Hacker News: jeremyjh

New comment by jeremyjh in "Why Software Factories Fail (or: harness engineering is not enough)"

jeremyjh — Thu, 23 Jul 2026 20:28:35 +0000

This entire article is about the fact that the codebase deteriorates over time when you work this way. If they are correct, it will take longer and longer to develop PRs, and more and more cycles to QA them. Eventually the code will be unsalvageable and it might get there before anyone notices a problem. I don’t know if they are right but I know codebases can get there with humans at the helm - and the humans can adapt and learn.

This is not a certain fate but it sure looks likely and is a serious risk. Some software is so low stakes that you can get away with it I’m sure - there are always trade offs.

New comment by jeremyjh in "Startup founders urge U.S. government not to shut off Chinese open weight AI"

jeremyjh — Thu, 23 Jul 2026 19:24:41 +0000

Is that why Linux failed? Just imagine the world in which an open source kernel dominated almost every compute market on the planet!

New comment by jeremyjh in "Qwen-Image-3.0: Rich Content, Authentic Details, Deep Knowledge"

jeremyjh — Tue, 21 Jul 2026 12:07:00 +0000

Yes and even if all they do is ask it to stage an empty room picture with furniture and decorations it will make improvements like adding a doorway to a room that doesn’t exist.

New comment by jeremyjh in "Claude Code uses Bun written in Rust now"

jeremyjh — Sun, 19 Jul 2026 16:43:06 +0000

I know, its so confusing. Its almost as if the bun rewrite to Rust has absolutely fuck all nothing to do with Anthropic's product strategy.

New comment by jeremyjh in "Claude Code uses Bun written in Rust now"

jeremyjh — Sun, 19 Jul 2026 16:34:34 +0000

Bun was not ported to Rust for performance reasons and its not clear to me how anyone can think it was. The reason they've given is memory safety.

New comment by jeremyjh in "OpenAI reduces Codex Model Context Size from 372k to 272k"

jeremyjh — Sun, 19 Jul 2026 15:25:44 +0000

Yes the idea is cool but this could really hammer usage, especially just leaving it up to the agent to decide when to do it. I'm not surprised though, considering the github account is named "Vibecodelicious" and became active in December.

Poking around in the repo the whole implementation is an unsupervised LLM fever-dream.

New comment by jeremyjh in "Gleam Is Now on Tangled"

jeremyjh — Sat, 18 Jul 2026 22:50:55 +0000

The more important thing about issues & PRs is that they explain why the code is the way that it is. Knowing why the code is the way it is, is almost the entire reason we have a repository to begin with. They belong to the code, not some social graph.

Yes, code is largely about communication as well - with others as well as our future selves. That doesn’t mean it’s a social media experience. There are lots of ways to communicate that aren’t that.

New comment by jeremyjh in "Gleam Is Now on Tangled"

jeremyjh — Sat, 18 Jul 2026 17:43:49 +0000

Issue and PR content are not social artifacts, they belong in the repo or at least its hosting instance. Likes and activity feed are the sort of thing that belongs in the social layer.

New comment by jeremyjh in "Gleam Is Now on Tangled"

jeremyjh — Sat, 18 Jul 2026 16:47:46 +0000

This looks like a much more sensible design for code repos: all the artifacts live in the repo.

New comment by jeremyjh in "Gleam Is Now on Tangled"

jeremyjh — Sat, 18 Jul 2026 16:44:26 +0000

It also has a pretty fundamental design flaw: issue /PR comments belong to the server where the commenter is hosted, not to the repo. I’m sure they will find a workaround but finding that reduced confidence they actually understand the problem they are solving.

New comment by jeremyjh in "Kimi K3: Open Frontier Intelligence"

jeremyjh — Fri, 17 Jul 2026 01:09:56 +0000

If they were a few weeks behind, Moonshot would release a Fable level model in a couple of weeks.

New comment by jeremyjh in "Sony deletes more movies from the accounts of people who ‘bought’ them"

jeremyjh — Fri, 17 Jul 2026 00:09:52 +0000

Movie distribution is a completely different game - the studios and distributors have far more leverage. These decisions are not made by Sony, and they can’t negotiate contracts that would give them that power for all content people would like to see.

So, given they have control over their game platform and the leverage to dictate these terms (and Steam has the same terms) - on what basis is it “extremely likely”?

New comment by jeremyjh in "Sony deletes more movies from the accounts of people who ‘bought’ them"

jeremyjh — Thu, 16 Jul 2026 20:29:32 +0000

Has Sony yanked a game? I know they have delisted some but I have not heard of one ever being removed from a user’s library and couldn’t find a reported instance of that happening. Sony’s agreement with game publishers does not leave them the right to do this.

I do not think there is any online platform that can make such guarantees for movies.

New comment by jeremyjh in "Sony deletes more movies from the accounts of people who ‘bought’ them"

jeremyjh — Thu, 16 Jul 2026 18:31:15 +0000

Yet people have been “buying”games from Steam the same way for more than two decades…

New comment by jeremyjh in "Prioritize mental health, and why communication is so important"

jeremyjh — Wed, 15 Jul 2026 21:09:16 +0000

Your description does sound like ADHD to me but even a diagnosis and treatment is only one part of what you’ll need. In addition to a therapist I think you really need a mentor to help you navigate the professional side of things. There are lots of shapes that can take but I would definitely prioritize it; if a lot of your stress and unhappiness is due to work there is a limit to what most types of traditional therapy can help with.

New comment by jeremyjh in "SpaceX bond worth 10% less than issue price – heading for junk bond status"

jeremyjh — Wed, 15 Jul 2026 15:43:42 +0000

This is about a bond issue - a large drop in value there so soon after issue is far more unusual and much worse news.

New comment by jeremyjh in "Cursor 0day: When Full Disclosure Becomes the Only Protection Left"

jeremyjh — Wed, 15 Jul 2026 03:14:34 +0000

No, read the report. It will call an executable called git.exe in the untrusted repo.

New comment by jeremyjh in "Cursor 0day: When Full Disclosure Becomes the Only Protection Left"

jeremyjh — Wed, 15 Jul 2026 00:57:40 +0000

> syntax highlighting, typechecking, linting, automatic git refreshes

In most languages, none of those things involve execution of code in the repo. In languages that do - for example Elixir - it prompts you to trust them first.

New comment by jeremyjh in "Cursor 0day: When Full Disclosure Becomes the Only Protection Left"

jeremyjh — Wed, 15 Jul 2026 00:55:56 +0000

Cursor is an editor. Why is it executing untrusted code when a folder is opened? Opening a folder is not equivalent to executing a script contained in the repo. How can you think that is ok?

New comment by jeremyjh in "Cursor 0day: When Full Disclosure Becomes the Only Protection Left"

jeremyjh — Wed, 15 Jul 2026 00:48:49 +0000

You think opening a git repo I just cloned should execute code in that repo?