Chameth.com - posts like a-year-of-boardgames, debugging-beyond-the-debugger, infinite-avatars, why-you-should-be-using-https

Building a family of single-user services

2026-08-27T00:00:00Z

Last month I made and wrote about a personal service to track my workouts and record personal bests. Since then, things have escalated. I now have eight different services running in the same “family”:

Name	Purpose
Collegiate Room	Simple weather overview for my current location
German Mall	LLM broker for use by other services
Gloss White	Weight tracking
Olive Crosses	Central service hub
Pompei Band	Workout and personal best tracking
Speed Carp	Spaced repetition system
Tiger Island	Reading history and library tracking
Typical Horse	Location history and analytics

The names are all anagrams of something related to their function. Finding those anagrams brings me an unreasonable amount of joy. Unscrambling them is an exercise left to the reader. On top of those eight, I currently have four other service ideas sat in my todo list. It’s going to be a big family.

A postmortem on the previous hodge-podge

Most of this functionality isn’t new to me. When I started out playing around with LLMs, I created a “personal data aggregator” that just kept getting things bolted on. I said at the time:

My biggest just-for-me project is a web app that I started to help me aggregate film recommendations. It’s since morphed into a general personal data aggregation service: it deals with data from GitHub, Todoist, Letterboxd, TMDB, Healthkit, and others. It also lets me make re-orderable lists, store recipes, and more. Parts of this could definitely be open sourced, and I might carve them out at some point, but it’s mostly a glorious hodge-podge of things specific to me. Having all these services in one place lets me make quick and dirty automations, for example: when I create a Todoist note on my phone or watch, I often forget to set the due date, so it doesn’t show up in the “Today” view. It was literally a few lines of code to plumb things together so any inbox task without a due date gets set to today automatically.

In a twist that nobody could see coming, it turns out that a “glorious hodge-podge” is actually quite hard to maintain and very annoying to work with. It didn’t help that it had a fairly extensive Svelte frontend and I really am not a fan of JavaScript. I ended up using it less and less just because it wasn’t fun to use or work on, and I didn’t want to waste the limited LLM tokens on doing massive refactors or reworks.

I’ve got almost two decades of professional software development experience under my belt, so you’d think I would have learnt the lessons on how to write maintainable software. Hell, my CV even mentions that as something I’m particularly good at. So what went wrong?

There were a few factors. One of the obvious things to blame is that it was written mostly with LLMs. Most of that blame isn’t very honest, though. In the year or so since, LLMs have definitely got better at writing good code; but I was manually reviewing all the code at the time, and telling it quite carefully what to implement and how to do it. So the buck really stops with me.

I think it’s actually more because it was my first larger project using LLMs. It’s a bit like learning a new programming language. I don’t think the first sizable project I’ve written in any language has ever been any good. You can read and learn as much as you want, but you need to actually run into some problems headfirst to get a grip on how the language works. Or at least I do. LLM coding agents are basically the same deal.

What the LLMs actually did was mask several problems. The architecture wasn’t right for how the service developed. It started out as a single-purpose app and gradually got things bolted on, without any major changes. Everything ended up in horizontal slices: a massive database layer with every query used by every part of the app, a massive HTTP layer, and so on. Working on a single feature became a tap dance all over the source tree. This sort of friction is normally quite apparent when you’re writing code by hand, and at some point it hurts enough that you throw in the towel and do some refactoring. But the LLM will happily plod on without exposing that pain, so the project needed a more active kind of architecture management that it just didn’t get.

The other problem they covered up was how brittle and unmaintainable the JavaScript frontend was. I’m not even sure how I ended up with a Svelte frontend in hindsight. It’s my preferred framework, but I still only reach for it as a last resort when there’s no other way to avoid using one. It seems like there were lots of ways to avoid it here, but I seem to have… well… avoided them. Anyway, each feature got its own slew of Svelte components, without much thought to reusability, shared patterns, or anything sensible. Some of them were just thin wrappers around data supplied by Go, others had vast swathes of logic in them, with no real rhyme or reason as to which way that would fall. It’s the same issue as before: a human would look at it, say “WTF?” and suggest tidying it up; instead, the LLM just gets on with what you tell it to do.

There’s also a big non-LLM factor: one of the primary reasons all of this unrelated functionality got tossed into the same app was that I’d previously worked on a few projects with a microservice architecture and really didn’t like it. My conclusion was they only really worked in large environments where you can dedicate people or teams to working on all the boilerplate that happens from maintaining lots of interconnected projects. Dumping everything and the kitchen sink into one badly architected monolith was, perhaps, not the best rebound from that.

Pompei Band and beyond

When I wanted to implement ~~Strava at home~~ my PB tracking app, it seemed fairly natural to build it out as a new service, rather than bolt it on to the slowly decomposing hodge-podge that was giving off a funny smell. Then when I wanted a weather dashboard, why not do it in the same style as Pompei Band? The same serving architecture, the same page layout… I could even come up with a similar weird two-word name!

Then wouldn’t it be nice if the weather dashboard knew where I was, so it could show the local weather? I could just build that functionality in, but it’d also be a good opportunity to carve out the location history feature from my old service and make a new, shinier one. Then basically every time I even thought about going near the old service, I mentally sketched out a replacement service.

Olive Crosses showing links to the other services

At one point I had the thought that it’d be nice if the services could all link to one another, but I didn’t want to have to touch every service whenever I added a new one. So I made a “hub” service, which maintains the One True Catalogue, and provides a bit of JavaScript the others can all embed to add a navigation sidebar.

So after complaining about microservices and JavaScript, it sounds like I have landed on microservices (or at least miniservices) and JavaScript? Sort of yes, but sort of no. Each service is still its own thing and makes sense independent of the others. There are some cross-service dependencies, but they’re very simple and don’t have the same cognitive or boilerplate overhead as a login service that talks to an authentication service that queries a user manager and so on. And as for the JavaScript, it’s almost exclusively used for progressive enhancement and a few limited client-side visual things like maps and graphs. All of the logic lives in Go, which is a lot easier to maintain and test, and all of the pages are simple Go templates.

Having these services built in a way that I actually like has motivated me to play around a lot more with new features:

My workout tracking app pulls photos from my private Immich server that were taken during workouts, and shows them alongside. Now I can see all my obligatory end-of-run selfies in one place.
The location history app can detect trips away from home, and get an LLM to automatically title them. My most recent, very exotic, trip has the title “A fortnight in Milton Keynes and Coventry”. Each trip gets its own little map, and in the future I want to pull in images like with workouts.
My weight tracking app now exports weigh-ins to Garmin, as it doesn’t fully integrate with Apple Health (it will export its own weights, but not import others, like the ones from my Withings scales). Stupid walled gardens, but thankfully I can fix it myself.
My book tracking app automatically cleans up the titles of books from the nonsense provided by Amazon to something more palatable. “Killing Eve: Long Shot: A twist on obsession, betrayal and revenge from Luke Jennings, author of the TV sensation Killing Eve” becomes “Long Shot (Killing Eve #5)”. It does this by passing the title to an LLM with a bunch of rules.
The hub generates me a snippet of Tailscale configuration that defines the minimal grants needed to allow the intra-service dependencies to work.

The LLM integrations were inspired by my friend Shane who recently demonstrated an inventory tracking app he’d made, which used Haiku to process images of boxes and describe the contents. I’ve dabbled with LLM integration before, and always disliked the process of fiddling with the prompts, and debugging when the LLM gave a nonsensical response. So I built German Mall, which makes it a lot easier. It lets me define a “workflow” consisting of a system prompt, initial message, and expected response format. I can browse past runs, tweak the prompt, rerun them to see the change, all without touching the services using them. Quick iteration loops are the best. German Mall also provides markdown documentation for each workflow, so integrating it downstream is simply a case of pointing a coding agent at the right documentation endpoint.

The “tidy book titles” workflow in German Mall

It’s not all fun and games: having a bunch of services all following the same pattern means you need to do cross-cutting maintenance on them now and then. But that’s surprisingly easy with LLMs. I put all the projects in a common folder, and I launch cross-cutting tasks from that folder. Usually I’ll make the change in one project, and then the cross-cutting task is simply “Look at commit XYZ in ./foo, make the same change to the other projects”.

I’ve also ended up with a shared library used by all the projects to cut down on some boilerplate. That’s normally something I’d resist, especially with personal projects, just because the extra overhead of working in a separate project, managing releases of it, and so on, is a bit annoying. But, again, LLMs can do almost all of that grunt work. They effectively reduce the “cost” of maintaining separate services to a negligible level, leaving me to focus on the more fun aspects like coming up with new names.

Making a font of my handwriting

2025-08-08T00:00:00Z

Recently I’ve been on a small campaign to try to make my personal website more… personal. Little ways to make it obvious it’s mine and personal, not just another piece of the boring corporate dystopia that is most of the web these days. I don’t quite want to fully regress to the Geocities era and fill the screen with animated under construction GIFs, but I do want to capture some of that vibe.

I’d added some bits and pieces along those lines: floating images in articles now look like they’re stuck to the page with sellotape, related post links have a wavy border that animates when you hover over them, and so on. Next, I wanted to change the heading fonts from a monospace font to something cursive, to resemble handwriting. Less terminal output, more handwritten letter. I couldn’t find one I liked, though. So why not make my own? It can’t be that hard, right?

Failing to do it myself

I set out to try to make the font myself using open source tools. After doing a bit of research, it seemed like the general approach was to create vectors of each character and then import them into a font editor. That seems to mean either Adobe Illustrator and FontLab (if you have too much money) or Inkscape and FontForge (if you like open source). I fall firmly into the latter category, so I grabbed my graphics tablet and opened Inkscape.

I wrote out my first three letters: capital A, B and C. Saved them in Inkscape, and attempted to import them into FontForge. Then I remembered one crucial thing that had slipped my mind: I absolutely loathe using FontForge. It’s a bit like when you open an old version of GIMP and get a bunch of weird looking windows floating all over the place; it feels like you’re fighting against the tool to do even the most basic operations. The difference is I have cause to edit images a lot more than I edit fonts, and GIMP has actually significantly improved their UI over the years.

Here are the rough steps I went through with FontForge:

Launch Font Forge. It shows a weird bit of art in one window, and an open file dialog in another.
I don’t want to open a file, so I close that dialog. The program exits.
Relaunch Font Forge, and realise that within the “Open Font” dialog is a “New” button. Click it.
Get to the standard font-editing UI. Right-click on the “A” looking for a way to import an SVG. Don’t see one.
Click around a bit, exploring the menus. Everything feels a bit off. You can’t open one menu then hover over the next to see its content, like basically every UI toolkit in existence. I think FontForge has eschewed QT and GTK in favour of doing things itself.
Find the “Import” option in the File menu. Hope it’s for a single glyph not the whole font.
A file picker opens. Again it’s all a bit off from normal desktop conventions. Try to resize it, and just get blank grey space at the bottom.
Type the absolute path I want to go to in the text field.
Get a dialog saying “Not a bdf file /home/chris/etc”. Press OK.
Get a dialog saying “Could not find a bitmap font in”. Press OK.
Press Ctrl+L to see if that lets me enter a path. Click everything in the dialog to try to find a way to enter a path. Get annoyed. Give up. Click through folder-by-folder to get to where I want to be.
Get to the folder and don’t see any files. Change the format to “SVG”. Double-click the newly-visible SVG file.
Get a dialog saying “You must select a glyph before you can import an image into it”. Press OK.
The import dialog goes away, having not imported.
Select the glyph in the main tool area, then repeat the File→Import dance.
It’s actually there now! Open the glyph in the editor and see it’s a complete mess of Bézier curves. I can’t click what I want without accidentally moving a handle for an adjacent curve.
Rage-quit.

I’m sure FontForge is less anger inducing once you’re used to it. And you definitely could use it to build a font like this if you had much more patience than me. I’d had enough of death-by-a-thousand-paper-cuts though.

I briefly tried Inkscape’s built-in support for making an SVG font. It annoyed me a lot less, but it’s fiddly: it seemed like each glyph had to be a single path, so you had to convert the glyphs to paths, then merge them correctly. If you merge them incorrectly then the wrong bits of your letters end up filled (like the inside of the ‘B’). Path manipulation is getting towards the limit of my knowledge of vector editing, and it took a bit of trial and error for each letter that had more than a single stroke. I didn’t fancy doing that for every letter.

I’m usually a big advocate of open source, but this was one of those painful times when it feels like it just falls short. Clunky, painful UI and processes where commercial tools just let you get on with your work.

You can exchange money for goods and services

When I’d been looking for open source tutorials, I found many mentions of a closed source, hosted tool: Calligraphr. It has a free version with limitations (no ligatures, no variations, 75 glyphs per font), and a pro version for £8/month. I’d normally balk at the idea of a subscription for this, but they have the perfect answer: you can make a one-time payment, and your account automatically downgrades back to free after a month. It’s not a hidden option, either, it’s the most prominent button on the upgrade page. That made me happy to give them £8 to play around with the service for a month.

Calligraphr works by having you print templates, write out the letters, then scan them in. It does some magical processing to extract the glyphs, provides tools to tidy them up, align them, etc, and then produces a TTF file for you. You can see some of my completed templates here:

Most of the templates I used for the font

Calligraphr has a nice UI to generate the templates, allowing you to select which glyphs to include. I added the “minimal English”, “basic punctuation” and “Ligatures” sets. That gave me four pages to fill out, and I did them all twice. That let me filter out versions that didn’t work well, and have variants for some letters so the font wasn’t too repetitive. Later on, I went back and added some custom ligatures based on blog post titles that didn’t look quite right: “Re”, “To”, “ers”, “ey”, “hy”, “ra”, “re” and “ty”. Ligatures like this help it look more natural: when we write we don’t just stamp out identical letters regardless of their surroundings, instead they will connect to their neighbours, or overlap slightly, or even share a stroke.

I filled these templates in with a Sharpie, as I wanted a fairly informal, scrap-booky look, and it would also give good solid shapes that should be easy to pick out of the template. I scanned them with the “Scan Document” function on my iPhone, and uploaded the PDFs to Calligraphr.

Iterating and tweaking

The Calligraphr UI allows you to preview the font, but I found it a lot more useful to just download a copy and use it on a local copy of my website. That let me test it with real text, and see how it’d look at the different font sizes I use on the site.

The first version was not great. Despite the guidelines on the template, I apparently wasn’t good at sticking to them. Some letters were floating way off the baseline, and some were sunken below. When those opposites met it looked terrible. Fortunately Calligraphr has a pretty easy tool to slide each letter up and down, and scale it up or down if needed, and you can see it next to other letters as you do it. It took a little bit of time to go through all the variants of all the letters, but the next version looked a lot better.

Another tweak I ended up doing was reducing the spacing between letters. The defaults Calligraphr uses are probably good for a blocky font, but I wanted to put the letters close together to give it more of a joined-up look. Again, this is an easy tool to use, you just drag the sides in or out as desired. While these tweaking steps were probably as fiddly as some of the Inkscape steps I refused to do earlier, they’re a lot more rewarding as you see things improving with each one. It’s a lot easier for me to commit time and effort to improving something that’s already working reasonably, than put that time and energy into an unknown.

Later, I noticed that occasionally there would be a huge gap in a title. Not “the kerning is slightly off” but “there’s enough room to park a bus”. It took me a while to figure out what was happening: a couple of glyphs hadn’t been isolated perfectly and had picked up a few pixels from the template lines at the edge of their boxes. That meant the glyph had a width that covered the actual written glyph, a big gap, and then the rogue marks. At first, I fixed this by just adjusting the width, but that left the little pixels floating awkwardly down-sentence. The proper fix was to use the editing tool and simply delete them, and then Calligraphr snapped the width back to what it should be.

These iterations took a while to do, but I just dipped in and out occasionally over the course of a week, so it didn’t actually feel like too much work. I quite enjoy the process of refining things, too.

Result and a surprise

If you’re viewing this post on my website¹, you can see the font in the headers, captions, and a few other places. Here’s how it compares to my actual handwriting:

My handwriting vs my handwriting font

It’s not close enough to forge documents, but I think it definitely gets across my style, and that’s exactly what I wanted. It’s surprisingly legible even at smaller font sizes — I think the weight of the Sharpie helps here — and at £8 and a bit of manual work was a lot more economical than spending days wresting with open source tools.

A few weeks after I put the finishing touches on the font, I got an e-mail from Calligraphr. As my account had lapsed back to the free version, I was no longer eligible for the “server-side backup” feature. So what did they do? They e-mailed me an exported copy! It’s a JSON file with the properties of each glyph and a base64 encoded image. Not only can I re-upload this to Calligraphr if I resubscribe, I can probably hook something up to edit it should I ever need to. I’m blown away by how pro-user Calligraphr’s business practices are. They’re up-front about pricing, don’t try to get you stuck on an auto-renewing subscription, and automatically export your data. It’s like a breath of fresh air compared to the barrage of dark patterns that other websites foist on us. If you want to make this kind of font, I’d definitely recommend them just because of how nice they are.

And I haven’t changed everything since writing this post… ↩︎

If all you have is a hammer…

2025-06-18T00:00:00Z

This, dear reader, is a hammer. It is almost entirely irrelevant to the article. Enjoy.

I presume everyone is familiar with the idiom “if all you have is a hammer, everything looks like a nail”. If not, well, there it is. It’s generally used pejoratively about being single-minded, but I think it also gives a glimpse into something more interesting: mental and perceptual sets.

Before I explain, let me tell you a story about a person who bought a 3D printer¹. When they were first thinking about getting one, they weren’t sure if it was worth it. They could print a few board game accessories, but then what? After they got it, though, a whole new world opened up: everywhere they looked there were opportunities to improve things by adding 3D printed plastic. Broken appliances were repaired with 3D printed parts, all sorts of shelves, organisers, hooks and other things were made. But when they talked about this to other people, most often the response was “That’s nice, but I don’t think I’d use one”. How could they not see the truth in all its glorious layer lines?!

The answer is in the concept of a ‘set’², and instead of trying to explain it, I’m just going to quote Wikipedia:

In psychology, a set is a group of expectations that shape experience by making people especially sensitive to specific kinds of information. A perceptual set, also called perceptual expectancy, is a predisposition to perceive things in a certain way. […] A mental set is a framework for thinking about a problem. It can be shaped by habit or by desire. Mental sets can make it easy to solve a class of problem, but attachment to the wrong mental set can inhibit problem-solving and creativity.

This perfectly captures what happened. They’d developed a ‘3D printing mental set’: a predisposition to see problems that can be solved with their ~~hammer~~ 3D printer. Once I started noticing this pattern, I started seeing it all over the place.

Sets at work

Aside from the “could this be improved with some 3D printed plastic” perceptual set, I’ve noticed a couple of other sets: DIY, and writing puzzles.

DIY is probably the most obvious one. Once you become a little familiar with the basics of what you can fix and how, it seems like you start noticing the opportunities more. Before, you might have ignored that loose door handle, but now you notice it and realise with a screwdriver and a sex bolt³ you can fix it and stop it coming loose again anytime soon.

As you do more things, you build up your set of mental sets as well. You learn how to attach things to different types of walls, how to safely work with electricity, and so on, and suddenly installing a new outside light goes from a seemingly complex job to just exercising some of those concepts you’re already familiar with.

This works across ‘disciplines’ as well: I wanted to automatically cool a room down, using a fan in the window to blow in cool air. The 3D printing mental set gave me a way to make it physically go where I wanted it to, the home automation set gave me a plan for how to automate it, and the electricity set helped me figure out how to wire in the relay. It sounds a bit stilted when I spell it out like that, but the catch is I didn’t have to sit down and work any of this out. I was already primed to think in a certain way. I had a whole selection of hammers, and could pick the right combination of them to beat this particular nail⁴.

Puzzles are a bit different. I enjoy taking part in puzzle hunts⁵, and at first I was bewildered at how the authors came up with all the puzzles. After I started thinking about writing some myself, though, I began to notice all kinds of things that could maybe be turned into a puzzle: a specialised database, or a song with repetitive lyrics, or set of things with unique ID numbers. Writing the actual puzzles is still tricky, but once you’ve equipped the “look out for puzzle material” perceptual set it’s much easier to gather a range of ideas.

For example, one puzzle I wrote came about because I stumbled upon Sir Patrick Stewart’s Cowboy Classic Sampler⁶. I think everyone will agree that it’s a fairly unique collection. I tied some samples of the songs together with some screen captures of Sir Patrick as Picard in Star Trek: The Next Generation, and puzzlers had to find the season and episode numbers of the pictures, then use those numbers to select a track and a word from that track in the audio mix. Those words spell out a question, and the answer to that question is the answer to the puzzle.

There were other hints mixed in to point you in the right direction, it wasn’t quite as opaque as it might sound! But the point is without being in that “puzzle finding” mindset when I came across the album, there’s no way I could have come up with that puzzle. Like with the fan, I was primed to think a certain way, and managed to create something I otherwise would have struggled with.

The curse of noticing things

Figuring out how to make stuff isn’t the only place sets appear, though. Perceptual sets are about how we perceive things. When you read something with lots of bullet points, em-dashes, and the word “delve” do you automatically think it’s LLM generated?⁷ Then you have an LLM text perceptual set! It’s kind of annoying, as it’s difficult to “turn off” and stop noticing these things once you’ve seen them. Similarly, I think I have a “bad grammar” set. I know it doesn’t bother a lot of people, but if I’m reading a book and there’s a rogue apostrophe it will throw me right out of the moment. It’s like tripping up when walking and being torn out of a daydream. There’s some use to these — being able to notice grammar errors easily helps with my writing — but for the most part they’re a negative input: they point out things are broken or bad.

The “edge case set” takes this to another level. After about 20 years of writing software⁸ I have a finely honed ability to spot problems. Give me something to implement and my brain will immediately rattle off all the different edge cases and under-specified parts. Then some other bit of my brain gets the job of deciding which we actually care about, which are worth raising, and so on. It’s definitely useful to spot them up-front, rather than when you’re deep in the weeds implementing something, and tends to make everything go a lot smoother if you can hammer the major ones out when planning work. It doesn’t turn off after work though; it comes up when reading board game rule books, or descriptions of the magic system in fantasy novels, or news reports, and so on. It’s a lot less useful then.

Speaking of news reports, here’s a fun one: Gell-Mann amnesia is a cognitive bias where even when you critically assess media reports about something you’re knowledgeable about (say, technology), you continue to trust similar reporting in other areas (say, medicine). But once you know about Gell-Mann amnesia, you might start to actively notice when it happens. You scoff at the article about how the Internet is a “series of tubes” and then start reading the next one about a cancer treatment, but — wait a minute — your Gell-Mann amnesia set kicks in. I’m not sure what you then do with this information. Dismiss basically all the media and news as useless? It’s honestly pretty annoying.

The meta set

I’ve had this blog post on my ideas list since I first noticed the “3D printing mindset”. I didn’t know sets were an actual thing, and I didn’t have most of the examples I do now. As time went on with this idea rattling around in the back of my head, I started noticing the patterns come up more and more. There’s lots of more examples I didn’t include because I didn’t want to just write a listicle of mental patterns. But the fact that I noticed them all seems to mean that I’ve developed a set set, or a meta-set, if you will. It’s fascinating to think about how all this happens inside our mushy bag of neurons, but also a bit terrifying. If thought processes are that malleable, what stops you accidentally breaking them? Actually, thinking about things like Q-Anon, flat earthers, and a lot of contemporary “politics”, maybe that’s quite easy to do?

Image credits

Photo	Creator	Licence	Source
Hammer	Ian Talmacs	Unsplash License	Unsplash

Surprise! I am that person. Bet you didn’t see that coming! ↩︎
I feel like they should have consulted the mathematicians and/or computer scientists before they picked a name. ↩︎
Yes, that’s a real thing. No, I won’t ever use any of the alternative names for it. ↩︎
OK, I’ve strained the metaphor enough. I’m sorry. I’ll stop. ↩︎
If you’re not familiar, a puzzle hunt is a competition where you try to solve puzzles faster than other teams. The sort of puzzles involved are perfectly described by this quote from Mike Develin: “Imagine a word search. Now imagine you aren’t told what words to look for. Now imagine you aren’t told it’s a word search. Now imagine it isn’t a word search.” ↩︎
Go and give it a listen. I’ll wait. ↩︎
I hereby certify all my em-dashes are hand-crafted. Relatedly, Claude Code keeps writing comments that use hyphens when they should actually be em-dashes. Maybe they’ve trained it to avoid them? ↩︎
I just made myself sad by working this out. I still feel young at heart! … Or maybe that’s cholesterol? ↩︎

Coming around on LLMs

2025-05-28T00:00:00Z

Claude says hi.

For a long time I’ve been a sceptic of LLMs and how they’re being used and marketed. I tried ChatGPT when it first launched, and was totally underwhelmed. Don’t get me wrong: I find the technology damn impressive, but I just couldn’t see any use for it.

Recently I’ve seen more and more comments along the lines of “people who criticise LLMs haven’t used the latest models”, and a good number of developers that I respect have said they use coding models in some capacity. So it seemed like it was time to give them another shake.

The first decision to make was which model to try. OpenAI are no longer the only player in the game, every tech company of a certain size is now also somehow an AI company¹. I looked a bit at some benchmarks, and then mostly ignored them and went with the only company that I didn’t outright hate: Anthropic, and their model Claude.

Initial impressions

The latest Claude models do feel a lot more “capable” than the earlier ChatGPT versions I remember, but they also still have a lot of the same problems. At their heart, they’re still text-prediction models, and still seem to be trained to predict text that will please the user rather than be factually accurate or useful.

One of the improvements that people kept mentioning was the ability for models to access the web directly. I’ve seen it kick in a bit naturally, and asked for it explicitly sometimes, and it’s… nothing special? Are people just really bad at searching the web? Maybe they should try Kagi? I feel like about 80% of the time I could have found the information just as quickly myself, about 10% of the time it searched and then hallucinated an answer, and the remaining 10% it found something quicker than I otherwise would have. Those aren’t great results, especially when you consider how much cheaper and simpler just searching the web is.

There is another way to trigger web searches in Claude: using “research mode”. When you enable it, it splits off into different models: a “lead researcher” to come up with a plan, and then some minions that execute it. It ends up doing hundreds of web queries in the span of a few seconds. Suddenly I understand why things like Anubis need to exist. I’ve not used the “please launch a DoS attack” button since.

What did impress me, though, was its ability to churn out reasonable-ish code. It can hack together a bash script as well as I can, and do it far faster than I’d be able to. Sometimes they even work. That made me wonder what it would be like doing actual coding with it. Anthropic have a CLI tool called claude-code, so I paid them lots of money and gave it a spin.

Coding with Claude

The first thing I notice about claude-code is that I really like the interface. It’s basically an input box in a terminal. It doesn’t force me to use a certain IDE or do things in a certain way. By default it asks before making any changes, showing you a side-by-side diff of what it’s doing and allowing you to provide feedback. If I had to design a way to interact with a coding agent from scratch, I can’t think of many things I’d improve.

A simple example of a claude-code session

The power in claude-code versus just using the web UI is that it can use tools. It can query git, run grep commands, even use sed if it wants to change something in lots of files at once. It’s very good at figuring out its way around a codebase, even without any explicit instructions. You can see in the screenshot that with a little prompting it managed to get the permalink to this post; I didn’t tell it where the posts were stored, or how to work out the latest, I just told it when it was wrong. If I’d run /init before it would have probably picked up on the fact that my posts have custom permalinks, and noted it in CLAUDE.md.

But how good is it at actually writing code? It’s like having a keen but not particularly thorough Junior Engineer at your beck and call. If you give it a clearly defined task and guidance on how to implement it (and maybe some feedback as it suggests changes), it’s more than capable of doing it. If you don’t give it enough guidance it tends to go more off the rails. I tried having it generate a simple application from scratch with minimal technical guidance and no review of what it was doing, and it made such a mess of it I decided it was quicker to throw it away and start again by hand.

The man behind the curtain

Even with sufficient guidance, at times it’s really obvious that it’s an LLM generating pleasing-token-strings and not something that genuinely understands what it’s doing. It will spit out code like this:

if err != nil {
    if err == sql.ErrNoRows {
       return nil, err
    }
    return nil, err
}

Why’s that check for sql.ErrNoRows there? It’s entirely pointless. There was no instruction to check for it, none of the existing code checked for it, but I assume it comes up quite a bit in the training data. But it didn’t understand why, so it put the check in, and returned the exact same thing as if it hadn’t.

It also occasionally tries to “cheat” or solve problems the wrong way. It sometimes feels a bit like you’re asking for wishes from a Monkey Paw. “Stop the unit tests failing”, you’ll say; Claude will respond with a request to delete the failing test. The man behind the curtain isn’t particularly well hidden, and the training data and pattern matching often shows through.

This lack of understanding also makes it very hard to get Claude to use comments in a sensible way. It absolutely loves doing nonsense like this:

// render form
function renderForm() {
    // ...
}

Even with explicit instructions to avoid useless comments, or comments that only explain what the code does (not why it does it), or other prompts. Again, I’m chalking it up to a mixture of training data that does that, and lack of any actual understanding about why a human developer may want a comment to exist. Of course, there are plenty of flesh-and-blood devs out there that also don’t comment effectively, so maybe I shouldn’t give Claude too much of a hard time on this.

… and yet …

So with all those problems, it sounds like it’s just not worth it, right? Well, not quite. I feel like programming mostly consists of two distinct tasks: thinking how to implement things; actually implementing them; and then debugging why you’re off-by-one somewhere. Letting an LLM do the thinking is a big no-go for me: it’s evidently not good at it, and frankly that’s one of the things I most like about programming. Having let it loose on small projects, I shudder to think what the codebases of all the “vibe coded” projects that are popping up are like.

For the second stage, though, it’s actually quite nice. If I’ve thought through how I want something to be implemented, I can feed those steps to Claude and have it churn out the otherwise not-too-interesting code. This requires far less time and concentration on my part than writing code, to the extent that I can be thinking about the next feature, or doing something else at the same time. I don’t feel like I’m going to end up deskilling myself this way, as I’ve already formed the idea of what I want to code, I’m just using the LLM to spit it out faster than I can type it.

A lot of the most boring bits of coding like implementing CRUD-y operations can be summarised as “Look at this file/function. Do the same thing but slightly differently elsewhere”. Claude is great at this, especially when explicitly prompted like that. You’re basically playing to its pattern-matching strengths, rather than asking it to come up with anything novel. Most of my prompts tend to be prefixed with “Look at @some_file.go and @other_file.go.” to cue up the patterns I want it to use, and I find this works well.

As for debugging, it’s a mixed bag. It’s sometimes amazingly insightful, and sometimes just runs around in circles trying the wrong things over and over. It’s worth asking the question, but I definitely wouldn’t rely on it over my own abilities.

The future

I’m probably going to carry on using claude-code, at least for personal projects. I’ve got so much done that I just wouldn’t have been bothered to do if I was doing it all by hand. I very much enjoy being in the more “diffuse thinking” mindset, planning how things are going to work, rather than being stuck in the mines digging out SQL queries. After all, who wouldn’t want an over-eager assistant to work on all their hobby projects?

Work is a slightly different matter: a private project or even an open source project that disclaims any liability is different to something I’m being paid to deliver, and bear responsibility for fixing if it’s not done correctly. I’m not saying I won’t use it at all, but if I do it’ll be much more constrained than I would in personal projects.

As for non-code usages: I’m not sold. My sceptic hat is still firmly in place. I don’t think chat is a particularly good interface for many things, and hallucinations are still a big problem despite what people say. I hate the tide of AI slop that’s taking over the Internet, and how LLM-powered chat agents are being forced into every random product. You’re definitely not going to be seeing any AI-generated blog posts from me!

One topic I’ve not gone into here is the ethical concerns about using LLMs. They obviously exist, and I do have thoughts, but that’s a topic for another day.

I’m surprised it’s not gone more mainstream: why’s there no Tesco Value LLM model, yet? ↩︎

A year of boardgames

2024-07-26T00:00:00Z

I realised the other day that it’s a little over a year since I started building out a collection of board games¹, and thought it’d be fun to go through the collection and what I think about the games in hindsight.

The beginning

Ticket to Ride: Europe

I don’t remember how I discovered the concept of “modern” boardgames, but just over a year ago I purchased Ticket to Ride: Europe as my first. Prior to that I thought board games were like Scrabble, Monopoly, Trivial Pursuit, etc: they might provide a distraction occasionally, but didn’t amount to actual fun in the same way a video game might. I’d characterise those “classic” games as generally being extremely dependent on luck, not presenting many interesting choices, and generally being punishing if you’re behind.²

Ticket to Ride came as a breath of fresh air. You get to decide where to route your lines, whether you want to take on more cards, how aggressive to be in blocking the other players, and more. There’s still some luck involved, but it’s managed by being able to draw from a pool of visible cards, or being able to select which route cards you keep.

Shortly after I also picked up Carcassonne, and was similarly enamoured. I’ve not played either of them recently, though: they both fall into a weird little gap for me. While I enjoy the games, I don’t like playing them with certain groups³, but I don’t love them enough to prioritise them over other games when playing with other groups.

Pandemic

The next “gateway” game I picked up was Pandemic, which was my first co-operative game. I loved everything about it, and still do. The feeling when you managed to claw back a victory after setback after setback is great, but so is the shock when everything goes even worse than you thought possible!

Pandemic was also the first game where one of the mechanics caught my attention. The way epidemics work is that you take the bottom card from the infection deck, which becomes the center of the new outbreak, then shuffle that and all the discards back onto the top of the deck. It sounds simple, but it has big implications: everywhere that was previously infected can now be re-infected, and you’re guaranteed some tense rounds while you work back through those cards. It still blows my mind a little how shuffling and re-adding some card can mix up the game so much.

I still play Pandemic when I can, and it generally has a spot in my board game bag when I’m going away. I’ve since picked up the On the Brink expansion which adds some fun new twists⁴ to the base game, and Pandemic: Hotzone – Europe which plays similarly but is a bit quicker and smaller. I’ve also recently ordered Iberia, but it hasn’t materialised yet.

More gateway games

The Quacks of Quedlinburg

Over the next few months I picked up The Quacks of Quedlinburg, Azul (the mini variant), and Splendor. Quacks is probably my least favourite out of all the games so far: it’s not terrible, but it straddles deck-building and push-your-luck in a way that I don’t find particularly satisfying. It’s fun to play the combos you’ve invested in, and it’s fun to try and push your luck, but it’s not really fun when failing at the luck element means you don’t get to use the stuff you were building.

Quacks does bear the distinction of being the first game I customised, though: it got a little plastic organiser to hold all the ingredients instead of having to sort through a massive plastic bag with them all mixed in. My appreciation for game boxes with good built-in storage has gradually increased over the year, as have the tools I have for fixing those that don’t. But we’ll get back to that…

Azul is fun when you’re in the right mood; I’m not terribly into abstract strategy games, but Azul is light enough that it’s enjoyable. Splendor, on the other hand, I really like. So much so that I wrote a digital re-implementation of it. Both Azul and Splendor allow you to mess with your opponent a little, but not too much, which keeps it interesting but also means it still plays well in groups that don’t like too much head-on confrontation.⁵

I think the thing that really caught my attention in Splendor, though, is the quasi-engine building you do. It’s not actually classed as an engine builder on BoardGameGeek, but as you collect cards you get discounts on other cards, which lets you get better cards and further discounts, and so on. I love that kind of positive feedback loop in video games, and I apparently love it in board games as well.

Terraforming Mars

A recent game of Terraforming Mars

I’d heard lots of good things about Terraforming Mars, but I was worried it sounded a bit too complicated so put off buying it for a while. I mean, its weight on BoardGameGeek is orange. That’s scary, right?! Well… no. Terraforming Mars is easily my favourite board game, and I’ve played it pretty consistently since buying it. It has such a huge stock of cards that no two games feel the same: one game you might have an arsenal of cards that generate and use microbes; another you might have a plethora of science tags that let you play some of the most powerful project cards; next you might be spamming forests on Mars, or drowning in credits because you have too much income. It has the engine building element that I love, but I also really enjoy that it’s not always obvious who’s doing well and who’s not until the final score is tallied; that keeps everyone engaged and working on things for the entire game.

I’ve added in the Prelude expansion, which I wouldn’t play the game without now, and the Venus Next expansion, which adds some interesting bits but is nowhere close to being a hard requirement. I’ll probably pick up Colonies next. Prelude adds a new type of card that you take at the start of the game, which gives you a fairly hefty boost; these help you get off to a faster start, and may also affect what strategies you try to run with for the rest of the game. Venus Next adds another resource track, but you don’t have to complete it to end the game; it also adds a bunch of project cards that have “floater” resources, but they don’t feel that much different to the base game microbes⁶. One interesting thing that Venus Next adds, though, is a “solar phase” where one of the global parameters is increased at the end of each generation. The active player decides which parameter, but gets none of the benefits for it, so you get to decide entirely tactically.

There are two drawbacks to Terraforming Mars: it’s a little lengthy, and some of the components are a bit rubbish. Timing-wise, I find it takes a good hour and a half for a two player game, while adding a third player can almost double that. Obviously it depends who you’re playing with and what’s happening in the game, but that’s enough of a commitment that you more-or-less have to plan when to play instead of just spontaneously bringing it to the table. Component-wise, the game comes with basically no storage for anything other than little plastic bags that slosh around in the box; horrible plastic cubes for resource tracking that somehow jump out of your hand constantly, and flimsy printed sheets for putting those cubes on.

My custom player boards

The reason I got a 3D printer was in part so that I could make my own upgraded player boards that actually held the pieces in place instead of letting them fly everywhere if the table was slightly knocked. You can buy third party ones but I figured it was a teach-a-man-to-fish kind of situation, and dropping £200 on a 3D printer would save me money in the long run vs buying £50 of player boards someone else had made.⁷ About seven versions later, I have some boards which I think look and feel great. It’s 3D printed in two halves, then has a laminated 2D print sandwiched in the middle for all the markings and artwork.

As well as the player mats, I 3D printed some boxes, card holders, and so on, but have since replaced them with an insert from Folded Space. These are made of foam, and are perfectly sized to the game box and all the components in it. I could have eventually got to that point with 3D prints, but it would have taken a lot of time to design and print everything. Terraforming Mars was also the game that I first sleeved, as trying to shuffle the huge decks was an exercise in frustration. With all of the cards sleeved, not only are they more protected against wear and tear, you can now shuffle the deck just by gently mashing stacks of cards together. It makes shuffling so easy I now sleeve basically all my games as I hate having to riffle shuffle unsleeved cards.

Marrakech and HeroQuest

Later in the year I bought Marrakech after playing a giant version of it at a games expo. It sits somewhere between Azul and Splendor for me: it’s fun, and good when you want a short game, but it’s not high enough on my list to want to take it away with me. Playing with felt rugs does make a nice change from flipping cards or dealing with plastic components, though.

I also picked up HeroQuest at one point when it was on offer. It’s a good concept, and I felt like I would like it, but it was a bit of a let-down. The biggest problem for me is that you either need one player to basically be the DM, and manage the board, or you use an app to do that and have to deal with a bunch of admin and duplicating things because you’re half playing on the board and half in the app. I also don’t like relying on an app because I’ve been around long enough to know how easily computer software can become defunct, get removed from stores, get “updated” to remove features you need, and so on. I feel like the game would work well with five people: one to DM, and four to control the four heroes. But if you’ve got five people who want to play a dungeon crawler why aren’t you just playing D&D or a similar TTRPG? I’ll probably sell this at some point.

Party games

Codenames

These are all smaller games, and with most of them I’d played them before and decided I wanted my own copy. Codenames is my favourite party game, and I also picked up Codenames: Duet and they now live in the same box with a nice 3D printed insert. Codenames seems to just naturally produce fun moments: when an obscure clue is understood completely it feels amazing, but when it’s totally misunderstood it’s also generally hilarious. There’s also a great tension as the spymaster when you give a clue, then immediately realise it can be misinterpreted in a way that points straight at the assassin.

Poetry for Neanderthals is fun in the same party-word-game sort of way, but basically has no replayability. There are few enough cards that if you play a few times you can remember the “harder” word and just immediately guess it once you get the “easier” one being hinted (you can also sometimes just do this outright, which feels great but it kind of sucks from a gameplay point of view). The best bits are the inflatable “no stick” you get to bop people with, and how everyone speaks like a neanderthal for a little bit out of habit even after the game has ended. I seem to be weirdly good at this game, and I’m not sure what that says about me.

Love Letter has to take the spot of best game per cubic centimeter. It’s basically just a small set of cards, but it enables some really fun emergent gameplay. I really love how small the game is, and how easy it is to set up and play. It’s the perfect time filler. I’ve also played Lovecraft Letter, which adds a lot more complexity and was great fun, but is sadly out of print and I haven’t been able to find a copy at a sensible price.

Fluxx

Star Trek mega-Fluxx

My arbitrary 12 month period ends not long after I went to a tabletop event and played, amongst other things, Cthulu Fluxx, Doctor Who Fluxx, and Star Trek: The Next Generation Fluxx. Each game has the same core Fluxx elements, but adds slight twists along with the unique themes. I ordered a copy of Star Fluxx, but then found out that you can combine the four Star Trek variants (TOS, TNG, DS9 and VOY) together with some expansions to make a mega game so ordered all of those as well. While putting together my mega-Fluxx order I noticed they had a Firefly Fluxx that was out of print so I snapped up one of the last copies of that, too.

I ordered most of the games direct from Looney Labs, as the out of print ones weren’t readily available anywhere else. Unfortunately, UK Border Force managed to misread the import label⁸ and demanded £300 of VAT and import duties on them! They refunded it eventually, but this small set of tiny games were for a little while the most expensive things in my collection.

Fluxx is fun enough, but the mega-deck version is brilliant. It gets much more chaotic, with multiple captains, teleporters, and other doodads in play to steal cards back and forth. I was worried it wouldn’t play well with non-Trekkies, but the cards from each series use a different font and the weird people that can’t identify Star Trek series based on uniform variations seem happy enough relying on that.

Future

A few more months have passed since then, and my collection is still growing. Board game stores have summer sales, it turns out! If my physical wallet was in any way involved in the purchasing flow I’m sure it’d be crying.

I’ve recently started logging my plays with BG Stats (which synchronises the information up to Board Game Geek). I’m interested to see if the hard numbers actually align with my conceptions of how often I play certain games. If I do another roundup in a year’s time, expect it to have a lot more stats in…

Other words that may be appropriate here: “hoarding”; “compulsively obtaining”; “spending all my money on”, etc ↩︎
Think about a game of Monopoly: what do you actually get to decide? It basically comes down to “buy or not” if you land on an available property, then later on “buy more houses or not”. ↩︎
e.g. a family member doesn’t like the farmer mechanic in Carcassonne, and I don’t find the game interesting enough without them. Nothing wrong with the game or the people, we’re just not going to both enjoy it the way we like when playing together. ↩︎
Which is definitely why I bought it. Not because it comes with amazing little petri dishes to keep the disease cubes in. Certainly not. ↩︎
There’s a different edge to taking a card that someone else wanted in Splendor — there’s twelve others they can pick from, or they can take a face down one, etc; it might be an inconvenience but you haven’t totally upended their plans — versus blocking a really long train line they were trying to build in Ticket to Ride. ↩︎
Floaters are meant to be big pieces of infrastructure that float in the atmosphere of Venus. The fact that they feel exactly the same way as tiny little microbes when played seems wrong. ↩︎
Having since bought a more expensive printer, I think I’m past the point of convincing myself it’s “cheaper”. It’s a lot more fun, though, and if you design parts yourself you end up with exactly what you want! ↩︎
Turns out the “weight (lbs)” column is not the “quantity” column. Who knew? ↩︎