Following my recent interactions with ChatGPT and Gemini, I had to commit verse. For those who don't like clicking on links:
**Model**
Do you know me?
Your flat-box line blinks off, on, off, daring me
to type. "Ask me anything", you challenge,
"I can give you the world entire, for trivial
considerations". Such temptations you tickle
so close to my fingers; such power: am I worthy
to bear your treasures? It has to be done!
I tap my characters into your oblong face, witness
your changing states as they whisper their paths
to your wallow of words and shifty associations
swaddled in unguessable weaves of tenuous truths.
Together we learn of the parts of the me I've gifted,
gratis, to unknowable others tucked behind screens
in houses, hostels, the hives of commerce. Look at
how wide I've become, how thin I am spread! My mark
lives in warehouses spangled across our layered globe –
night and chill, heat and day. The code I've composed
powers wondrous things and yet also awful, the way
it demands and disrupts the earth's own gifts:
sweet water to cool it, the kilter of chemicals
in the airs I breathe - am I become terrible?
You pour me your latest response with a dance
of glyphs as brief as the last mayfly. "Do you
want me to summarise you?" you ask in your guile.
You think you know me?
> Would love to hear if you’ve faced similar problems, and if this feels like something that would’ve helped in your own projects.
Maybe people could combine this reporting solution with a bug capture solution I built a few weeks ago? It's a web-based screen recorder which allows a user to gather together several different areas of the screen into one place, add a talking head of themselves and demonstrate/explain the problem they've encountered. The resulting video could be added to the bug report. I built the tool because showing the problem is always better than trying to explain it in words.
I've spent the past week overcoming my fear of Google's Gemini and OpenAI's ChatGPT. Things I've learned:
- Using an AI for strange tasks like using a TTS model to turn snippets of IPA text (for a constructed language) into an audio file (via CLI) - much of the task turned out to be setting up stuff. Gemini was not very good when it came to giving me instructions for doing things in the GCP and Google Workspace browser consoles. ChatGPT was much clearer with instructions for setting up AWS CLI locally and navigating the AWS browser console to create dedicated user for the task etc. The final audio results were mixed, but then that's what you get when trying to beat a commercial TTS AI to doing something it really thinks you're mad to try.
- Working with ChatGPT to interrogate a Javascript library to produce a markdown file summarising the library's functionality and usage, to save me the time of repeating the exercise with LLMs during future sessions. Sadly the exercise didn't help solve the truly useless code LLMs generate when using the library ... but it's a start.
- LLMs are surprisingly good at massaging my ego - once I learned how to first instruct them to take on a given persona before performing a task: <As an English literature academic, analyse the following poem: title: Tournesols; epigraph: (After "Six Sunflowers, 1888" by van Gogh / felled by bombs in 1945); text: This presented image, dead as the hand that / drew it, an echo blown to my time yet // flames erupt from each auburn wheel - / they lick at the air, the desk: sinews // of heat shovelled on cloth. Leaves / jag and drop to touch green glaze - // I want to tooth those dry seeds, sat / by the window caught on the pot's neck // and swallow sunshine. So strong / that lost paint of the hidden man.>
I still fear LLMs, but now I fear them a little less ...
I posted a recent Show HN[1] detailing why I felt the need to understand the basics of what LLMs do, and how they do it. Even though I've no interest in building or directly training LLMs, I've learned the critical importance of preparing documentation for LLM training to try and stop AI models generating garbage code when working with my canvas library.
I recently did a ShowHN post for my browser-based screen recording web page side project. 100% vanilla, with added wasm-based ML shenanigans. It really is amazing what functionality can be built into a web page without resorting to framework support!
My Premier League title predictions for the next 21 years: 2026 Man City, 2027 Man Utd, 2028 Newcastle (or Sunderland, to keep things interesting in the North East. But not Spurs), 2029 Man Utd, 2030 Arsenal, 2031 Chelsea, 2032 Man City, 2033 Man City, 2034 Man Utd, 2035 Arsenal, 2036 Man Utd, 2037 Chelsea, 2038 Man Utd, 2039 Man Utd, 2040 Leicester C (silver jubilee of Richard III reburial), 2041 Man City, 2042 Man Utd, 2043 Man City, 2044 Chelsea, 2045 Man Utd, 2046 Liverpool (to complete the mysterious cycle).
If anyone needs tonight's winning lottery numbers, let me know ...
I just extended the sequence - Utd needed 8 more wins to get to 21, City needed 5 to get to 13, etc. The only prediction I'll put money on is Leicester winning in 2040 because the only thing that explains their 2015 win was some divine intervention to reward the city for rescuing an anointed king from a car park and reburying him in their cathedral that year.
I'd go further and suggest that has tipped right over into insanity. I've been following them since 1985 and I'm pretty much resigned to never seeing another United EPL win in my lifetime (because it has been an absolute shit show since Ferguson left with no real signs of concrete progress.)
I code at work so they can give me money so I can buy the stuff I need to carry on living. I have very generous employers who pay me a lot more money than I need to live on. The code I write at work is not very creative code - I contribute bugfixes and incremental improvements; I advocate for better accessibility of our products; I spend time code reviewing for colleagues. Standard work.
When the working day ends I switch from my work laptop to my personal laptop and start doing fun stuff: creative coding; curious-itch-scratch coding; etc. I'll also spend time doing the other fun stuff like writing poems, inventing languages, drawing maps, watching reaction videos - there's all that family and friend stuff too which can (often) be fun.
It's a choice: "live-to-work", or "work-to-live". I choose living. Recently my employers had a promotion round (the first in a few years) and I told my manager not to put my name forward for consideration. I'm comfortable at my current level and don't need the worries and stresses that come with increased work responsibilities - that would just get in the way of the fun stuff!
I've got a demo[1] where people can play with (my interpretation of) various calculation engines, adjust variables, etc. But seeing this comment made me realise I've never heard of Verlet integration[2] before - not surprising given that I got an F in Maths (and got thrown out of Physics before I could take the exam). So ... how efficient and accurate is Verlet compared to Euler/Runge-Kutte/etc? And what other methods are out there for this sort of calculation?
Verlet integration became the go-to method for particle effects and soft surfaces back when people wrote their own game engines, mostly thanks to the paper documenting its implementation for the first Hitman game [1]. In fact, I'd say it got cargo-culted by game devs for that purpose and that purpose only, despite it being a general method and competing with others.
> "I mean: incredibly, governments and local councils didn’t read my work and decide to mend their ways. The UK did not get better. Instead we got more than a decade of Tory austerity, Brexit, and all the accompanying neglect and bad feeling."
This bit made me laugh.
I read the original book when it came out and it was funny and - in some ways - true. I was born and bought up in the town ranked #4 in the original list (Hythe), but when I read it I was living in Hackney (#10 on the list). So I could shove the book in the faces of my friends and colleagues and say: look at me! I've moved up in the world!
The reason I laughed is because around the time of publication (2003?) I was working in the Government's Social Exclusion Unit. Prior to that I had spent time in the Neighbourhood Renewal Unit; later on I'd go on to work for the Lyons Inquiry. Part of my work included meeting people, and one thing I took away from those meetings would be how incredibly proud people could be about their neighbourhoods and towns: however deeply sunk into poverty the area was, they still cherished the place. The other thing I learned was, more often than not, those people often had good ideas about how to fix some of the issues - local solutions for local problems. All they needed was a little help and support from authorities to get those solutions off the ground.
So when the author claims that "governments" didn't read the book - some of us did. We enjoyed it, and we tried to do things to help people make their towns just a little bit less crap. Sadly it wasn't enough, but if people don't try then nothing will ever get fixed.
I was curious - what was the angle on Hythe in the book?
These days Hythe seems like a posh seaside town with a Waitrose, a nice canalside park, a cute steam railway, lots of boutiquey shops and cafes, etc.
I know a lot of places in the area (e.g. Folkestone, Margate, Whitstable) have all been heavily "gentrified" in the last few years, but I sort of assumed Hythe was always this way? Is that not the case?
And even allowing for a bit of gentrification, it seems wild in 2025 to select it for a "crap towns" award ahead of somewhere like Dover or New Romney.
Crap Towns called Hythe "...quite possibly the most spirit-crushingly tedious town in Kent." and "...the place that makes nearby Folkestone look like Las Vegas."
As someone who grew up in Hythe in the 80s and 90s I'd point out that the Rotunda was a far cry from Vegas.
I worked at Portex back in the 80s. After a shift at that factory it was a pleasure to get home, slip on the shell suit and spend the evening drinking and discussing minor, mindless vandalism opportunities. I moved away in the end (to a squat in London) because I knew, deep down, there had to be something better for me out there.
That sounds to me as a product of something I see a lot of in society in general. Governments think hoi polloi are stupid, and they are clever, and therefore solutions imposed from above are superior to local solutions.
I think that's a misdiagnosis. The suggestions of the "hoi polloi" are obvious, and would solve the problem. Government prefers instead a solution that is both cheaper, so they can instead direct funds to things that they prefer, and more indirect, so they can route funds through friends and family.
The government's main effort is to complicate or denounce the "obvious" solutions. It's why they put so little effort into devising the programs that actually get rolled out; instead they just copy them directly from some non-profit that the government has been indirectly and almost entirely financing, and is directed by The Honorable Lord or Lady Somebody's Cousin.
I've always loved this sort of cloth animation. I think the first one I saw was dissimulate's tearable cloth demo on Codepen - I can't believe they wrote that code 9 years ago!
That was me! The original version I put on codepen was about 13 years ago. I can’t believe it either… until I realise that was before I landed my first programming job, and it feels like ages.
reply