HN user

tajd

85 karma

https://tom-dickson.com - personal

https://verdient.co.uk - software development and consultancy

Posts5
Comments89
View on HN

A cloudflare hosted task and wiki deployment - https://tajd.github.io/projektor/ - free to host because it's on your infra on cloudflare. Built it because I could, and it was simpler and cheaper than using atlassian etc.

A tool for enforcing code architecture and conventions - https://tajd.github.io/cofferdam/ - these sorts of tools come along every week, but this one is mine and it's v quick to run and extendable.

And then I'm currently working on a game for winning the start of sailing races as that's quite a tactical and fun problem to break down. Will be releasing something there soon!

Personally I apply it every so often to audit codebases but it’s mainly sonnet and Gemini for most tasks. That and a harness for an agent to keep iterating on a goal is perfect. I’m more concerned buy getting hooked on the 50% increase on all models, not just Fable.

ah epic - please drop an email to the address on the website and I'll give membership to you and whoever else would like to take it for a drive

I'll try and get your feedback addressed today/tomorrow!

A few things:

- https://www.ironvolume.com/ for generating crossfit, hyrox and athx workouts and warm ups, with a few posts on topics I've found really helpful to train around.

- https://pokerchallenges.com/ for practicing the maths behind different aspects of playing texas hold'em poker, trialling it at the moment so can give anyone a free membership if they want to get to grips with it

- https://typst.app/universe/package/calendaring/ a typst calendar grid generation package

- https://github.com/TAJD/cofferdam a tool to help with implementing code architecture compile time checks, helps give agents relevant context on how well their code fits in with the wider codebase without them having to read it all

I built https://www.ironvolume.com/ to help plan workouts for crossfit, hyrox and building it out for athx.

https://tajd.github.io/cofferdam/ - is an experimental compile time time checker that can be run for typescript/rust to be able to impose architecture, design principles that can be forgotten by agents. The goal is to be able to make it possible for an agent to be reminded about design conventions where it forgets, so it reduces the amount of context that needs to be used initially to define those sorts of conventions.

A redesign that gets replaced 2 years later is a catastrophe

I mean, given how quickly things can change I think the language and sentiment here isn't quite right, it's just how businesses can change and we can't necessarily control that.

There has got to be opportunities here for abstracting over regulation to make it easier to comply with and prove compliance so that risk owners/govt can enact change faster. Now to figure out who would pay for that.

I feel like articles like this need to come with a diagram like this to put it in the context of relevant tradeoffs.

                       High Scale/Revenue
                              │
                              │
      Managed Services        │      Self-hosted K8s
      (Overpaying but         │       (article is
      no choice)              │       pitched here)
                              │
  ────────────────────────────┼────────────────────────────
      Low capacity            │         High capacity
                              │
                              │
      Managed Services        │      Managed Services
      (Right choice -         │      (Wasting money on
      focus on product)       │       platform team)
                              │
                        Low Scale/Revenue
Or something like that. Maybe as a function of time as well but that might be harder to do in 2d.

Sure I can absolutely manage my own k8s, but there is no doubt it's easier for me to spin up postgres and ship faster on my own. At enterprise scale it's definitely a lot easier to do everything in k8s and be able to manage as many aspects as possible. I have experience of both.

I ported my personal website from Jekyll to Astro a few weeks back and I really liked it. Astro is much easier to build and extend for me (and that is a personal preference point - I (and by I mostly mean claude) - and it's cool to add react components in to create more interactive points (but I haven't deployed that element yet).

Speed is probably the same as jekyll - but relative to my react vite and nextjs apps it's about 10 times faster.

I would definitely use Astro for more complicated websites than content driven - but would probably return to nextjs or more hefty full stack solutions for complicated web apps.

I think hopefully it raises that awareness that these systems are imperfect and do not allow people to delegate their decision making totally. I'm afraid rise of AI and the increase in more plausible sounding yet incorrect behaviours won't help us there. We should use this as an instructive example on how we can built more transparent/useful systems for all of our users.

This looks really cool! I can't remember what article I was reading the other day but it was advocating for local first software as being the true goal of creating user friendly software -> this app absolutely meet's that goal. Looking forwards to using it.

I really like using UV. I introduced it at work for builds and it made everything a lot faster which was awesome. Now I can remove the other components of the build process and just use one.

I am interested in how they're going to make money eventually, but right now it's working for me.

Does anyone have an idea about how they're going to monetize?