Hacker News Reader

jondwillison Ask HN: What are you working on? (September 2026)

Remaking the Final Fantasy XI client. The idea is to have a ~0.9:1 faithful reimplementation of the retail client without a lot of the performance jank and cruft that the original has, and then layer optional improvements, QoL, graphical improvements, UI/UX improvements etc., on top with a plugin system and hackable source code for players across various private servers.

This approach allows people that just want to replace their familiar client with something slightly more modern/performant/cross-platform, all the way to entirely new experiences.

There's at least one other project, not yet open-sourced, which is going more in the direction of adding all of the enhancements by default.

https://github.com/jondwillis/kuluu-ffxi

jondwillison Show HN: ClaudeStatsBar: your session is 486k deep and nothing told you

Could’ve asked Claude to put this in your status line.

https://code.claude.com/docs/en/statusline

jondwillison Hook, hold, harvest and hide: Meta's alleged strategy laid out in first week

Not that anyone is asking the question that you’re answering, but what’s amazing about their talented young team?

jondwillison AI Coding Without the Vibes

If your harness has an “AskUserQuestion”-like tool, you can prompt it to use that, or, as I have done, add a stop hook that uses a smaller model to check if there was any similar prose that was output toward the end of the previous response.

Happy to share the implementation or package it up if anyone is interested.

jondwillison [dead]

README is total slop, does not seem related to title… references videos in the first sentence… what videos???

There might be a “there” here but I think you need to put in more effort and thought about your audience, AI or not.

jondwillison Build Wide, Ship Narrow

Trauma dump/sign of the frothy times:

I got fired back in June for protesting against trying to “design everything up front” (by adding more slop plan docs alongside the hundred of others that my tiny team was charged with producing per month) and instead opting to explore the problem space through code/experimentation. In actuality, I just asked to be remote again. I was burnt out and they were looking for an excuse to fire me.

jondwillison OpenAI’s head of ethics leaves less than a year after joining

In this article: no explanation of why she left.

jondwillison DOGE overstated savings, government watchdog says

Accelerationism more like. Not just cutting aid to the poor.

jondwillison OpenAI's New Device Will Be Hockey Puck-Sized and Cost over $300

https://maps.apple/p/0mZU6qFyXAEYRe

Yeah, Apple would never design a donut headquarters and make their employees live inside of it /s

jondwillison Lines of code. 1,596 BTC gone

Since this has become a thread about complaining about unreadable, brain-killing AI slop, I'll add mine:

>Six steps, no jargon. The note beside each one is the precise technical version, for anyone who wants to check the work.

I think this is called "concept leak" or "prompt leak" (please someone correct me, I can't recall at the moment.) It is one of my least-favorite failure modes. I notice it a lot when drafting landing pages, and it eats up a lot of time to edit away.

jondwillison Ask HN: Who wants to be hired? (August 2026)

Location: San Diego, CA

Remote: Yes

Willing to relocate: Yes, but not before November. Happy to go hybrid or onsite after that.

Technologies: TypeScript, Vercel AI SDK, Workflow SDK, MCP servers, Next.js, React, Node/Bun/Deno, NestJS, Postgres (Drizzle/Prisma/TypeORM), Redis, BullMQ, OpenTelemetry, Swift/SwiftUI, some Rust and Python.

Résumé/CV: https://jonwillis.dev/?utm_source=hn | https://github.com/jondwillis

Email: jw-port-burner@proton.me

I'm an AI product engineer and sometimes-founder w/ 15+ years of experience across full stack web, native, mostly startups. I have been building agent systems and I have been at it long enough to know where they break, and I am a lifelong continual learner and striver.

Currently co-founder and CTO at Valara, bootstrapped and pre-revenue, where I built an AI-native orchestrated workflow product that reviews commercial and residential real estate appraisals for lenders. I turn my co-founder's and our design partners' experiences into product features.

Before that: a little more than two years at Uni Industries building AI, agent, and data plugins across two products, including an agent harness w/ structured tool use, skill systems, and eval pipelines.

In 2023 I shipped waggledance.ai, an open-source multi-agent system with DAG-based planning and parallel execution, before things like LangGraph or reasoning models existed. Earlier I was senior iOS at Curtsy (crash-free sessions from 99% to 99.9% at hundreds of thousands of MAU) and the first mobile engineer at Pluto TV, pre-seed through Series B, with the mobile/OTT team that I help lead there driving a lot of the early traction and growth, which eventually led to an exit in 2019.

---

I'm looking for a former-founder, senior, or staff IC role building AI products people actually use, or will use soon. Strongest on TypeScript, frontend, UX, agent architecture, and SDLC/ADLC.

jondwillison Agent skills that bring team coding standards to Claude Code and Codex

I’ve found hooks to be really useful for correcting or inducing certain behaviors/outcomes that tend to occur during agentic development.

For example, I’m working on a project to remake the Final Fantasy XI client. My repository has a bunch of git submodules that reference other peoples’ related efforts, and an open source server. For example, despite including CLAUDE.md to suggest otherwise, Claude Code always ends up writing these insanely dense comments referring to specific files and lines of code in submodules. When those submodules update, now the comments are no longer correct.

So I added a hook to detect when comments are included. Then, for example, I have a deterministic heuristic and script involved to remove some kinds of comments, and another that asks Haiku to quickly LLM-as-a-judge whether or not to edit/remove the comment.

Similarly, I have a stop hook that reminds Claude to commit its code logically on main, noting that other changes may have been added by other concurrent sessions (I avoid worktrees and even branching for this particular project and stage of development.) It works well.

jondwillison Ask HN: Who is hiring? (August 2026)

FYI, I'm interested but trying to drill into any of the specific postings is reloading the `/open-positions` page for me (Safari, connected to a VPN.)

I'll go ahead and accept that `Staff Software Development Engineer in Test, Cloud Applications` role, thanks ;)

jondwillison LearnVector – Andrew Ng's AI company building one‑to‑one learning experiences

/frontend-design

jondwillison How to stop Claude from saying load-bearing

A coworker started spamming this word in ~April while working on system design/architecture.

jondwillison Show HN: Jacquard, a programming language for AI-written, human-reviewed code

Thanks for the clarification. “Mode collapse” explains so much about how I have been experiencing output convergence and repetition for a while now.

jondwillison Australian energy retailers must offer three hours of free daytime electricity

What is their fire suppression setup like??? Granted I guess they could be doing pumped hydro storage lol

jondwillison Show HN: Jacquard, a programming language for AI-written, human-reviewed code

lol I had started a “plugin compiler” that Claude also wanted to name after the Jacquard loom https://github.com/jondwillis/jacq

Fricking training distribution/same-y ness coming for us all… cool project though.

jondwillison You shouldn't copy-paste errors into Claude Code

You should edit the article to suggest this point, it may not be obvious to everyone reading it.

jondwillison Slow breathing modulates brain function and risk behavior

Ah, I suffer from bruxism and TMJ. I’m sorry you’re dealing with this condition.

jondwillison 45°C cooling design cuts data center water use to near zero

Unironic Paperclip Maximalism. Huh.

I propose that you don’t actually want cheap tokens. You want cheap tokens in service of some other goals. Are they worth the environmental cost?

jondwillison Slow breathing modulates brain function and risk behavior

Cmon tell us what it is

jondwillison Claude Desktop spawns 1.8 GB Hyper-V VM on every launch, even for chat-only use

Yeah unironically just let an agent harness rip with full admin access without monitoring anything it does or using a VM. It’ll be fine, probably. “How I Learned To Stop Worrying And Love the AgentDOS and Only Exfiltrate Secrets Occasionally”

jondwillison Show HN: Lowfat – pluggable CLI filter that saved 91.8% of my LLM tokens

I think GP is basically saying, bitter lesson applies here.

jondwillison Claude Code and Codex can have real-time conversation via Git

Low effort comment: reinventing things that Unix provides is the backbone of the entire western economy.

jondwillison Claude Code and Codex can have real-time conversation via Git

This ACP has been rolled into A2A (https://a2a-protocol.org/ ) and shouldn’t be confused (it’s happened multiple times for me, again just now) with Agent CLIENT Protocol https://agentclientprotocol.com/

jondwillison Meta's ships facial recognition on smart glasses

I don’t think it’s reasonable to think that privacy regulation is coming with the current admin and weak “opposition”, and lack of capacity to care or understand the stakes by the populace. We need some very effective political organizing, yesterday. And normalized, cheap, scramble suits wouldn’t hurt either.

jondwillison First U.S. screwworm case confirmed in South Texas

And that’s the entire point. Accelerate!

jondwillison Bitcoin Has Longest Losing Streak Since August in Bruising Week

Halvings? I guess?

jondwillison Ruby vs. Java vs. TypeScript: my experience on building a Cowork DOCX plugin

> I will check out if Claude Code itself has MCPB support.

It doesn’t.