Firehose

There is a common idea that our devices are using their microphones to spy upon us. I’m certain that some devices do, but most don’t.

Under wiretapping and recording consent laws in many jurisdictions, such listening is forbidden. Many platforms, devices, and operating systems also indicate to the user when the camera or microphone is in use. Instead of listening in via a microphone, tracking is handled by more subtle means.

Thinking about travelling to France? While on your phone, you might stop for a moment longer on adverts or posts about travel or about French cuisine. You might linger on a post by a French speaker. Perhaps you research travel plans online on a site that has a partnership with an advertising network or which has social widgets tied to social media. Eventually, you’ll see more and more posts about France and travelling there.

Every little interaction you have across the web – no matter how seemingly inconsequential – can be collated to build a profile on you and target content towards you. It is the entire strategy and purpose of many billion-dollar companies.

Take the story of retailer Target identifying a teen girl as pregnant before even her father knew. There was no recording there; it was careful identification of patterns of behaviour.

There is some great discussion on the topic of devices tracking without listening from Simon Willison in his post, I still don’t think companies serve you ads based on spying through your microphone, and from Jonathan Zeller in McSweeney’s with Calm Down—Your Phone Isn’t Listening to Your Conversations. It’s Just Tracking Everything You Type, Every App You Use, Every Website You Visit, and Everywhere You Go in the Physical World.

Whenever people attempt to test microphone listening, they often end up scrolling around trying to find the thing they were testing for anyway, thus tailoring what they’re served and playing into confirmation bias.

Thinking a microphone is listening to what you say is simple and easy to comprehend. The idea that everything you do is being tracked and analysed by systems so complex you can barely conceptualise them is much more abstract.

I personally find this sort of ever-present and inescapable tracking much more insidious than any microphone listening in.

Personally extremely disappointed to see Mozilla proudly and cheerfully announcing that Firefox is active again on X (formerly Twitter).

Always confused by people/organisations who measure productivity or effort with Git commits.

There are so many different approaches to committing that it seems a completely meaningless metric. Certainly nothing to brag about.

Those tormented screams? Believe it or not, from the damned.

I’ve been asked what framework I used to make the dashboard for my Stoat bot AutoMod so fast and snappy.

It is genuinely just slick HTML/CSS/JS that communicates with an API layer I developed. The dashboard itself doesn’t depend on even a single external library.

The framework-addled mind simply can’t comprehend not reaching for a bloated framework and a handful of libraries for everything.

Squeezing code and prose out of myself like a toothpaste tube on its last legs.

Can’t wait to see AI-driven dementia become a big thing with people offloading all their thinking and their brains thusly deteriorating.

A confession: sometimes I’m too lazy to type </div> so I just copy the closing tag from elsewhere.

One day Google will release the one hundred and eighty second version of Blink, and I’ll be ready with the jokes.

~15 months away if my maths are correct and I’ve got the releases cadences correct. Around October 2027.

I find the period where television was transitioning from 4:3 to 16:9 an interesting one. When broadcasting began in 16:9, but people were still actively using 4:3 televisions, broadcasters would commonly either letterbox the content or expect the left and right edges to be cropped. This led to a period where television series direction avoided putting anything too important at the horizontal edges of the frame. Cinema faced similar challenges, and either embraced pan-and-scan, letterboxing, or careful reframing from a wider image.

Content made exclusively for television broadcast, such as news shows, often had all interface elements centred in the image with nothing of substance in the 16:9 exclusive portions:

You can see this effect in action with The Onion’s morning show, Today Now!. Compare Disney Geneticists Debut New Child Stars, which is in 4:3, to Missing Teen’s Friends Go On TV To Plead For Her Release, Gossip About Ugly Classmates, which is in 16:9 but includes all important information within a 4:3 safe zone.

To mitigate the complexities of working with two aspect ratios, broadcasters (mainly in Europe) opted for an aspect ratio of 14:9 as a compromise. It wasn’t optimal on 16:9 displays or 4:3 displays and was pillarboxed or letterboxed, respectively.

The cropping of a wider image reminds me of open mattes, which is where film is captured at a larger size and then matted down to a specific aspect ratio, such as by applying bars on the tops (letterboxing) or sides (pillarboxing). Take the 35mm open matte of Spider-Man (2002). Extra content is there; it is just hidden by the aspect ratio. It is by this method that more full frames can be achieved by merging various source mediums.

The embrace of 16:9 from 4:3 is also tied up in the switch from standard definition to high definition and analogue to digital, meaning that content is large on the screen – especially text, which can be blurred by low-resolution displays and fuzzy CRTs. It is an interesting period of transition, and as a front-end developer I can’t help but think of it in terms of progressive enhancement.

For people without widescreen displays, content functioned and continued as normal without interruption. For people with capable television sets, however, they got an improved experience. As time progressed and 4:3 became less common, it switched from progressive enhancement to graceful degradation on displays still using the old aspect ratio, as we often see as newer technology reaches wider adoption.

Currently voting on conference talks based entirely on which talk is less about AI.

I know that the conference I’m voting for had guidelines when submitting papers, so I’m also voting against the talks which blatantly ignored these guidelines and are full of AI tells.

Colon split title, verbose, too long, tricolons, not X but Y, and an unreasonable number of em dashes.

Why submit a talk if you can’t be bothered to even think up or describe the premise of it yourself?

The use of ‘he/she’ really annoys me.

‘They’ is much more inclusive than ‘he/she’ and is also quicker to both say and write. Anyone using ‘he/she’ today is either forced to by some obtuse requirements, or is purposefully being bigoted.

‘(s)he’ is no better and is just as needlessly difficult. Also, it is always ‘he/she’, never ‘she/he’. ‘(s)he’ literally demotes the feminine pronoun with parenthesis.

Don’t get me started on cases where an entire document is written with masculine pronouns as ‘default’.

I just saw a marketing page refer to a product as the ‘final solution’.

I didn’t think it needed saying, but don’t go around advertising the ‘final solution’. There are some connotations there…

Information architecture? More like, ‘Blimey, I’m 90% through organising this crap, and now I’ve just found a stupid edge case that means I’ve got to reorganise the entire hierarchy and structure.’

The sudden sinking feeling when you’re given a codebase to work on, then you notice the HTML tags are all written in uppercase, that there is no document declaration, and that there are conditional comments for Internet Explorer.

‘Hi. Yeah, it’s me. I couldn’t help but notice when I opened the folder you sent me that boss music started playing, and in the distance I heard what I think might be the eternal screams of the damned. Can you just confirm this is expected behaviour?’

With Xbox struggling and the PlayStation physical media fiasco, Sega has the chance to do the funniest thing.

I’m not quite sure where to go or what to do. I’m watching my peers slowly burn out, go quiet, quit the industry, or struggle on hoping it’ll get better in time. Even if the storm is weathered, what comes out the other side?

Businesses seem wary of investing in people or just spending money at all. I feel I have to sell the very basic concept of doing ‘good work’, rather than ‘good enough work’. Accessibility feels like something I discuss but don’t get to see through.

Things are stagnating for genuine professionals.

Many conferences are coming to their end, seeing lower attendance, and are dominated by AI talks. Of those that are surviving, many are cutting budgets or catering to smaller, focussed audiences. Most publications are seeing declining readership, and many are turning to AI for their content.

This is a clear result of education by humans being in a struggling place, with people turning to generative AI for quick solutions (and notably not to comprehensively learn). Speaking to many folks with courses, their sales have plummeted – dropped completely to nil in a disturbing number of cases.

It is gutting to feel I’d see greater success if I didn’t care who I worked for, gave up on my morals, and embraced the latest harmful fad. Grifting has long been a thing, but at present it seems there aren’t any consequences beyond one’s own conscience and that it is the optimal way to make a buck.

Pair all those issues with the dire state of everything. Computers and their components’ costs continue to inflate, the USA has a fascist government, there are major ongoing wars (and new ones being started), and everything is increasingly morally ambiguous – especially in tech.

Couple all that with the erosion of rights and companies being increasingly anti-consumer. Forcing features unwanted, removing features wanted, taking away physical media, making all a subscription, taking away our control of our devices and what software we can run on them, and eroding privacy.

Things are dire, and I’m thinking I need to pivot, because front-end developer is a job that companies are happy to have AI complete, regardless of if it is on par with a professional’s work. My work is better than sloppy AI output, but that can be hard to distinguish if one doesn’t investigate.

Perhaps a freelancer collective is needed? A distributed agency. Different people with different specialisations sharing resources and referring work to each other based on their capabilities. Every freelancer I know is reporting difficulty finding work, and they’re long established in the industry.

I don’t know how to feel or what to do. Things are bad, and while I’m sure they’ll get better to a degree eventually, I’m not sure how much better they’ll get. What does ‘better’ look like when the greatest minds in the industry have left and don’t come back? So much lost skill and knowledge.

I suppose Google Search’s full pivot to AI explains why Google dropped support for non-JavaScript usage at the end of 2025.

They should do a show like Air Crash Investigations, but instead of investigating aeroplanes, they investigate what went wrong at last Tuesday’s all-hands.

AutoMod used to have a red accent, but the logo was rainbow-ified for pride month a few years ago and it stuck.

Sometimes people are critical of it which weeds them out, and sometimes people message me because they appreciate the representation and it makes them feel safe.

Another pride month has come and gone, but AutoMod will remain as colourful as ever. True pride is deeper than recolouring a logo for a month.

Reddit is slowly winding down Old Reddit. Functionality is being removed and features are discreetly disappearing. They’re quite obviously eroding its functionality so that when they finally kill it there is less outrage.

Sucks as someone who makes money from Old Reddit stylesheets…

https://arstechnica.com/gadgets/2026/06/reddit-will-require-you-to-log-in-to-use-old-reddit-com/

CSS started out as something a lot simpler. Originally it was just a bare-bones means to lightly simple, majority text pages, but it quickly took off and had to keep up with the swiftly evolving web.

Along the way, mistakes were made – both genuine blunders and at the time reasonable choices which time hasn’t been kind to and which are flawed in retrospect. The CSS Working Group published the Incomplete List of Mistakes in the Design of CSS, which outlines a great many of these suboptimal features. A similar story is etched into the history of many languages; they started simple with a tight scope, but decades of usage have expanded the cases beyond the original design decisions.

These mistakes obviously cannot just be fixed, for the plethora of existing software relying on them would break, and even if fallbacks were implemented, there are edge cases galore which could cause unacceptable disruption. Thus, to peer into the alternate timeline where CSS didn’t make these mistakes, I’ve created a library: FixCSS.

FixCSS tackles the mistakes and the knock-on implications of them to present a CSS where they never happened – a CSS that could have been. It works client-side, as a Node package, and via the command line (CLI). It is mostly a novelty, rather than something for production. We must live with the stylesheets we’ve cascaded ourselves into.

The source code is on Tangled: https://tangled.org/vale.rocks/FixCSS
The library can be found on NPM: https://npmx.dev/package/@declanchidlow/fixcss

There is talk of regulating smart glasses, perhaps outright banning them. I can agree with regulation, but not with a ban.

For a lot of folks with severe visual impairments, smart glasses paired with image recognition are a great aid. For general usage by other people, a convenient glasses-based camera is great for capturing the exact perspective of the wearer while avoiding bulky head mounted camera gear.

However, what isn’t right is creepy, discreet, non-consensual recording, especially given who the recordings are in the hands of. Social media companies are not good custodians of data, Meta especially not. They should not share people’s eyes into the world. It is true that in many places you are legally able to record in public, but the legal fact doesn’t change the ethical and moral correctness. Flock has surveillance cameras across the United States for tracking people’s every move, and while legal, I think them reprehensible. Same applies to privacy-eroding face based cameras.

Smart glasses should use local or data-protections vetted object recognition models, be disconnected from surveillance operations and big tech, and have large, bright recording lights that are directly integrated with the camera, such that the camera cannot operate in isolation. It should also be made clear they are camera equipped, perhaps with an overt reflective patch as part of their design.

I don’t think a ban it a good idea, as bans are very rarely good ideas, but I do think that regulation is necessary and prudent. I fear that restricting their availability so they can only be used as an accessibility aid would end commercial investment into them entirely, but then removing the data-collecting ‘perv glasses’ aspects would also dissuade companies from further developing them.

I don’t actually understand why everyone keeps talking about using AI for coding.

I simply can’t see how artificial insemination helps with building software.

I play a fun game when checking domains that have expired called ‘Is it now a porn site, a gambling site, or did they fail a pivot into crypto/NFTs?’

Some applications and websites have inbuilt theming capabilities. Others lack anything of the sort but can be themed by external means.

I’d prefer people don’t theme or alter my work, as a painter prefers people not to alter their paintings. Please don’t theme our apps pops to mind. Of course, the web is a medium in which some modification can be expected. From zooming, to forcing fonts, to high-contrast mode, etc, people will modify sites. Particularly on the web, people will theme and modify things as they see fit. It is very much part of the platform’s ethos – user agent stylesheets and extensions and whatnot.

I know that I personally alter websites all the time. Whether that be deleting an element, forcing dark mode, stopping an annoying visual effect, or something more substantial. I also know that people have altered websites I’ve worked on (and icons I’ve designed), and that doing so usually makes me feel something adjacent to bad, conjuring thoughts of ‘Have I failed, such that people must undo or change my work?’.

Another line of thought in my own theming of sites (and indeed, creation of my own bespoke apps) is that I’m being exposed to less design. Exposure to different interfaces and flourishes is inspiring. You find new interface patterns and stay up-to-date with trends by seeing what is popular and in use. As a developer and designer, keeping track with what is ‘in’ is important, as is finding inspiration to motivate oneself and apply to other contexts. By forcing everything existing into my own rigid pre-existing structures, I’m walling myself off. In some cases, this is necessary, such as when the design is really bad or user-hostile, but in many other cases it is something I’m doing out of pure comfort with the known.