Even more important: stop using personas, start using actual people. I've experienced many startups make unforced errors by conflating people into personas. A better way is to tag people with attributes, such as specific interests, explicit concerns, tasks to be done, usage goals, learning preferences, and the like.

When you switch from personas to actual people, it opens up many more product experiments-- many of which are surprising and may even feel counter-intuitive to founders. Increase your startup chances of success by carefully connecting with your actual users.

I'm seeking your help please for suggesting ideas, articles, tactics, and the like for improving ways of working - especially if you have experience and opinions.

Comments URL: https://news.ycombinator.com/item?id=48600733

Points: 2

# Comments: 0

Also shoutout to ratatui because even though it's technically a TUI not GUI, it's superb.

Comments URL: https://news.ycombinator.com/item?id=48472677

Points: 2

# Comments: 3

https://crates.io/crates/assertables

I would like to donate toward getting Assertables into Debian stable. I can afford to offer $100 to any Debian Developer who has the suitable skill and trust to accomplish this.

Success looks like this: long term, Rust crates in Debian stable can optionally use Assertables for their testing i.e. all the source code is available in Debian.

I know $100 isn't enough to cover all the work. It's what I can do to help move things forward.

I'm doing this because I believe in funding open source as much as possible.

Comments URL: https://news.ycombinator.com/item?id=48409145

Points: 2

# Comments: 0

Setup is simply 3 steps:

1. Sign up on each service, ideally with the same username.

2. For each repo you want to share, create the same repo name as a blank repo; do not automatically create a README.

3. Edit your local file .git/config to add push URLs, then push as usual.

Example:

    [remote "origin"]
        url = git@github.com:foo/bar.git
        pushurl = git@codeberg.org:foo/bar.git
        pushurl = git@github.com:foo/bar.git
        pushurl = git@gitlab.com:foo/bar.git
        fetch = +refs/heads/*:refs/remotes/origin/*

The Rust crate: https://crates.io/crates/assertables

All the macros have forms for different outcomes:

    assert_gt!(1, 2) // panic
    assert_gt_as_result!(1, 2) return Result (Ok or Error)
    debug_assert_gt!(a, b) // panic when running in debug mode

I'm seeking to stop supply chain attacks as described at https://wiki.debian.org/Rust

Could there potentially be a way for a program to include all the source code of all its dependencies, at least for any that aren't on Debian?

The concept is called a "package manager" and it greatly helps me achieve a concept called "reproducible builds", as I gradually migrate the work toward a package manager for reproducible builds called "Nix".

https://en.wikipedia.org/wiki/Package_manager

https://en.wikipedia.org/wiki/Reproducible_builds

https://en.wikipedia.org/wiki/Nix_(package_manager)

I want to use Fresh Editor (https://getfresh.dev/). This is because Fresh is easier to learn quickly than vim & emacs, and more of an IDE than micro & nano, and more favorable for compliance auditing because of its core-and-plugins architecture and open source code using Rust.

I would like to donate toward this goal. I can afford to offer $100 to any Debian Developer who has the suitable skill and trust to accomplish this. The software author must say yes too, of course.

Success looks like this: create a new Debian 13 server, type "apt install fresh-editor", and it works. No third-party apt sources needed, because I have compliance reasons. Fortunately Fresh already has a *.deb file and explains it on the Fresh installation page.

Here's the GitHub issue: https://github.com/sinelaw/fresh/issues/2169

I'm not affiliated with Fresh in any way, just a early adopter user. If you view the GitHub issues, I'm also offering $100 each for helping with Fresh display issues on MacBooks, and for adding Fresh ePub capabilities.

I'm doing this because I believe in funding open source as much as possible.

Comments URL: https://news.ycombinator.com/item?id=48347180

Points: 29

# Comments: 14

Comments URL: https://news.ycombinator.com/item?id=48316470

Points: 3

# Comments: 0

AI agents, provided as markdown files, that implement United Kingdom (UK) Government Digital and Data (GDAD) Profession Capability Framework (PCF) roles and their skills.

Purpose: The United Kingdom is implementing a jobs classification system for digital technology work, such as "Software developer", "Test analyst", "Incident manager", etc. The UK provides extensive documentation about the roles and the skills, along with naming conventions.

I'm experimenting with multi-agent software enginnering, so I extracted the documentation into markdown files that I can use as AI skills.

It turns out this works very well in practice, with equivalent results to popular multi-agent generic systems. The big win is that using the UK naming convention and exact roles and skills makes teams understand it immediately - including executive teams, goverance teams, and legal teams.

The markdown files are especially easy for novices to copy/paste into AI chat systems, and also can of course work with more-advanced tooling e.g. I use Claude Code and swarms of subagents.

I'm sharing the markdown files in this repo. Constructive feedback welcome.

Comments URL: https://news.ycombinator.com/item?id=48305822

Points: 2

# Comments: 0

AI use is mixed: I write everything by hand in the specifications, and the components.tsv file, and in the Svelte version because it's my stack of choice. If you dig in, you'll see lots of hierarchies that are all hand-written to help Svelte caching, for example. I also research other major design systems, especially government-oriented public-sector systems such as GOV.UK and Reuters, and fold them in.

Then Claude Opus tranforms the Svelte version into the other stacks. Claude does a lot of the documentation text because I'm aiming for clear and consistent explanations, suitable for novice developers.

I'm developing Lily Design System because I work with multiple teams that each use a different tech stacks, each with their own ad hoc HTML tag names and semantic names. As a salient example, for a hospital form one team used terminology "health banner area" and another used terminology "medical red box". Lily is my attempt to converge these into something that works better internationally and across multiple stacks.

That said, I'm seeking help doing human proofing and improvements for the stacks, because we all know AI isn't perfect, and needs tuning, guardrails, expert feedback, and the like.

Comments URL: https://news.ycombinator.com/item?id=48274749

Points: 4

# Comments: 2

See https://github.com/bugzmanov/bookokrat

Update: a developer just now created a Fresh IDE pull request, thanks to this incentive.

See https://github.com/sinelaw/fresh/pull/2093