bug due to this? No thread for text to be attached so was dropped as comment to the next/previous live submission?

forgotpwd16··on Show HN: Free PDF Editor by TechRex – client-side PDF editing, OCR, compression
Note, there's has been a wave of similar sites recently (2mo+ after you first submitted yours). See: https://news.ycombinator.com/item?id=46675926.
forgotpwd16··on Show HN: Klaus – a Claude Code native delegating agentic harness
Confused. Is this a software project, a specification for one, a prompt for Claude Code to read, an essay, ...?
forgotpwd16··on Vibe coding kills open source
Jujutsu?
← PreviousPage 4 of 34Next →