Hidde's links

Filler text no one wants to read or write

2023-07-04T00:00:00Z

Sci-fi writer Ted Chiang: ‘The machines we have now are not conscious’:

Chiang’s view is that large language models (or LLMs), the technology underlying chatbots such as ChatGPT and Google’s Bard, are useful mostly for producing filler text that no one necessarily wants to read or write

Originally posted on 4 July 2023 on Hidde's blog.

Reply via email

Algorithmic thatcherism

2024-01-03T00:00:00Z

Dan McQuillan says AI is algorithmic Thatcherism:

“Case after case, from Australia to the Netherlands, has proven that unleashing machine learning in welfare systems amplifies injustice and the punishment of the poor. AI doesn't provide insights as it's just a giant statistical guessing game. What it does do is amplify thoughtlessness, a lack of care, and a distancing from actual consequences.”

(via Ethan Marcotte)

Originally posted on 3 January 2024 on Hidde's blog.

Reply via email

Encapsulating components

2024-01-04T00:00:00Z

Nolan Lawson on the problem of component encapsulation:

Overall, what I would love to see is a thorough synopsis of the various groups involved in the web component ecosystem, how the existing solutions have worked in practice, what’s been tried and what hasn’t, and what needs to change to move forward.

Originally posted on 4 January 2024 on Hidde's blog.

Reply via email

AI terminology

2024-01-07T00:00:00Z

I've been using “AI”, with quotes, a bunch on this website. I feel the industry is calling things artificially intelligent way beyond the scope of what that (admittedly hard to define) phrase actually means. That makes criticial analysis harder… that's good for marketeers, not so much for others.

Simon Willison agrees that “spicy autocomplete” is a good analogy for how LLMs work today, but at the same, it's ok to call it artificial intelligence:

We need an agreed term for this class of technology, in order to have conversations about it. I think it’s time to accept that “AI” is good enough, and is already widely understood.

Originally posted on 7 January 2024 on Hidde's blog.

Reply via email

How engineers see the web

2024-01-07T00:00:00Z

In Weird things engineers believe about Web development, Brian Birtles talks about different assumptions of developers of websites and and web browsers:

it’s easy to assume our experience of the Web is representative of Web development in general

Yup, checks out.

Originally posted on 7 January 2024 on Hidde's blog.

Reply via email

Losing the imitation game

2024-01-07T00:00:00Z

Jennifer Moore on what LLMs can and cannot do:

The fundamental task of software development is not writing out the syntax that will execute a program. The task is to build a mental model of that complex system, make sense of it, and manage it over time.

From: Losing the imitation game

Originally posted on 7 January 2024 on Hidde's blog.

Reply via email

Design systems and the promise of solving inconsistency

2024-01-10T00:00:00Z

Design systems aren't a silver bullet to align teams and magically make digital products more consistent. In Through a design system, darkly, Ethan Marcotte describes two issues:

Design systems haven’t “solved” inconsistency. Rather, they’ve shifted how and when it manifests.

Many design systems have introduced another, deeper issue: a problem of visibility.

Originally posted on 10 January 2024 on Hidde's blog.

Reply via email

Inspecting a scam site

2024-01-11T00:00:00Z

Hui Jing had fun inspecting a site:

I just got a scam SMS and thought it’d be fun to inspect how the phishing website works

Hope you recover soon, my friend!

(From: Let's inspect a phishing site)

Originally posted on 11 January 2024 on Hidde's blog.

Reply via email

More unnecessary AI

2024-01-11T00:00:00Z

One of the things that I keep circling back to when reading about ‘AI’ is the kind of problems people are trying to solve with it, so many of which are completely futile.

Chris Person on the ‘rabbit’:

What’s most annoying about all of this is the sheer repeated imposition of this horseshit. I’m sick of being forced to think about generative AI, large language models and large action models. I’m tired of these adult toddlers who need an AI to tie their shoes and make bad Pixar characters for them. Microsoft and Google keep shoving AI features into their software, and I absolutely should not have to worry about this garbage from Firefox of all places.

(from: Why Would I Buy This Useless, Evil Thing? - Aftermath)

Originally posted on 11 January 2024 on Hidde's blog.

Reply via email

WCAG 2.2 in GOV.UK design system

2024-01-11T00:00:00Z

Design systems can be a super effective way to propagate a lot of accessibility at once, across many services. Not just as part of components that have good defaults, but also, maybe especially, in written documentation that helps people understand better what to do.

The GOV.UK Design System was updated to add an accessibility section and specific guidance on meeting WCAG 2.2 across the different component pages.

Originally posted on 11 January 2024 on Hidde's blog.

Reply via email

Meta's Fediverse plans

2024-01-12T00:00:00Z

Tom Coates went to a Meta event about Threads and the Fediverse:

Just before Christmas I was lucky enough to be invited to a Data Dialogue event at Meta’s offices in San Francisco. The event was designed to reach out to people in the ‘Fediverse’ community, tell us their plans for their product “Threads” and get a bit of feedback about the policy and privacy implications.

His post has a detailed analysis of the pros and cons: How Threads will integrate with the Fediverse – plasticbag.org

Originally posted on 12 January 2024 on Hidde's blog.

Reply via email

Human curation is best

2024-01-14T00:00:00Z

Cassidy Williams:

In the earlier internet days, you went to a fun website or read the latest thing because you decided to go do it. Now, all of this content is pushed in your face, designed to be as addicting as possible, so you keep coming back.

This is from her recent post in which she explains why she misses human curation, it ties in nicely with why I got into reading (and, as of recently, ow writing) link blogs.

Cassidy is right about algorithmic curation, I feel the same when scrolling modern social media. The sites found out I liked a couple of things, then bombard me with just more of those things. It's better to optimise for human context and intention than for algorithms, they quickly get boring.

Originally posted on 14 January 2024 on Hidde's blog.

Reply via email

A system of common components

2024-01-15T00:00:00Z

In A Global Design System, Brad Frost makes a case to:

centralize common UI components, reduce so much of this unnecessary duplication, integrate with any web-based tech stack, and create a connected vehicle for delivering front-end best practices to the world’s web experiences.

His proposal explicitly isn't “new HTML features”. We're working on that over at Open UI, and it's fruitful and good, but also hard, because web compatibility is complicated. It's challenging to get accessibility “built-in” in: these web platform features, like a fully customisable select, need to be flexible and styleable enough so that people actually want to use them, but also inflexible enough so that people can't accidentally use them to make something inaccessible. Or maybe “inflexible” isn't the right word… it's a matter of adding “guardrails” to these features: what sort of ARIA relationships and states should apply when? Browsers can't guarantee what developers are trying or going to do.

Anyway, Brad's “global design system“ is not that, it's proposed as a layer on top of HTML, a common library between design systems. That too resonates with me. In fact, it is close to what NL Design System is setting out to do. Different government departments and layers maintain their own design systems, but share a common architecture and reuse themeable components and tests from one another.

Originally posted on 15 January 2024 on Hidde's blog.

Reply via email

Books with music

2024-01-16T00:00:00Z

Melanie made a new website called Lit Tapes:

I read a couple books in the past few months that heavily featured music throughout the plot, and thought it would be fun to create playlists for those books.

(From: New project: Lit Tapes | Melanie Richards)

It looks gorgeous.

Books that feature a lot of music are my jam, so I love this idea. I'm currently reading Whites can dance too by Kalaf Epalanga, but in this genre I also liked Mix tape by Jane Sanderson and everything by Haruki Murakami.

Originally posted on 16 January 2024 on Hidde's blog.

Reply via email

AI images look cheap and easy

2024-01-17T00:00:00Z

iA write excellent posts that put “AI” into context. In their latest, they compare these images, that ‘often miss realness, depth, and originality’, to stock photos. This comes with a business risk: your content looks cheaper, of less value:

using AI images makes all of your content feel ordinary. Good images enrich your article, bad images devalue it. Your audience thinks: “If they use AI for images, they probably use it for content, too.”

(From: AI Art is The New Stock Image)

Unless there's a load of generated images that are so good that we can't recognise them, and we don't realise, I think iA are right: they're super obvious to spot and already look old.

Further down in the post, they predict the lack of creativity in machines may spark more human creativity:

Photography has made us question traditional art. Similarly, AI can make us question empty off-the-shelf communication. Ironically, machine-generated content might catalyze a fresh wave of humane creativity and hand-crafted innovation in verbal and visual storytelling.

I sure hope so. If we are to create things worth having around, we've got to make our choices and intentions matter.

Originally posted on 17 January 2024 on Hidde's blog.

Reply via email

Billions for “AGI” and “metaverse”

2024-01-18T00:00:00Z

“Artificial general intelligence” is a phrase different people assign different meanings to. Few think is actually within the realm of possibility. Yet, Zuckerberg talked to The Verge to announce Meta's new focus on trying to find out:

While he doesn’t have (…) an exact definition for it, he wants to build it.

(From: Mark Zuckerberg’s new goal is creating artificial general intelligence at The Verge)

In the same interview he also wanted to “unequivocally state” they're still focused on “the metaverse” and will spend more than 15 billion dollars per year on that. Imagine that sort of budget to go to solving some of the world's more clearly defined problems.

Originally posted on 18 January 2024 on Hidde's blog.

Reply via email

Standards in 2024

2024-01-26T00:00:00Z

In a blog post, W3C's new CEO Seth Dobbs outlines a focus on putting people first. He explains global standards, like the W3C's, are essential to:

[ensure technologies] are accessible by all, secure, maintain privacy, respect the planet, and work anywhere in the world

Originally posted on 26 January 2024 on Hidde's blog.

Reply via email

Work and life

2024-01-26T00:00:00Z

Mandy Brown suggests we should give less fucks about work:

Because here’s what I’ve learned: if you give your fucks to the unliving—if you plant those fucks in institutions or systems or platforms or, gods forbid, interest rates—you will run out of fucks. One day you will reach into that bag and your hand will meet nothing but air and you will be bereft.

(From: A unified theory of fucks)

Originally posted on 26 January 2024 on Hidde's blog.

Reply via email

Costs of running Signal

2024-01-28T00:00:00Z

Signal shares what it costs to make Signal:

We estimate that by 2025, Signal will require approximately $50 million dollars a year to operate (…)

Here we review some of these costs and where this money goes, [and] help clarify just what is required to fulfill the dream of privacy-preserving alternative technology, and contribute to establishing a solid foundation from which we can grow alternatives that contest tech surveillance and the incentives behind it.

(From: Privacy is Priceless, but Signal is Expensive on fhe Signal blog)

They do all this with donations, not investors.

Originally posted on 28 January 2024 on Hidde's blog.

Reply via email

Custom elements in React

2024-01-28T00:00:00Z

It took a while, but it's happening!

React 19 […] will have direct support for custom elements. Developers can expect that all tests on custom-elements-everywhere.com will pass by default going forward, like they currently do in the experimental channels. Release date, as well as docs for what's supported, still to be announced.

(From: RFC: Plan for custom element attributes/properties in React 19 · Issue #11347 · facebook/react)

Originally posted on 28 January 2024 on Hidde's blog.

Reply via email

Indie web and IndieWeb

2024-01-28T00:00:00Z

The web is yours. You can put up a website where you share whatever it is that you want to share with others.

(From: The indie web | James' Coffee Blog)

Originally posted on 28 January 2024 on Hidde's blog.

Reply via email

The LLM search engine

2024-02-01T00:00:00Z

Ben Werdmuller tried Arc's new “AI”-based search and shares his concerns in Stripping the web of its humanity.

Like all these tools, it outputs falsehoods. But that isn't the worst issue, he explains. Without attribution, the tool gives a false sense of objectivity and hides away bias:

If I search for “who should I follow in AI?” I get the usual AI influencers, with no mention of Timnit Gebru or Joy Buolamwini (who would be my first choices). If I ask who to follow in tech, I get Elon Musk. It undoubtedly has a lens through which it sees the world.

It's a particular kind of bubble where Elon Musk is worth following and Timnit Gebru is not suggested (would very much recommend following her instead).

Ben also notes that when bots consume content instead of humans, that threatens the ecosystem of content and writing:

If we strip [payments or donations to writers] away, there’s no writing, information, or expression for the app to summarize.

Who's going to make the input these tools grab in order to generate their output? Google faced various legal issues around displaying excerpts of news outlets on their news website. But they did at least quote and attribute them, while linking to the original. The automated processing basically strips away any opportunity for writers to be paid (or known) for their work.

Originally posted on 1 February 2024 on Hidde's blog.

Reply via email

Bazaar vs cathedral people management

2024-02-06T00:00:00Z

As a manager, you should ask yourself: are you building a cathedral or a bazaar? And as an employee, you should ask yourself: do you prefer working in a cathedral or a bazaar?

In Cathedral vs Bazaar People Management, Ben Balter explains how these two styles of software development are indeed also styles of people management. He also explains how they can differ by industry (eg army needs more hierarchy than startup), individual (junior might need more structure than senior) and role (of course, it depends on what you do).

Originally posted on 6 February 2024 on Hidde's blog.

Reply via email

The tech a VC does believe in

2024-02-08T00:00:00Z

Molly White wrote an excellent review of a book written by Chris Dixon, an investor specialised in crypto companies and blockchain at Andreessen Horowitz. Like Molly, I’m very scepticical of crypto companies. While I understand people want to try and get rich from tech investments, and that creating hype could help with such goals, I continue to not get the premise of this particular hype.

We’ve now had many years of solid crypto and blockchain criticism. From Molly’s review I gather Dixon’s book doesn’t really engage with that criticism, at all. He just dismisses it. While bashing on other technologies, like RSS. Strikes me as ironic that Dixon can’t see the use cases for RSS, while advocating for blockchaim, a technology still in search for a compelling application. Is it that RSS isn’t ‘monetised’, and that the ’value‘ he’s looking for is ’value as in money‘? Rather than ‘is how podcasts work and lots of people use this to keep up to date with content’?

Or, as Ahmet A. Sabanc wrote, in RSS is not dead:

RSS was never dead. I don’t know why people claimed that just because a tech overlord decided something is not profitable for them. People are still using it and it’s as good and alive as it’s always been.

Originally posted on 8 February 2024 on Hidde's blog.

Reply via email

Apple removes PWA support

2024-02-09T00:00:00Z

Bruce on how Apple now seems to plan for breaking PWAs:

Presumably Apple doesn’t want PWAs to open in third-party browsers that have more powerful features than Safari, because those would directly compete with native apps in its own App Store. However, in the EU, it can’t privilege PWAs in Safari with its own private APIs any more. And so its solution, in its spirit of malicious compliance, seems to be “if we can’t have them, nobody can!”

(From: Bruce Lawson's personal site : Is Apple breaking PWAs out of malicious compliance?)

I loved to see Apple invest in web platform features and Safari a lot more in recent years. I hate to see the way they seem to want to place themselves above (what seems to
me is fair) European law.

Originally posted on 9 February 2024 on Hidde's blog.

Reply via email

Opportunities for AI in accessibility

2024-02-09T00:00:00Z

Aaron Gustafson:

AI can be used in very constructive, inclusive, and accessible ways; and it can also be used in destructive, exclusive, and harmful ones. And there are a ton of uses somewhere in the mediocre middle as well.

(From: Opportunities for AI in Accessibility – A List Apart)

In this post, Aaron shares some examples of where ‘AI’ could be used to make content more broadly accessible. This is a controversial subject, because there are many automated ‘solutions’ that don't actually remove barriers, so caution is warranted. Such solutions often focus on people who want to comply with accessibility instead of people with disabilities. And accessibility is about people with disabilities, period. Aaron acknowledges this in the post, and calls for including people with disabilities.

What if, he suggests, users could ask specific questions about complex charts? As Aaron acknowledges, hallucinations exist, there could still be a use, especially with more diverse training data. Other examples of where ‘AI’ could remove barriers in his post: voice preservation, voice recognition and text transformation.

I'm still sceptical, because I've seen too many claims from automated tools that don't work that well in practice, but I understand it's worth to at least explore different options, and weigh them against the reality of today's web. For the voice and text tools I am actually somewhat optimistic.

Originally posted on 9 February 2024 on Hidde's blog.

Reply via email

Questioning practices for a more ethical web

2024-02-09T00:00:00Z

Every day, I get to work with people who do good work, who care about things like privacy, ethics and accessibility. We exist. We just are drowning, drowning in algorithmically-guarded walled gardens that are nothing more than quagmires of enshittification, of AI-generated content, of snake-oil solutions.

(From: Have we forgotten how to build ethical things for the web? - Nic Chan)

This piece from Nic resonated. There's a lot of good people in the web industry, doing good things. But there's also a lot of very bad practices that have an impact way beyond their initial uses. From content optimised for search engines (and thus not for humans) to extreme tracking (very far from just figuring out how many people visit which pages).

To avoid bad practices and build more ethically as an industry, I think what Nic shares would actually be the most effective way: to actively question practices and orders we receive as individuals in teams. In What kind of ethics do front-end developers need? I listed a bunch of other things that individuals in teams can make a fuss about.

Originally posted on 9 February 2024 on Hidde's blog.

Reply via email

Don't disable form fields

2024-02-12T00:00:00Z

In Don’t Disable Form Controls, Adrian Roselli explains that, while it sometimes can be ok to disable buttons, it's never ok to disable submit buttons (or any form fields):

Telling authors not to disable submit buttons is too narrow. Authors should not disable any form fields.

Originally posted on 12 February 2024 on Hidde's blog.

Reply via email

Django's accessibility improvements

2024-02-14T00:00:00Z

In Django accessibility in 2023 and beyond, the accessibility team of Python framework Django reflects on their work in the year 2023 and looks to the future. This makes me happy to see, as I'm a bit of a CMS/built-in accessibility nerd (see this intro to CMS accessibility with ATAG and Your CMS is an accessibility assistant). CMS and framework accessibility projects have the potential to increase a lot of accessibility at once: better defaults, better guidance and better output can all lead to a ripple effect. Better admin interfaces can increase how many people can make content for the web, which, again, is great.

Django updated forms in their core, improved the admin UI in lots of ways, updated guidance and automated tooling (including CI/CD). On top, they improved their website and did a bunch of outreach. Great to see it, may more CMSes follow suit (shoutout to WordPress Accessibility Day and Drupal's Accessibility Project).

Originally posted on 14 February 2024 on Hidde's blog.

Reply via email

Webmentions and Sundays

2024-02-14T00:00:00Z

Wouter Groeneveld ran a Webmentions server then decided to stop running it, because it costs him too much time:

I’d rather spend my Sunday doing something else.

(From: Why I Retired My Webmention Server | Brain Baking)

He questions if it's worth having mentions at all if most are likes and most are sent from Bridgy, a centralised service.

Food for thought. I think I like even the mentions that are just likes… not as some vanity metric, but because it adds a layer of humanity and, community, at least in my bubble. But it is at the expense of privacy. And simplicity (shoutout to Mu-An and Robb's OpenHeart, which is indeed “much, much, much simpler”). And efficiency, especially if you recheck for deletion of replies, as Robb concludes you probably should (not the same Robb).

Originally posted on 14 February 2024 on Hidde's blog.

Reply via email

Software quality and developer leverage

2024-02-17T00:00:00Z

Isaac Lyman (Stack Overflow) looks at whether software gets worse. He notices software, unlike sandwiches and movie tickets, is hard to sell:

There are only two demographics that are willing to pay for good software: corporations and video gamers. We’ve somehow blundered our way into a world where everyone else expects software to be free.

(From: Is software getting worse? - Stack Overflow)

Developers have leverage, the post concludes, they have the power to insist on quality, but aren't really using it. This seems unlikely to change overnight, says Isaac.

Originally posted on 17 February 2024 on Hidde's blog.

Reply via email

Paying people to work on open source is good

2024-02-18T00:00:00Z

paying people to work on open source is good, full stop, no exceptions. We need to stop criticizing maintainers getting paid, and start celebrating. Yes, all of the mechanisms are flawed in some way, but that’s because the world is flawed, and it’s not the fault of the people taking money.

(From: Paying people to work on open source is good actually - Jacob Kaplan-Moss)

Originally posted on 18 February 2024 on Hidde's blog.

Reply via email

New cypher

2024-02-19T00:00:00Z

TIL that UK monarchs choose their own Royal Cypher, a “symbol to represent their personal authority”.

In September 2022, the College of Arms announced His Majesty King Charles III’s Royal Cypher, which features the monarch’s chosen crown. This Cypher features the Tudor Crown, rather than the St Edward’s Crown chosen by Queen Elizabeth II following her Accession in 1952. Her Royal Cypher was itself a change from her father King George VI.

On each accession, the monarch will choose a Royal Cypher, or symbol to represent their personal authority. You can see the Royal Cypher in many places, for example post boxes, on police and military uniforms or on the side of official buildings.

(From: Updating GOV.UK’s crown - Inside GOV.UK)

With a government wide design system, the digital version of that can be rolled out much more efficiently. Yay design systems!

Originally posted on 19 February 2024 on Hidde's blog.

Reply via email

Stitching together

2024-02-19T00:00:00Z

Brian Merchant explains in Let's not do this again, please that OpenAI's new image generating thingy is mostly a “promotial narrative” to try and seek more investment money (OpenAI's server spend, the article says, is over 1 million USD per day).

The tech stitches together imagery, rather than create new imagery, Brian says:

It’s not that Sora is generating new and amazing scenes based on the words you’re typing — it’s automating the act of stitching together renderings of extant video and images.

Originally posted on 19 February 2024 on Hidde's blog.

Reply via email

Design that encourages deletion

2024-02-21T00:00:00Z

In Design 11Patterns that Encourage Junk Data, Michelle talks about the environmental cost of creating and storing so much of our data in the ‘cloud’:

the need for limitless digital storage bumps up against the very real physical limits of our planet.

(From: CSS { In Real Life } | Design 11Patterns that Encourage Junk Data)

She explains it's not only a huge amount of data, a lot of it is probably unnecessary:

It’s estimated that up to 88% of the data stored in the cloud is ROT (Redundant, Obsolete or Trivial) data, or “dark data”: data collected by companies in the course of their regular business activities, but which is not used for any other purpose. It all amounts to a lot of junk data that has no purpose, that will never be needed or looked at again.

Yup, I definitely store a lot of photos and emails that I will never need to look at again. I should set aside some time for cleanup.

I agree with Michelle. Design could help consumers decrease their storage. I want my software to encourage deletion, not (or not just) addition, bring it on!

Originally posted on 21 February 2024 on Hidde's blog.

Reply via email

Invisible systems

2024-02-28T00:00:00Z

On the work the GOV.UK Design System team do:

it’s the invisible systems work that has a bigger impact. Reviewing. Advising. Organising. Co-ordinating. Triaging. Educating. Supporting. Allowing the innovation happening at the edges of the ecosystem to feed back into the centre, to be consolidated and standardised for the benefit of everyone.

(From: How far we’ve come: What it would mean to lose the GOV.‌UK Design System)

Originally posted on 28 February 2024 on Hidde's blog.

Reply via email

Switch in HTML

2024-02-29T00:00:00Z

Apple is experimenting with a new HTML form control: a switch (see WHATWG/HTML issue #9546). It is designed as an attribute for <input type="checkbox">, you'd turn a checkbox into a switch by adding the switch attribute:

<input type=checkbox switch checked>

In terms of pseudos:, they're experimenting with ::thumb and ::track pseudo elements for styling the parts of the switch. Unlike the checkbox, it has no ::indeterminate pseudo class, because it has no indeterminate state.

The colour can be set with accent-color. For browsers that don't support this new switch attribute, the element simply falls back to a checkbox.

There is some accessibility support: a switch gets a switch role under the hood, and the element respects the “differentiate without color“ setting in iOS and “on/off labels” on iOS.

Their blog post on when to use it:

Generally, we recommend using a switch when the end user understands the user interface element as a setting that is either “on” or “off”. A checkbox is well suited for when the end user would understand the element as something to be selected.

(From: An HTML Switch Control | WebKit)

Originally posted on 29 February 2024 on Hidde's blog.

Reply via email

Hallucination is inevitable

2024-03-01T00:00:00Z

Researchers show that hallucination is inevitable:

LLMs cannot learn all of the computable functions and will therefore always hallucinate. Since the formal world is a part of the real world which is much more complicated, hallucinations are also inevitable for real world LLMs.

(From: [2401.11817] Hallucination is Inevitable: An Innate Limitation of Large Language Models)

Originally posted on 1 March 2024 on Hidde's blog.

Reply via email

Jakob Nielsen's problematic claims about accessibility

2024-03-07T00:00:00Z

Jakob Nielsen wrote a post in which he states “the accessibility movement has been a miserable failure’ (his words) and claims that generative “AI” can somehow magically remove the need for accessibility research and testing.

Note, there's currently no evidence that what he proposes is desirable (by users) or possible (with the tech). It is, however, clear that testing with users and meeting WCAG is desirable and possible.

Léonie explains Nielsen needs to think again:

Nielsen thinks accessibility has failed.

Nielsen thinks that generative AI will make my experience better. Nielsen apparently doesn't realise that generative AI barely understands accessibility, never mind how to make accessible experiences for humans.

I think Nielsen needs to think again.

Matt May said we need to talk about Jakob:

This part of the post isn’t so much an argument on the merits of disabled access as it is a projection of himself in the shoes of a blind user, and how utterly miserable he thinks it would be. At no point in any of this—again, classic Jakob Nielsen style—does he cite an actual blind user, much less any blind assistive technology researchers or developers

Per Axbom wrote:

the published post is misleading, self-contradictory and underhanded. I'll walk you through the whole of it and provide my commentary and reasoning.

Originally posted on 7 March 2024 on Hidde's blog.

Reply via email

Broaden your frame of reference

2024-03-11T00:00:00Z

Sean Voisen recommends to not stick to a particular technology:

Lose the label and become T-shaped. Stay curious. Keep learning. Go deep in a specific technology or framework or programming language, but develop breadth in adjacent technologies that will help inform your work and develop new perspectives.

(From: On being a ‹insert favorite technology here› “guy” | Sean Voisen)

Coincidentally, Jonathan Snook posted similar advice this week, in Shifting identifies.

Originally posted on 11 March 2024 on Hidde's blog.

Reply via email

Simpler businesses

2024-03-11T00:00:00Z

All of this leaves me wanting simpler businesses with simpler motives — I'll pay, you provide a product or service commensurate with the value. No opaque policies, no concerns about data. I'd love for you to be profitable and sustainable, without being obsessed with scale. I'd love you to build products for the customers, not the speculators, that have invested in you.

(From: Of course AI is extractive, everything is lately • Cory Dransfeldt)

Originally posted on 11 March 2024 on Hidde's blog.

Reply via email

W3C and AI

2024-03-13T00:00:00Z

The W3C established articificial intelligence is having a “systemic impact on the web” and looked at how standardisation, guidelines and interoperability can help manage that:

Machine Learning models support a new generation of AI systems. These models are often trained on a large amount of Web content, deployed at scale through web interfaces, and can be used to generate plausible content at unprecedented speed and cost.

Given the scope and scale of these intersections, this wave of AI systems is having potential systemic impact on the Web and some of the equilibriums on which its ecosystem had grown.

This document reviews these intersections through their ethical, societal and technical impacts and highlights a number of areas where standardization, guidelines and interoperability could help manage these changes

(From: AI & the Web: Understanding and managing the impact of Machine Learning models on the Web)

Originally posted on 13 March 2024 on Hidde's blog.

Reply via email

MEPs adopt new and first AI law

2024-03-14T00:00:00Z

On Wednesday, Parliament approved the Artificial Intelligence Act that ensures safety and compliance with fundamental rights, while boosting innovation.

(…)

It aims to protect fundamental rights, democracy, the rule of law and environmental sustainability from high-risk AI, while boosting innovation and establishing Europe as a leader in the field. The regulation establishes obligations for AI based on its potential risks and level of impact.

(From: Artificial Intelligence Act: MEPs adopt landmark law | News | European Parliament)

Originally posted on 14 March 2024 on Hidde's blog.

Reply via email

Touchscreen accessibility

2024-03-17T00:00:00Z

Touch screens and buttonless designs on devices have become the norm, not a definition of the ultra-modern any more. Which means, as a blind individual, that finding accessible household appliances has become increasingly challenging.

(From: Devising devices with accessibility in mind)

Originally posted on 17 March 2024 on Hidde's blog.

Reply via email

Alt texts as meta data would and the need for context

2024-03-20T00:00:00Z

The idea of including alt text for images as metadata into image files pops up every now and then.

Eric Bailey explains some of the many reasons why this isn't as good of an idea as it seems:

The largest thing to grapple with is that images are contextual. Choosing to select and share one is a highly intentional act, and oftentimes requires knowing the larger context of how it will be viewed.

(From: Thoughts on embedding alternative text metadata into images – Eric Bailey)

He explains describing images is a human to human thing, not a “problem” that just needs some tech thrown at it. Even if some of the tech can in some ways be helpful and powerful.

Originally posted on 20 March 2024 on Hidde's blog.

Reply via email

Content that's worth our time

2024-03-26T00:00:00Z

Cory Dransfeldt explains that while we are developing technology that can generate and produce a larger amount of content, the real problem is the quality of that content:

I'm more and more concerned that we're heading to a place that will make it ever more difficult to find anything that's actually worth our time.

(From: We have a content quality problem, not a content quantity problem // Cory Dransfeldt)

Originally posted on 26 March 2024 on Hidde's blog.

Reply via email

AI uses too much energy

2024-03-30T00:00:00Z

If ChatGPT were integrated into the 9 billion searches done each day, the IEA says, the electricity demand would increase by 10 terawatt-hours a year — the amount consumed by about 1.5 million European Union residents.

(From: AI already uses as much energy as a small country. It’s only the beginning. - Vox)

This is from an interview with Sasha Luccioni, climate researcher at Hugging Face. In it, she explains what the power and water consumption of AI, specifically LLMs, looks like today. It's bad, the amount of energy required is enormous. One example in the post is that a query to an LLM cost almost 10 times as much energy as a query to a regular search engine. That's unsustainable, even if we manage to use 100% renewable energy and water that we really didn't need for anything else.

Once again, this begs the question if we really need all the AI applications companies are rushing into their products. It's often completely unnecessary.

It reminds me of eating animals. With all we know about animal welfare and climate impact, we've got to consider if (regularly) eating animals has benefits that outweigh those downsides.

Everyone can choose to do whatever they want with the information they have available to them. As a person or as a company. But if you're deciding for a company, the impact is larger, it's the decision times the amount of users. For me it's increasingly clear I don't want to use these “AI” solutions in personal workflows, suggest we might as well use them when I give talks, let alone push for integrating them into the products I work on.

Originally posted on 30 March 2024 on Hidde's blog.

Reply via email

WebAIM Million 2024

2024-03-30T00:00:00Z

The WebAIM Million 2024 report is out! More errors were detected, but also pages with fewer errors generally got better.

If this inspired you to go fix low hanging fruit in your projects, I previously wrote about ways to fix common accessibility issues, and a part 2 with more issues to fix. Making websites perfectly accessible can be hard, but reducing fruit that is both low-hanging and very common, is not.

Originally posted on 30 March 2024 on Hidde's blog.

Reply via email

What ARIA attributes do

2024-04-04T00:00:00Z

Kitty explains the difference between disabled and aria-disabled:

[disabled and the aria-disabled attribute] are both meaningful attributes with their own pros and cons

(From: On disabled and aria-disabled attributes | Kitty Giraudel)

There's a lesson in here that applies more generally: ARIA attributes always merely set ‘accessibility semantics’, they don't have side effects like affecting discoverability. It also means when you use them and want behaviours associated with the attributes, you need to add those yourself. So if you add a button role, it won't behave like a button upon adding that attribute, you need to add click and keyboard handlers (and more) yourself.

Originally posted on 4 April 2024 on Hidde's blog.

Reply via email

AI, accessibility and fiction

2024-04-05T00:00:00Z

This week, once again, someone suggested that “AI” could replace (paraphrasing) normative guidelines (ref: mailing list post of AGWG, the group that produces WCAG).

Eric Eggert explains why this seems unnecessary:

The simple fact is that we already have all the technology to make wide-spread accessibility a reality. Today. We have guidelines that, while not covering 100% of the disability spectrum, cover a lot of the user needs. User needs that fundamentally do not change.

(From: “AI” won’t solve accessibility · Eric Eggert)

I cannot but disagree with Vanderheiden and Nielsen. They suggest (again, paraphrasing) that we can stop making accessibility requirements, because those somehow “failed” (it didn't, WCAG is successful in many ways) and because generative AI exists.

Of course, I'm happy and cautiously optimistic that there are technological advancement. They can meet user needs well, like how LLMs “effectively made any image on the Web accessible to blind people”, as Léonie Watson describes in her thoughtful comment. If people want to use tools meet their needs, great.

But it seems utterly irresponsible to have innovation reduce websites' legal obligations to provide basic accessibility. Especially while there are many unresolved problems with LLMs, like hallucinations (that some say are inevitable), environmental cost, bias, copyright and social issues (including the working conditions of people categorising stuff).

Originally posted on 5 April 2024 on Hidde's blog.

Reply via email

Statistical illusion

2024-04-07T00:00:00Z

Baldur Bjarnason, author of the excellent “The intelligence illusion”, on business risks of Generative AI (recommended!):

Delegating your decision-making, ranking, assessment, strategising, analysis, or any other form of reasoning to a chatbot becomes the functional equivalent to phoning a psychic for advice.

(From: The LLMentalist Effect: how chat-based Large Language Models replicate the mechanisms of a psychic's con)

In his post, Baldur warns us once again not to imagine functionality that doesn't exist, he says it's all a ‘statistical illusion’.

Originally posted on 7 April 2024 on Hidde's blog.

Reply via email

Opening

2024-04-21T00:00:00Z

To convince a reader or conference attendee that your content is something to pay attention to, try opening strong.

I don't think I'm very good at this, so I loved Maggie Appleton's latest piece. It's full of useful advice:

For your writing to be worth reading, you need to be exploring something of consequence for someone. You have to have some kind of problem that matters.

(…)
Once you know you have a consequential problem for a community and some sense of a solution, you get to play with narrative details. This is the fun storytelling part.

(From: On Opening Essays, Conference Talks, and Jam Jars)

Originally posted on 21 April 2024 on Hidde's blog.

Reply via email

Attributes and properties

2024-04-25T00:00:00Z

Attributes and properties are fundamentally different things.

(From: HTML attributes vs DOM properties - JakeArchibald.com)

Originally posted on 25 April 2024 on Hidde's blog.

Reply via email

Popover in Baseline

2024-04-29T00:00:00Z

With Firefox 125 shipping the feature, good news on popover:

This web feature is now available in all three major browser engines, and becomes Baseline Newly Available as of April 16, 2024.

(From: Popover API lands in Baseline | Blog | web.dev)

Originally posted on 29 April 2024 on Hidde's blog.

Reply via email

Reducing complexity

2024-04-29T00:00:00Z

Tim Paul on how complexity can increase unexpectedly if we automate:

handling complexity isn't the same as reducing it.

In fact, by getting better at handling complexity we're increasing our tolerance for it. And if we become more tolerant of it we're likely to see it grow, not shrink.

(From: Tim Paul | Automation and the Jevons paradox)

Originally posted on 29 April 2024 on Hidde's blog.

Reply via email

Screenreader only component

2024-04-29T00:00:00Z

Donny D'Amato on making a design system component for content that is meant for screenreaders:

there has been one concept that I’ve stuggled to put into this component-driven ecosystem; screenreader only as it has traditionally existed as a class (eg., .sr-only) added to an otherwise benign element

(From: Screenreader only)

Originally posted on 29 April 2024 on Hidde's blog.

Reply via email

EN 301 549 vs WCAG

2024-05-03T00:00:00Z

Accessibility standards veteran and axe-core product owner Wilco Fiers explains how he sees EN 301 549 relate to WCAG:

EN 301 549 steadily gained importance. It is often dismissed as “WCAG with a different number slapped on it,” but it is far more than that.

(From: 301,549 ways to improve accessibility: EN 301 549 | Deque)

In the post, he explains the EN is broader than WCAG in various ways:

it has more requirements, like for browser settings to be respected by websites (11.7 User preferences)
it applies to more than web content (apps, kiosks), a scope WCAG explicitly doesn't support
it and derivatives of it apply in more and more places, way beyond the EU (Canada, Japan, Australia)

On the authoring tool requirements bit that Wilco mentioned, I wrote a blog post to summarise our group's thoughts: On authoring tools in EN 301 549.

Originally posted on 3 May 2024 on Hidde's blog.

Reply via email

Less perfectionism, more humanism

2024-05-03T00:00:00Z

In Branch, Michelle Barker suggests a new mindset for the web:

If “move fast and break things” is Silicon Valley’s rallying cry, then the flip-side is “move slow and mend things”.

(From: The perfect site doesn’t exist - Branch)

She says we should center sustainability in our work on the web, be intentional in content and code, and prioritise being human to being perfectionist. I could not agree more.

Originally posted on 3 May 2024 on Hidde's blog.

Reply via email

Public speaking tips

2024-05-04T00:00:00Z

I’ve got a few tips for you--and anyone--about speaking. I think I can sum it all up pretty well in one phrase: GET COMFORTABLE.

(From: “Get Comfortable,” an article by Dan Mall)

I hadn't seen this post by Dan Mall about speaking when it came out, but read it as it was shared on Mastodon today. This week I gave a talk that I was quite nervous for beforehand and quite happy with afterwards; from my experience I would second all the advice he gives here on public speaking.

Originally posted on 4 May 2024 on Hidde's blog.

Reply via email

Live regions

2024-05-07T00:00:00Z

Live regions have a reputation for being "flaky" and inconsistent. While this can be attributed in part to shortcomings in current implementations, the problem can also be caused by developers misunderstanding how live regions are intended to work.

(From: Why are my live regions not working? - TetraLogical)

Excellent post by Patrick Lauke!

Originally posted on 7 May 2024 on Hidde's blog.

Reply via email

Mastodon's DDoS'ing

2024-05-07T00:00:00Z

I've noticed that whenever I (or someone else) posts a link to this blog on Mastodon, the decentralised nature of the platforms effectively DDoS's me.

(From: Mastodon Is DDoSing Me | Kev Quirk)

Originally posted on 7 May 2024 on Hidde's blog.

Reply via email

Humane web

2024-05-10T00:00:00Z

Bravo to Michelle Barker putting this out there, putting into words what many of us feel:

The web is becoming hostile to humans. Users are tracked and their privacy is routinely violated. Search results are populated with ads. We are constantly spammed by bots. Generative AI threatens to turn previously useful public forums into soulless marketing soup, while sacrificing the livelihoods of the creators that unwittingly power them. Power-hungry data centres demand the burning of fossil fuels, and divert water and energy from communities, emitting tonnes of carbon in order to power this digital junkyard. Users abandon hostile websites that take too long to load on low-powered devices, or are forced to upgrade, as the pile of electronic waste grows. We need a web by and for humans.

(From: Manifesto for a Humane Web)

Originally posted on 10 May 2024 on Hidde's blog.

Reply via email

Want it all

2024-05-10T00:00:00Z

Ana Rodrigues:

I want it all, but it is impossible.

(From: Oh Hello Ana - I want it all but, it is impossible)

Originally posted on 10 May 2024 on Hidde's blog.

Reply via email

Inspiration porn

2024-05-15T00:00:00Z

Just in time for #GAAD, and in the week where we saw OpenAI share a video of a new feature that integrates the GPT-4o model with BeMyEyes to assist blind people, I want to share the definition of “inspiration porn”:

Inspiration porn is the portrayal of people with disabilities (…) as being inspirational to able-bodied people (…), on the basis of their life circumstances.

(From: Inspiration porn - Wikipedia)

The demo itself was impressive, but I couldn't not notice that OpenAI shared it without captions. It could be that they just forgot. But they clearly didn't forget to upload record, edit, upload and promote the video, so accessibility must have been lower or no priority. And that's likely to be a systemic issue, a bit like some of the other announcement videos that portray assistants as somewhat flirty women.

The term “inspiration porn” has been used in various places, I found the concept particularly well explained in “Against Technoableism” by Ashley Shew and “Handicap” by Anaïs Van Ertvelde. I believe it was originally coined by comedian and journalist Stella Young, see also her TED Talk.

Originally posted on 15 May 2024 on Hidde's blog.

Reply via email

Baseline and accessibility

2024-05-17T00:00:00Z

Currently, the Baseline support information does not include whether the feature is supported accessibly:

If you need to understand whether browsers support accessibility features as your own base level set of requirements, for legal or other compliance reasons, then Web Platform Baseline does not represent a baseline.

(From: Baseline Does Not Really Cover Baseline Support — Adrian Roselli)

I support Adrian's call for this to change. We should't assume features to be supported, if there are known accessibility issues in specific browsers or assistive technologies. I previously noted I feel Interop should include accessibility (and there is a separate Interop project for accessibility now, nice). I hope Baseline can meaningfully implement updates to make accessibility support more clear. A disclaimer, as Adrian suggested seems a good first step.

What would that it mean to deem a feature supported accessibly? WCAG's accessibility supported definition is a good start, but the W3C or Accessibility Guidelines Working Group (that makes WCAG) don't say exactly which browsers or assistive technologies a feature needs to work. For users of assistive technologies, ‘it works’ depends both their choice of assistive technology and the browser they use it with. But even if there's no clear cut answer to ‘which browsers or AT’, even the information that there are known bugs and where would tremendously help web developers decide whether to use a feature or not.

Originally posted on 17 May 2024 on Hidde's blog.

Reply via email

Design system dilemmas

2024-05-23T00:00:00Z

Building a component library is a challenging and rewarding experience, but there’s more to it than writing great-looking components. At every step of the journey, you’ll need to make choices to ensure your library is built appropriately for its intended audience.

(From: andrico1234/the-dilemmas-youll-face: The Dilemmas You'll Face When Creating Your First Component Library)

Originally posted on 23 May 2024 on Hidde's blog.

Reply via email

For idea guys

2024-05-23T00:00:00Z

Rachel Smith on makers vs idea guys:

Generative AI is like the ultimate idea guy’s idea! Imagine… if all they needed to create a business, software or art was their great idea, and a computer. No need to engage (or pay) any of those annoying makers who keep talking about limitations, scope, standards, artistic integrity etc. etc.

(From: Generative AI is for the idea guys)

Originally posted on 23 May 2024 on Hidde's blog.

Reply via email

Legitimate

2024-05-25T00:00:00Z

Jeremy noticed that an Instagram
notification said:

we’ll now rely on the legal basis called legitimate interests for using your information to develop and improve AI

(From: Adactio: Journal—InstAI)

That's not by any means reasonably what the word legitimate means, is it?

It's unfortunate many interesting people and businesses can mostly or only be followed on Instagram, as that's pretty much why I still keep an account. This feels like the social media equivalent of being kept hostage.

Originally posted on 25 May 2024 on Hidde's blog.

Reply via email

Masonry and reading order

2024-05-27T00:00:00Z

CSS can lay things out for you automatically, which is cool but could mean reading order and visual order end up not matching. Luckily, Rachel and others at the CSS Working Group are working on a solution, that lets developers choose which order makes most sense:

There’s a proposal however that aims to deal with this, that would let developers indicate to the browser that they want to follow the “visual” flow of items rather than source order. This is currently named reading-order-items, and I recently added a draft of the proposal to the CSS Display Level 4 editor’s draft. The specification deals with reordering both in an automatic sense, but also the reordering you might want to do when placing items on the grid.

(From: Masonry and reading order – Rachel Andrew)

Originally posted on 27 May 2024 on Hidde's blog.

Reply via email

Alt generation in Firefox

2024-06-02T00:00:00Z

Firefox experiments with automatic text alternative text generation, using a local and therefore privacy-preserving (?) machine learning model:

Until recently it’s not been feasible for the browser to infer reasonably high quality alt text for images, without sending potentially sensitive data to a remote server. However, latest developments in AI have enabled this type of image analysis to happen efficiently, even on a CPU.

We are adding a feature within the PDF editor in Firefox Nightly to validate this approach. As we develop it further and learn from the deployment, our goal is to offer it for users who’d like to use it when browsing to help them better understand images which would otherwise be inaccessible.

(From: Experimenting with local alt text generation in Firefox Nightly - Mozilla Hacks - the Web developer blog)

This is good to see as so many websites lack text alternatives and this may be the first of its kind made by a company that didn't take part in large scale user privacy violations.

Originally posted on 2 June 2024 on Hidde's blog.

Reply via email

Friends and AI

2024-06-02T00:00:00Z

Neven Mrgan received an email:

my friend had a question to ask me, and the email asked it over the course of a few paragraphs. It then disclosed that, oh by the way, I used AI to write this

(From: How it feels to get an AI email from a friend)

In his post he talks about what it feels like to be on the receiving end of AI generated content, in this case one where you'd hope these tools aren't used: an email from a friend. Not for grammar checks, but for the actual message. It felt off:

It felt like getting a birthday card with only the prewritten message inside, and no added well-wishes from the wisher’s own pen.

Originally posted on 2 June 2024 on Hidde's blog.

Reply via email

WPT and accessibility

2024-06-07T00:00:00Z

Rahim Abdi of Apple describes an effort to make it possible to test more web platform features for accessibility features:

what if we could regularly test the accessibility behavior of any and all web platform features on the latest browsers in an automated fashion? How much time and effort could this save?

(From: Improving Web Accessibility with Web Platform Tests | WebKit)

Originally posted on 7 June 2024 on Hidde's blog.

Reply via email

The opposite of human creativity

2024-06-11T00:00:00Z

Apple's ethos3 has always been about building tools to empower users to make art, to create, to be original. I don't know what is is, but it sure as hell isn't human creativity.

(From: roguelazer's website: oh, the humanity)

Originally posted on 11 June 2024 on Hidde's blog.

Reply via email

The ch unit for line lengths

2024-06-17T00:00:00Z

Richard Rutter explains that maybe the ch unit isn't the best choice if you're trying to follow Robert Bringhurt's guidelines:

The important part of Bringhurst’s guideline is not the ’66-characters’ but the ‘satisfactory length’. This is about readability, and readability is affected by the length of a line more so than the number of characters in it.

(From: Use of ch unit considered inappropriate (in certain circumstances) | Clagnut by Richard Rutter)

Originally posted on 17 June 2024 on Hidde's blog.

Reply via email

Synergy Greg

2024-06-20T00:00:00Z

This post is a bit violent at times, but has some very, very good points on the “AI” hype from someone who actually knows the technology and sees through the hype:

You either need to be on the absolute cutting-edge and producing novel research, or you should be doing exactly what you were doing five years ago with minor concessions to incorporating LLMs. Anything in the middle ground does not make any sense unless you actually work in the rare field where your industry is being totally disrupted right now.

(From: I Will Fucking Piledrive You If You Mention AI Again — Ludicity)

Originally posted on 20 June 2024 on Hidde's blog.

Reply via email

If statements in CSS today

2024-06-21T00:00:00Z

Lea Verou explains how we could do if statements, that are coming to CSS, today.

She explored various very clever ways to do it.

Great point on abstractions:

Ugliness is only acceptable if it’s encapsulated and not exposed to component users.

(From: Inline conditionals in CSS, now? • Lea Verou)

Originally posted on 21 June 2024 on Hidde's blog.

Reply via email

Philosophically bullshit

2024-06-29T00:00:00Z

LLMs don't hallucinate or lie, they ‘bullshit’, in the sense that the late philosopher Harry Frankfurt, explain Glasgow researchers in their recent paper:

The problem here isn't that large language models hallucinate, lie, or misrepresent the world in some way. It's that they are not designed to represent the world at all; instead, they are designed to convey convincing lines of text.

(From: ChatGPT is bullshit)

The paper explains Frankfurt's interesting distinction between ‘soft bullshit’ and ‘hard bullshit’, reasoning that ChatGPT is definitely the former and in some cases arguably the latter.

It's crucial to replace phrases like ‘hallucinate’ or ‘lie’ with a word like ‘bullshit’, not to try and be witty, but because the phrases shape how investors, policymakers and general public think of these tools. Which in turn impacts the decisions they make about using them.

Originally posted on 29 June 2024 on Hidde's blog.

Reply via email

Mediocre and derivative input

2024-07-01T00:00:00Z

Scott Riley:

Figma’s AI shit will suffer from the same problems every other company’s GenAI shit suffers from: the average input to its dataset is, almost by definition, mediocre and derivative. Especially when you consider the state we’re in by and large as an industry.

(From: On AI and the commoditisation of design – Scott Riley)

Originally posted on 1 July 2024 on Hidde's blog.

Reply via email

Reading page

2024-07-02T00:00:00Z

Melanie Richards:

In all my years as an absolute book fiend with a personal website, I’ve never hosted my reading list here! That’s now changed

(From: New Reading page, powered by the Airtable API | Melanie Richards)

Love these kinds of pages, one of the hardest parts about reading more is to find out what's worth reading. Another person's opinion is worth more than a thousand algorithms.

Originally posted on 2 July 2024 on Hidde's blog.

Reply via email

We need information architects

2024-07-05T00:00:00Z

Vicky Teinaki urges us to revisit the discipline of information architecture:

for anyone designing flows with any amount of complexity (such as things that have repeat use), please have a look at information architecture as a discipline.

(From: A plea for the lost practice of information architecture)

In her post, she explains how the field got mostly erased over the last 15-20 years, with the rise of the ‘UX designer’ first then, followed by the popularity of agile delivery and lean startup.

IA is considered as ‘unsexy’, she explains, especially compared to full blown prototypes in Figma, but critical when you're designing complex structures: they need to be thought out or it will show.

Originally posted on 5 July 2024 on Hidde's blog.

Reply via email

The problem is with energy

2024-07-06T00:00:00Z

The problem isn’t that AI is using "too much" power from our current grid; it’s that our current grid still overwhelmingly runs on fossil fuels in the first place.

(From: @Westenberg | A question of power)

Originally posted on 6 July 2024 on Hidde's blog.

Reply via email

The long closed site that got revitalised as a zombie AI version

2024-07-11T00:00:00Z

TUAW (“The Unofficial Apple Weblog”) was shut down by AOL in 2015, but this past year, a new owner scooped up the domain and began posting articles under the bylines of former writers who haven’t worked there for over a decade.

(From: Early Apple tech bloggers are shocked to find their name and work have been AI-zombified - The Verge)

The content on the relaunched site was LLM-generated, including author names and pictures. But then they used real author names from people who used to work at the site. Very uncanny.

After one of the former TUAW writers posted about what happened and threatened with legal action, the names have now been changed.

Originally posted on 11 July 2024 on Hidde's blog.

Reply via email

What users thing vs what corporations think

2024-07-13T00:00:00Z

The corporate branding, the new “AI-powered developer platform” slogan, makes it clear that what I think of as “GitHub”—the traditional website, what are to me the core features—simply isn’t Microsoft’s priority at this point in time.

(From: "GitHub" Is Starting to Feel Like Legacy Software - The Future Is Now)

Originally posted on 13 July 2024 on Hidde's blog.

Reply via email

30 times easier

2024-07-27T00:00:00Z

Accessibility is easier when you do it earlier, I can't emphasise that enough. In his latest post, Eric posts research that shows how much easier:

Accessibility audits almost always happen after launch, as an afterthought. That means that errors that could have been found in the requirement analysis are 30 times harder to fix2 . This makes accessibility audits seem very inefficient.

(From: The infuriating inefficiency of accessibility audits · Eric Eggert)

Audits aren't useless. They have their place and are essential for monitoring accessibility within large organisations and for governments. Still, for your website or team, finding and fixing issues early is ideal. And it makes those audits easier and quicker too, as they'll need to report less low hanging fruit.

Originally posted on 27 July 2024 on Hidde's blog.

Reply via email

Money, money, money

2024-07-29T00:00:00Z

The founders and namesakes of the famous Silicon Valley venture capital firm A16Z decided to back Trump, and discussed why on a podcast.

Elizabeth Lopatto of The Verge took one for the team and listened to the whole thing. It sounds like their sole reason for backing this extreme party, ultimately, is money:

this VC cabal is trading against the basic principles of America — not merely against personal freedom, but democracy itself — in the hopes of profit.

(From: The moral bankruptcy of Marc Andreessen and Ben Horowitz - The Verge)

Selfishness and motivation by money isn't new (or inherently wrong). But A16Z-backed companies used to allude to more inspiring ideals, like changing the world for the better by connecting everyone. The marketing/morality ratio increasingly seems off.

Originally posted on 29 July 2024 on Hidde's blog.

Reply via email

TAG on third party cookies

2024-07-30T00:00:00Z

Hadley Beeman of the Technical Architecture Group:

After reading Google’s announcement that they no longer plan to deprecate third-party cookies, we wanted to make our position clear.

(From: Third-party cookies have got to go | 2024 | Blog | W3C)

Originally posted on 30 July 2024 on Hidde's blog.

Reply via email

Community survey

2024-08-07T00:00:00Z

W3C is looking for community input:

We want to get to know our community better, investigate needs, and understand our community’s vision of how we fulfill our mission for the world-wide web.

(From: W3C opens community-wide survey | 2024 | News | W3C)

Originally posted on 7 August 2024 on Hidde's blog.

Reply via email

New browser

2024-08-12T00:00:00Z

Ladybird is a new browser, written from scratch, including the engine. It plans not to be funded by ads or ad companies:

The world needs a browser that puts people first, contributes to open standards using a brand new engine, and is free from advertising’s influence.

(From: Why we need Ladybird)

Originally posted on 12 August 2024 on Hidde's blog.

Reply via email

Apple to crush creation, again

2024-08-13T00:00:00Z

Thom Holwerda on the 30% fee Patreon is going to be forced to pay to Apple when users want to donate to creators on iOS:

Remember that ad Apple made where it crushed a bunch of priceless instruments and art supplies into an iPad – the ad it had to pull and apologise for because creators, artists, writers, and so on thought it was tasteless and dystopian?

Who knew that ad was literal.

(From: Apple forces Patreon to charge Patreons 30% tax, or have its iOS application banned from the App Store – OSnews)

What happened to the company that won respect with Think Different?

Originally posted on 13 August 2024 on Hidde's blog.

Reply via email

Spoilers

2024-08-23T00:00:00Z

Scott O'Hara shares his thoughts around the ideal way to code spoilers on the web:

I’m just going to tell you what I’d expect from a spoiler component if someone were to build one, or if one were to ever be standardized.

(From: Spoiler Alert: it needs to be accessible | scottohara.me)

Originally posted on 23 August 2024 on Hidde's blog.

Reply via email

inert in React

2024-08-23T00:00:00Z

Mayank explains how to use the inert attribute in React 19, in earlier versions and with or without TypeScript:

The answer is a bit tricky.

(From: Using the inert attribute in React)

Originally posted on 23 August 2024 on Hidde's blog.

Reply via email

Dimensions to meanings

2024-08-28T00:00:00Z

I think there are multiple implicit dimensions to the meanings of behaviour words. That compounds questions about where to draw boundaries, and it can lead to discussion at cross purposes and confusion.

(From: Do LLMs REALLY reason, understand, think, summarise...? — UlrikeHahn)

Originally posted on 28 August 2024 on Hidde's blog.

Reply via email

Centrism

2024-09-03T00:00:00Z

The stance that works in standards making and in the workplace in general, is ‘quicksand’ when it comes to politics, Joan Westenberg argues:

The fatal flaw of Radical Centrism is its obsession with the middle ground, as if that's inherently where truth and wisdom reside.

(From: @Westenberg | The Radical Centrism Fallacy)

Originally posted on 3 September 2024 on Hidde's blog.

Reply via email

Introducing carbon

2024-09-05T00:00:00Z

Through our carbon.txt project we propose a standardised, yet distributed place on website domains, to efficiently surface the structured data that companies now have to publish anyway.

Carbon.txt would become a single place to look on any domain for public sustainability data relating to that company, that would allow anyone to build a database.

(From: Introducing carbon.txt - Applying lessons from crowdsourcing net zero data - Green Web Foundation)

Originally posted on 5 September 2024 on Hidde's blog.

Reply via email

Social media

2024-09-05T00:00:00Z

On and off for the last several years I’ve been manually curating my roughly 40,000 lifetime tweets. I recently finished, and in the process embarked on an unexpected journey of self-discovery.

(From: Emily F. Gorcenski)

Loved this post!

Originally posted on 5 September 2024 on Hidde's blog.

Reply via email

Blandness vs absurdity

2024-09-10T00:00:00Z

Have been nodding along to this post, that touches on a lot of the themes I plan to bring to Beyond Tellerrand in November:

as AI gets better at mimicking human communication, the pressure on human creators to be weirder, more original, and more authentically human will only increase.

(From: @Westenberg | Shitposting Our Way Through the Singularity)

Originally posted on 10 September 2024 on Hidde's blog.

Reply via email

Real things by real people

2024-09-11T00:00:00Z

yup:

I want real things by real people. I don’t want more things averaged out by a language model that can only make likely sentences. I don’t want more creepy images directly sourced from thousands of copyrighted works. I want you to put yourself on the page.

(From: A short note on AI – Me, Robin)

Originally posted on 11 September 2024 on Hidde's blog.

Reply via email

Knowledge and context

2024-09-13T00:00:00Z

Artificial intelligence companies deeply underestimate how perfect the things around us are, and how deeply we base our understanding and acceptance of the world on knowledge and context. People generally have four fingers and a thumb on each hand, hammers have a handle made of wood and a head made of metal, and monkeys have two legs and two arms. The text on the sign of a store generally has a name and a series of words that describe it, or perhaps its address and phone number.

These are simple concepts that we learn from the people and places we see as we grow up, and what's very, very important to remember is that these are not concepts that artificial intelligence models are aware of.

(From: Subprime Intelligence)

Originally posted on 13 September 2024 on Hidde's blog.

Reply via email

Making it easier to build good UX

2024-09-13T00:00:00Z

Mathia⁠s Schäf⁠er wrote a balanced post about the JavaScript discourse and where to put the blame for poor user experience.

I like what he wrote about the impact of tooling:

We need tools that make it easy to do the right thing and hard to impair the UX. Tools that inform the developer about the impact of a decision on performance and reliability.

(From: Something went wrong · molily)

Originally posted on 13 September 2024 on Hidde's blog.

Reply via email

No personal data harvesting in Europe

2024-09-19T00:00:00Z

Glad the AI Act seems to effectively protect my rights:

Anyone living in the EU, EEA or Switzerland will not have their data harvested. LinkedIn has not yet confirmed why it has spared the citizens of Europe, but it may be due to rules introduced under the EU AI Act.

(From: LinkedIn trains GenAI models on personal data by default)

Originally posted on 19 September 2024 on Hidde's blog.

Reply via email

The temporality of making

2024-09-22T00:00:00Z

But in eliminating the effort, in refusing the temporality of making, the outcome of an “AI”-driven creative process is a phantasm, an unsubstantiality, something that passes through the world without leaving any trace. A root that twists back upon itself and tries to suck the water from its own desiccated veins.

(From: Coming home | A Working Library)

Originally posted on 22 September 2024 on Hidde's blog.

Reply via email

Value of subtle imperfections

2024-09-22T00:00:00Z

In our rush to digitize the world, we often underestimate the value of the patina, subtle imperfections, and otherwise visible history of the physical objects we choose to digitize.

(From: Wendell Berry on the benefits of writing without a computer | Sean Voisen)

Originally posted on 22 September 2024 on Hidde's blog.

Reply via email

Who requested this feature?

2024-09-26T00:00:00Z

This is creepy, dull and useless. I wish they didn't:

If you think avoiding AI-generated images is difficult as it is, Facebook and Instagram are now going to put them directly into your feeds. At the Meta Connect event on Wednesday, the company announced that it’s testing a new feature that creates AI-generated content for you “based on your interests or current trends” — including some that incorporate your face.

(From: Meta’s going to put AI-generated images in your Facebook and Instagram feeds - The Verge)

Originally posted on 26 September 2024 on Hidde's blog.

Reply via email

Human rights and the next 30 years of web

2024-10-01T00:00:00Z

Last week I saw Nick Doty from the Center of Democracy and Technology give an excellent short talk at the W3C's 30tj birthday event, in which he said:

We need to consider human rights in all the work that we do at W3C in the next 30 years.

(From: Happy 30th Birthday, W3C - Center for Democracy and Technology)

He just published the full text.

Originally posted on 1 October 2024 on Hidde's blog.

Reply via email

Overrate

2024-10-08T00:00:00Z

Iris van Rooij on Ada Lovelace:

Two centuries later, as we are living through yet another AI summer where AI hype and promises of artificial general intelligence (AGI) abound, Ada’s wise words remain relevant as ever. When writing about the “AI” of her time, called the Analytical Engine, she wrote: “It is desirable to guard against the possibility of exaggerated ideas that might arise as to the powers of [AI]. In considering any new subject, there is frequently a tendency [...] to overrate what we find to be already interesting or remarkable”.

(From: Editorial AI Inside: Celebrating Ada and Women in AI | Radboud University)

Originally posted on 8 October 2024 on Hidde's blog.

Reply via email

The selected option

2024-10-19T00:00:00Z

Jake Archibald presents some options for how <selectedoption> would work:

what if the selected

(From: How should work? - JakeArchibald.com)

Originally posted on 19 October 2024 on Hidde's blog.

Reply via email

Gladwell

2024-10-22T00:00:00Z

Gladwell writes like someone who doesn't care about being correct because he doesn't care about being correct! His spitballs are truly spitballs, and he doesn't care where they land.

(From: Forget Gladwell)

Originally posted on 22 October 2024 on Hidde's blog.

Reply via email

Unlicensed use of creative works

2024-10-22T00:00:00Z

The unlicensed use of creative works for training generative AI is a major, unjust threat to the livelihoods of the people behind those works, and must not be permitted

(From: Statement on AI training)

Originally posted on 22 October 2024 on Hidde's blog.

Reply via email

Behind the facade

2024-10-28T00:00:00Z

If the pursuit of an easier, slower and more pleasant life comes at the expense of others, is staying where you are and suffering the right thing to do? Maybe.

(From: How Digital Nomads Are Exploiting the World - Thrillist)

Originally posted on 28 October 2024 on Hidde's blog.

Reply via email

LLMs also hallucinate in medical contexts

2024-10-29T00:00:00Z

This shouldn't surprise anyone, but it turns out LLMs also make up stuff when used by doctors:

[Professors Allison Koenecke and Mona Sloane] determined that nearly 40% of the hallucinations were harmful or concerning because the speaker could be misinterpreted or misrepresented.

(From: Researchers say AI transcription tool used in hospitals invents things no one ever said | AP News)

The article lists some examples: the tools made up violence, racial details and medication out of thin air.

Originally posted on 29 October 2024 on Hidde's blog.

Reply via email

We learn

2024-11-04T00:00:00Z

Humans aren’t trained up. We have experience. We learn. And for us, learning a language, for example, isn’t learning to generate ‘the next token’. It’s learning to work, play, eat, love, flirt, dance, fight, pray, manipulate, negotiate, pretend, invent and think. And crucially, we don’t merely incorporate what we learn and carry on; we always resist. Our values are always problematic. We are not merely word-generators. We are makers of meaning.

(From: Can computers think? No. They can’t actually do anything | Aeon Essays)

Originally posted on 4 November 2024 on Hidde's blog.

Reply via email

Hitting a wall

2024-11-10T00:00:00Z

Just as I argued here in April 2024, LLMs have reached a point of diminishing returns.

The economics are likely to be grim. Sky high valuation of companies like OpenAI and Microsoft are largely based on the notion that LLMs will, with continued scaling, become artificial general intelligence. As I have always warned, that’s just a fantasy.

(From: CONFIRMED: LLMs have indeed reached a point of diminishing returns)

Originally posted on 10 November 2024 on Hidde's blog.

Reply via email

Localising icons

2024-11-18T00:00:00Z

Creating effective, trustworthy communication with language communities means doing the work to make sure your content meets them where they are.

A big part of this is learning about, and incorporating cultural norms into your efforts.

(From: Don’t forget to localize your icons – Eric Bailey)

Originally posted on 18 November 2024 on Hidde's blog.

Reply via email

Laundry and dishes

2024-11-23T00:00:00Z

Writer Joanna Maciejewska on Threads (29 March 2024):

You know what the biggest problem with pushing all-things-AI is? Wrong direction. I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do my laundry and dishes.

(From: Threads)

Originally posted on 23 November 2024 on Hidde's blog.

Reply via email

Steve

2024-11-23T00:00:00Z

Love this new series by Tetralogical, who photographed and chatted with people with disabilities about what they do and how they use the web.

Meet Steve, a photographer from London who is deaf and low vision. He is an ex-civil servant who then went on to do freelance technology journalism and travelled the world.

(From: Meet Steve: a photographer who is deaf and low vision - TetraLogical)

Originally posted on 23 November 2024 on Hidde's blog.

Reply via email

Survival independent of the company

2024-11-26T00:00:00Z

I’m a fan of the Bluesky leadership and engineering team. With the VC money as fuel, I expect their next 12 months or so to be golden, with lots of groovy features and mind-blowing growth. But that’s not what I’ll be watching.

I’ll be looking for ecosystem growth in directions that enable survival independent of the company. In the way that email is independent of any technology provider or network operator.

(From: ongoing by Tim Bray · Why Not Bluesky)

Originally posted on 26 November 2024 on Hidde's blog.

Reply via email

Autofocus

2024-12-02T00:00:00Z

Kilian explains that you usually want to avoid autofocus:

there's a place where autofocus shines: On single-purpose pages containing forms.

(From: Starting off right: Where autofocus shines - HTMHell)

Originally posted on 2 December 2024 on Hidde's blog.

Reply via email

Culture change

2024-12-03T00:00:00Z

culture change is creating the conditions that make it easier to choose to do the right thing, to close the gap between your enacted values and your stated values

(From: the values of work | The Roof is on Phire)

Originally posted on 3 December 2024 on Hidde's blog.

Reply via email

Contributing to browsers

2024-12-04T00:00:00Z

Keith Cirkel:

I recently made my 100th patch to browser engines, which I consider somewhat of a milestone.

(From: 100 patches to 5 browsers in 18 months - Keith Cirkel)

Yeah, wow, that is a milestone and very impressive read. Interesting to learn more about what it involves.

Originally posted on 4 December 2024 on Hidde's blog.

Reply via email

Tech for bad

2024-12-04T00:00:00Z

Big tech doesn't seem against building technology to aid deportation:

The Intercept posed the same question to each company, and requested a yes or no response: Would your company provide the Trump administration with data or other technical services to help facilitate mass deportation operations, either voluntarily, in response to a legal request, or via a paid contract?

(From: These Tech Firms Won’t Tell Us If They Will Help Trump Deport Immigrants)

Originally posted on 4 December 2024 on Hidde's blog.

Reply via email

OpenAI pivots to military

2024-12-05T00:00:00Z

Today, OpenAI is announcing that its technology will be deployed directly on the battlefield.

(From: OpenAI's new defense contract completes its military pivot | MIT Technology Review)

They weren't going to, but now they are…

Originally posted on 5 December 2024 on Hidde's blog.

Reply via email

Accessible components

2024-12-14T00:00:00Z

The value of component libraries that have accessibility ‘built in’ is both immense and often overstated. Two sides of the same coin: yes they're good, but yes it's also risky to say they are some sort of one size fits all solution. They're a blessing and a curse.

Michael Matthews at A11y Quest explains:

merely using an accessible library or framework doesn’t mean your team has somehow outsourced all its accessibility responsibilities

(From: The Myth of 'Accessible Components and Done' | A11y Quest)

He goes on to say in addition to maybe benefitting from an accessible component library, you need to make informed decisions, test with users and ensure expertise within your team. Amen to that!

Originally posted on 14 December 2024 on Hidde's blog.

Reply via email

Features

2024-12-19T00:00:00Z

Like features, tokens can create little piles of technical debt.

(From: Every token is a feature - daverupert.com)

Originally posted on 19 December 2024 on Hidde's blog.

Reply via email

System

2024-12-24T00:00:00Z

Dave:

CSS wants to be a system and it’s not an impossible task to make your own.

(From: CSS wants to be a system - daverupert.com)

Originally posted on 24 December 2024 on Hidde's blog.

Reply via email

Accessibe fined

2025-01-04T00:00:00Z

Accessibe fined 1 million dollars:

The U.S. Federal Trade Commission has fined accessiBe, a startup that claims to make websites more compatible with the screen readers blind people rely on to access the internet, for false advertising and compensating reviewers without disclosing that it sponsored the reviews.

In a proposed order, the FTC would require accessiBe to pay $1 million that may be used to refund the company’s customers, and prohibit accessiBe from overstating the capabilities of its tools. The order would also mandate that accessiBe “clearly and conspicuously” highlight connections to endorsers of its services.

(From: FTC orders AI accessibility startup accessiBe to pay $1M for misleading advertising | TechCrunch)

Originally posted on 4 January 2025 on Hidde's blog.

Reply via email

WordPress sustainability team

2025-01-10T00:00:00Z

Matt Mullenweg is at it again:

Members of the fledgling WordPress Sustainability Team have been left reeling after WordPress co-founder Matt Mullenweg abruptly dissolved the team this week—an action prominent tech journalist Kara Swisher has described as “bizarrely heinous behavior.”

(From: Mullenweg Shuts Down WordPress Sustainability Team, Igniting Backlash - The Repository)

Dude shows once again he has too much power over too large of a project. Fershad Irani wrote that he likely never cared about sustainability anyway.

In Why should there be a WordPress Sustainability Group? Chris Adams explains open source CMSes should have a sustainability group for many reasons, including reducing risk:

If you expect people to use your product or service, you might be expected to help people responsible for compliance when using it now.

Originally posted on 10 January 2025 on Hidde's blog.

Reply via email

Tax holidays and human dignity

2025-01-25T00:00:00Z

Why do the richest and powerful people on earth struggle to inform their decisions with basic ethics? Surely they could use their wealth and power against fascism instead of supporting it, it would show leadership.

Joe Steel says Tim Cook is failing us, and should be using the same skills he used to negotiate tax holidays to negotiate human dignity:

what I would like to see is Tim standing up for what he purports to believe in. Not speaking about things that have inspired him, or donating to charities as some kind of fascist-neutral offset, but devising a way to argue for the protection of human dignity for his customers and employees from this federal government. If he can transact a tax holiday, or tariff carve-outs, he can figure out what to wheel and deal to protect people. Many of whom are his employees and customers.

(From: Tim Cook Is Failing Us)

Originally posted on 25 January 2025 on Hidde's blog.

Reply via email

Real time accessibility

2025-01-27T00:00:00Z

What can we do to make apps for real time collaboration, like Google Docs, more accessible?

This month, W3C published a new document called CTAUR (Collaboration Tools Accessibility User Requirements). Previous docs in this ‘series’ include XAUR (on XR), NAUR (on natural language interfaces like voice agents and chatbots), SAUR (on synchronisation in media, like that of captions and video) and RAUR (on RTC). The Accessible Platform Architectures (APA) Working Group bas been busy in the last few years. Note, these are all non-normative.

CTAUR covers a number of aspects that go beyond WCAG:

Real-time and asynchronous co-editing (including things like where are other people in the document currently)
Annotation of content with comments
Comparing revisions
Suggested changes
Access controls

For each, it describes users needs and corresponding requirements.

Originally posted on 27 January 2025 on Hidde's blog.

Reply via email

Find in page

2025-02-17T00:00:00Z

Christian shares how he used hidden=until-found to make hidden content findable via find in page:

This episode taught me the importance of accommodating diverse navigation patterns and highlighted (once more) how evolving standards can improve accessibility for everyone. But while hidden="until-found" is a good step forward, there's still work to be done to make find-in-page truly universal.

(From: Rethinking Find-in-Page Accessibility: Making Hidden Text Work for Everyone)

Originally posted on 17 February 2025 on Hidde's blog.

Reply via email

Igalia and Reference Target

2025-02-21T00:00:00Z

Igalia got a grant to work towards shipping a new proposal to solve for cross root ARIA:

we’ll work on all the process to get the Reference Target proposal ready to be shipped in the web rendering engines

(From: Solving Cross-root ARIA Issues in Shadow DOM)

Originally posted on 21 February 2025 on Hidde's blog.

Reply via email

Aesthetic for the right

2025-03-06T00:00:00Z

It's embarrassing, destructive, and looks like shit: AI-generated art is the perfect aesthetic form for the far right.

(From: AI: The New Aesthetics of Fascism // New Socialist)

Good post (even if it falsely claims The Hague is the capital of The Netherlands).

Originally posted on 6 March 2025 on Hidde's blog.

Reply via email

Heist

2025-03-10T00:00:00Z

When you use AI, you are probably benefiting from stolen material, explains Toby Walsh, who has been an AI researcher for 40+ years. He calls these tools the “greatest heist in human history”:

I am outraged at the tech companies like OpenAI, Google and Meta for training their AI models, such as ChatGPT, Gemini and Llama, on my copyrighted books without either my consent or offering me or Black Inc any compensation.

The tech companies claim this is “fair use”. I don’t see it this way. Last year, at the Sydney Writers’ festival, I called it the greatest heist in human history. All of human culture is being ingested into these AI models for the profit of a few technology companies.

(From: I have been an AI researcher for 40 years. What tech giants are doing to book publishing is akin to theft | Toby Walsh | The Guardian)

Originally posted on 10 March 2025 on Hidde's blog.

Reply via email

Enlighten those new to the field

2025-03-14T00:00:00Z

In defense of technical writing:

If reading philosophy taught me something, it is that thought can be valuable in any form, provided it’s original. (…) Narrate your failures and your wins. Telling your day-to-day routine might reveal patterns or enlighten those who are new to the field.

(From: Technical writing has a depth issue - passo.uno)

Originally posted on 14 March 2025 on Hidde's blog.

Reply via email

Roles and responsibilities for accessibility

2025-03-14T00:00:00Z

Who doesn't love RACIs? When it comes to ensuring accessibility of your products, almost all roles in the team can contribute something.

The W3C's Web Accessibility Initiative published a mapping of roles and responsibilities to WCAG Success Criteria.

It's still a “draft”, and feedback is welcomed, but it's been worked on for many years.

They've included:

6 role groups: Business, Author (Content), Design, Development, Testing, Admin
a mapping to WCAG Success Criteria of who is primarily or secondarily responsible, and who can contribute
what can everyone do? A list of tasks and who can do them
a decision tree for defining your own responsibility mapping

More info on the project and its aims:
Accessibility Roles and Responsibilities Mapping (ARRM) | Web Accessibility Initiative (WAI) | W3C

Originally posted on 14 March 2025 on Hidde's blog.

Reply via email

All LLM output hallucination

2025-03-16T00:00:00Z

Tante makes the point that actually, all output of LLMs is a hallucination, the true and the false parts:

If using the term hallucination is useful to describe LLM output it is to illustrate the quality of all output. Everything an LLM generates is a hallucination, some might accidentally be true.

(From: It's all hallucinations)

Originally posted on 16 March 2025 on Hidde's blog.

Reply via email

More ethics of AI

2025-03-18T00:00:00Z

Richard wrote about a number of different aspects of AI, including sales people complaining they don't sell, erosion of copyright, design tools and mediocrity, AI as a trick to sack humans and bias:

I read a couple of posts about AI recently, which seemed to hold opposing ideas, but I agreed with them both to some extent. (It’s a radical idea, I know).

(From: Another uncalled-for blog post about the ethics of using AI | Clagnut by Richard Rutter)

Good post, I am glad practitioners continue to share their thoughts beyond the hype.

Originally posted on 18 March 2025 on Hidde's blog.

Reply via email

Tech bros misunderstand stuff

2025-03-19T00:00:00Z

Aaron Ross Powell explains he isn't an AI skeptic and that he finds LLMs “powerful tools with real world use cases”, but that the idea that AGI is near or that art can be made with these tools comes down to a misunderstanding on the part of tech bros:

What’s going on is a confluence of two features of Silicon Valley tech bro culture. First, Silicon Valley tech bros believe that they aren’t just skilled at computer programming, but that they are geniuses to a degree that cuts across all disciplines and realms of accomplishment. (…) The second feature is a basic lack of taste.

(From: Why Tech Bros Overestimate AI's Creative Abilities)

Originally posted on 19 March 2025 on Hidde's blog.

Reply via email

Sysadmins and LLM crawlers

2025-03-20T00:00:00Z

The crawlers that collect data to train LLMs on cost sysadmins lots of time, writes Drew DeVault:

instead of working on our priorities at SourceHut, I have spent anywhere from 20-100% of my time in any given week mitigating hyper-aggressive LLM crawlers at scale

(From: Please stop externalizing your costs directly into my face)

Originally posted on 20 March 2025 on Hidde's blog.

Reply via email

Pick your battles, green software edition

2025-03-22T00:00:00Z

Thomas Broyer on what battles are most worth picking when you want to make software more sustainable:

So, what have we learned so far?

It's important that end users keep their devices longer,

we can't do much about networks,

the location (geographic region and datacenter) of servers matter a lot, more so than how and how much we use them.

(From: Climate-friendly software: don't fight the wrong battle)

Originally posted on 22 March 2025 on Hidde's blog.

Reply via email

Careless people: courageous but incomplete?

2025-03-29T00:00:00Z

Sabhanaz Rashid Diya, who was head of public policy for Bangladesh for Meta, reviewed Sarah Wynn-Willams memoir Careless people, which I'm currently reading.

She says it is incomplete:

the author glosses over her own indifference to repeated warnings from policymakers, civil society, and internal teams outside the U.S. that ultimately led to serious harm to communities.

She explains how the people at headquarters were detached:

Every visit to a country or a high-profile meeting at the World Economic Forum in Davos or the U.N. was the product of weeks of intense coordination across regional policy, legal, security, business, and operations teams. When they left after a few days, teams on the ground like my own had to spend months cleaning up the mess they left behind. That included frequently expending local policy and diplomatic relationships built over a decade, and chasing promises made to policymakers and civil society for more resources that rarely got approved.

She does call the book brave and interesting:

Despite telling an incomplete story, Careless People is a book that took enormous courage to write. This is Wynn-Williams’ story to tell, and it is an important one. It goes to show that we need many stories — especially from those who still can’t be heard — if we are to meaningfully piece together the complex puzzle of one of the world’s most powerful technology companies.

(From: Careless People book review: Sarah Wynn-Williams’ Facebook memoir reveals Meta’s global problems - Rest of World)

Originally posted on 29 March 2025 on Hidde's blog.

Reply via email

Display of power

2025-04-01T00:00:00Z

Tante thinks that Open AI didn't just steal Studio Ghibli's art to show they're still relevant, they did it to move the goalposts, and stretch what people will accept as behaviour:

It’s not that they just picked something cute and accidentally the co-founder of that studio hates their whole approach from the bottom of its heart. OpenAI picked Studio Ghibli because Miyazaki hates their approach.

It is a display of power: You as an artist, an animator, an illustrator, a writer, any creative person are powerless. We will take what we want and do what we want. Because we can.

(From: Vulgar Display of Power)

Originally posted on 1 April 2025 on Hidde's blog.

Reply via email

Europe and US tech

2025-04-04T00:00:00Z

Fascinating read on the weakening position of US big tech firms in Europe:

Technology companies such as Alphabet, Meta, and OpenAI need to wake up to an unpleasant reality. By getting close to U.S. President Donald Trump, they risk losing access to one of their biggest markets: Europe.

(From: The Brewing Transatlantic Tech War | Foreign Affairs)

Originally posted on 4 April 2025 on Hidde's blog.

Reply via email

Standardising AI crawler consent

2025-04-11T00:00:00Z

IETF works on building blocks to let websites declare if crawlers can take their content for traininy:

Right now, AI vendors use a confusing array of non-standard signals in the robots.txt file (defined by RFC 9309) and elsewhere to guide their crawling and training decisions. As a result, authors and publishers lose confidence that their preferences will be adhered to, and resort to measures like blocking their IP addresses.

(From: IETF | IETF setting standards for AI preferences)

Originally posted on 11 April 2025 on Hidde's blog.

Reply via email

World wide web fonts

2025-04-24T00:00:00Z

on web we could simply start our font stacks with Verdana, pick a couple of reasonable fallbacks, and get IKEA branding effectively for free. Everyone wins.

Or at least that was the plan, but there turned out to be a problem that developed over time.

(From: IKEA’s web fonts - Robin Whittleton)

IKEA happily used Verdana on the web. Until they expanded business across Asia and the Middle East, and found the supported languages of Verdana were lacking.

In this post, Robin Whittleton explains how this situation lead to Noto IKEA.

Originally posted on 24 April 2025 on Hidde's blog.

Reply via email

The human input is more interesting

2025-05-05T00:00:00Z

Computer scientist Clayton Ramsey shares reasons why people use generative AI to write: they don't care enough, they believe LLM results are better or the writing was never meant for human consumption anyway.

Focusing mostly on the first two, he concludes he would rather read the prompt than the result:

The resulting output has less substance than the prompt and lacks any human vision in its creation. The whole point of making creative work is to share one’s own experience - if there’s no experience to share, why bother?

(From: I'd rather read the prompt)

This was one of the conclusions in my talk Creativity cannot be computed,too. The point of a lot of art is that some human wanted, intended, decided to do something… not as much the artifact they created.

Originally posted on 5 May 2025 on Hidde's blog.

Reply via email

Value of conferences

2025-05-07T00:00:00Z

Ana is on point re the value of conferences:

The point of a conference shouldn't be to provide ready-made solutions to specific workplace problems. A 40 minute code-heavy presentation might offer some technical pointers, but these will never truly address your specific challenges.

(From: Oh Hello Ana - A forty minute tech talk might not fix a very specific code problem at work, but it might leave you inspired to fix everything else)

Originally posted on 7 May 2025 on Hidde's blog.

Reply via email

WCAG, but for mobile

2025-05-07T00:00:00Z

Can WCAG be applied to mobile? Well, sort of… it maps quite well, as long as you carefully assess how each criterion works in the mobile context.

That's what the W3C's Mobile Accessibility Task Force has done, with a group of experts. They've just published their first public working draft of WCAG2Mobile: Guidance on Applying WCAG 2.2 to Mobile Applications (WCAG2Mobile)

Originally posted on 7 May 2025 on Hidde's blog.

Reply via email

App stores and accessibility info

2025-05-15T00:00:00Z

Blogging ✨ works ✨!

4 years ago, I wrote A case for accessibility statements in app stores.

This week, Apple announced “Nutrition Labels“:

Accessibility Nutrition Labels bring a new section to App Store product pages that will highlight accessibility features within apps and games. These labels give users a new way to learn if an app will be accessible to them before they download it, and give developers the opportunity to better inform and educate their users on features their app supports. This includes VoiceOver, Voice Control, Larger Text, Sufficient Contrast, Reduced Motion, captions, and more.

jk, but very glad to see it, I think the visibility it brings will be helpful!

Originally posted on 15 May 2025 on Hidde's blog.

Reply via email

Privacy Principles

2025-05-15T00:00:00Z

The W3C published Privacy Principles as a Statement, meaning a document the Members approved.

It has definitions, meant to aid policy, as well as principles, to be more precise:

a set of privacy principles that should guide the development of the web as a trustworthy platform

(From: Privacy Principles)

Originally posted on 15 May 2025 on Hidde's blog.

Reply via email

Human dignity

2025-05-25T00:00:00Z

Tina He explains our discomfort with AI:

What unsettles us is AI's supreme indifference, or commonly called "slop," which is also synonymous with the lacking of soul.

Software doesn’t hate, plot, or hold a grudge. It optimizes. That optimization, crucially, includes the effectiveness of language—words as system output, not as evidence of intent.

(From: AI, Heidegger, and Evangelion - by Tina He - Fakepixels)

And then explains that this indifference, that leads to humans being reduced to function, ‘cuts at the heart of our dignity’. Yup yup yup, and that's very, very dystopian.

Originally posted on 25 May 2025 on Hidde's blog.

Reply via email

Heard

2025-05-31T00:00:00Z

When you have numbers to look at, it’s easy to care only about the numbers. If we want to truly care, we gotta kill the metrics in our heads.

(From: kill the metrics in your head | The Roof is on Phire)

Originally posted on 31 May 2025 on Hidde's blog.

Reply via email

Intelligence cannot be measured

2025-05-31T00:00:00Z

Intelligence has never been an objective quality that can be ascertained the way we measure the (actually increasing) carbon in the atmosphere. It is a political device that preserves power and care for those deemed worthy of it, and which simultaneously withdraws such care from everyone else.

(From: Toolmen | A Working Library)

Originally posted on 31 May 2025 on Hidde's blog.

Reply via email

LLMs link back less

2025-07-03T00:00:00Z

Websites allow search engines to crawl them, because these search engines then end up linking to the website. You'd get traffic in return.

LLM providers also crawl, but they link back less, say David Belson and Sam Rhea of Cloudflare:

These Large Language Models (LLMs) do their best to read the web to train a system that can repackage that content for the user, without the user ever needing to visit the original publication.

(From: The crawl before the fall… of referrals: understanding AI’s impact on content providers)

They share the numbers and it is, unsurprisingly, not a little bit less. LLMs refer back magnitudes less.

This is a problem, it makes writing on the web, as a business, even less attractive.

Originally posted on 3 July 2025 on Hidde's blog.

Reply via email

We can be fooled

2025-07-04T00:00:00Z

James Gleick reviewed The AI Con and points out that Turing didn't (try) to define intelligence, he defined how we can be fooled into not recognising machines. And it that's so:

The Turing test is done. We’ve proven that we can be fooled.

(From: The Parrot in the Machine | James Gleick | The New York Review of Books)

Originally posted on 4 July 2025 on Hidde's blog.

Reply via email

What good means

2025-07-12T00:00:00Z

At the UN's AI for Good conference, Abeba Birhane was forced to remove any mention of Gaza and genocide.

She explains what happened in her post, and observes what trend this is an example of:

what “good” means is overwhelmingly shaped, defined, and actively curated by the tech industry that inherently holds invested interest in societal uptake of AI regardless of its risk or harm

(From: AI for Good [Appearance?] | AI Accountability Lab)

Originally posted on 12 July 2025 on Hidde's blog.

Reply via email

Reasonable

2025-07-15T00:00:00Z

Vales says we should be more reasonable in critiqueing or hyping up AI:

Let’s strive for informed discussion, respectful critique, and a focus on addressing the real issues surrounding AI rather than going at each other’s throats.

(From: Advising Reasonable AI Criticism | Vale.Rocks)

I do agree with this and try to be reasonable, even if I mostly am a sceptic with regards to the majority of applications of the technology.

Originally posted on 15 July 2025 on Hidde's blog.

Reply via email

Billonaires and hearing mostly yes

2025-07-17T00:00:00Z

tante makes a good point regarding billionaires and LLMs:

I keep thinking about how super wealth makes you lose getting to hear no like a normal person when I see multi-millionaires and billionaires talk about how they use LLMs.

(From: Never hearing No)

Originally posted on 17 July 2025 on Hidde's blog.

Reply via email

How AI features are pushed

2025-07-24T00:00:00Z

Researchers Nolwenn Maudet, Anaëlle Beignon and Thomas Thibault looked at how tech companies are forcing AI on us.

They can deploy it faster than before:

Unlike previous technological evolutions, companies can now, through updates and softwares hosted in the cloud, transform interfaces overnight, affecting all users in an instant.

(From: How tech companies are pushing us to use AI | Limites numériques)

Their guide, based on their research (linked in the guide), shows how most of these tools are put front and center of UIs, are always on and hard to disable, and are presented as magic or a smart assistant.

Originally posted on 24 July 2025 on Hidde's blog.

Reply via email

Friction

2025-07-30T00:00:00Z

Tante nails it again:

The idea of frictionlessness has very narcissistic, “player character” vibes: You don’t experience friction if the whole world is build around you and your needs. When you get whatever you want when you want it.

This is about recognising value:

Because being actually touched, being inconvenienced, being emotionally moved, having your mind and perception changed means acknowledging your fellow human beings around you, realizing their differences to you and to recognize their value.

(From: Friction and not being touched)

Originally posted on 30 July 2025 on Hidde's blog.

Reply via email

Critical AI

2025-08-05T00:00:00Z

Maggie sees a use for AI as a possible intellectual tool:

I think we’ve barely scratched the surface of AI as intellectual partner and tool for thought. Neither the prompts, nor the model, nor the current interfaces – generic or tailored – enable it well. (…) It’s a problem I need to work on in some form.

(From: A Treatise on AI Chatbots Undermining the Enlightenment)

Originally posted on 5 August 2025 on Hidde's blog.

Reply via email

Glorifying ignorance

2025-08-28T00:00:00Z

I don't think I would self describe as an AI hater, though there are a lot of things I hate about what we make computers do today. For all the reasons Anthony Moser lists in his post. Yet I am also careful, because I don't want to absolutely close the door to everything. Recently, to give one example, I saw multiple friends with disabilities use AI in meaningful ways and I don't want to discount that, who am I to hate on the technology that is useful to them in ways previous technology was not?

Among other points, he explains very clearly how creatives and AI companies are different:

Miyazaki tells stories that blend the ordinary and the fantastic in ways people find deeply meaningful. Altman tells lies for money.

(From: I Am An AI Hater | moser’s frame shop)

Originally posted on 28 August 2025 on Hidde's blog.

Reply via email

Teach humans instead

2025-09-15T00:00:00Z

Craig Abbott about why we can't solve screenreader problems with AIs:

Empowering humans to communicate in a more accessible way prevents the problem at its source, avoiding the need to rely on expensive, unreliable machines to try and patch it in real-time afterwards.

(From: Screen readers do not need saved by AI - craigabbott.co.uk)

Originally posted on 15 September 2025 on Hidde's blog.

Reply via email

AI in dating apps

2025-10-12T00:00:00Z

AI dehumanises us, part 82: dating apps:

“I’d already been ChatGPT-ed into bed at least once. I didn’t want it to happen again.”

(From: ‘I realised I’d been ChatGPT-ed into bed’: how ‘Chatfishing’ made finding love on dating apps even weirder | Dating | The Guardian)

Originally posted on 12 October 2025 on Hidde's blog.

Reply via email

Friction matters

2025-10-16T00:00:00Z

Jenny:

The problem with AI “art” is that it was not the expression of a mortal being choosing to spend its one wild and precious life clawing its way through mediocrity to try and imperfectly communicate a feeling with other mortal beings who, by definition, can never fully comprehend it, and therefore it is fundamentally uninteresting to me.

(From: choosing friction | The Roof is on Phire)

Originally posted on 16 October 2025 on Hidde's blog.

Reply via email

250

2025-10-23T00:00:00Z

Researchers from Anthropic, the UK AI Security Institute and the Alan Turing institute, found it is easier than thought to poison very large models:

we demonstrate that by injecting just 250 malicious documents into pretraining data, adversaries can successfully backdoor LLMs ranging from 600M to 13B parameters

(From: A small number of samples can poison LLMs of any size \ Anthropic)

Originally posted on 23 October 2025 on Hidde's blog.

Reply via email

Notes from Green IO

2025-12-14T00:00:00Z

Chris Adams went to Green IO and wrote up his notes

I figure it’s worth sharing a few takeaways from sifting through about a bajillion pics of slides, and all notes scribbled down over the last three days, for others and my future self. Off we go.

(From: Takeaways, trends and notes from Green IO Paris 2025 - Reads, Takes and Links)

Originally posted on 14 December 2025 on Hidde's blog.

Reply via email

A machine that gentrifies the English language

2026-01-02T00:00:00Z

Robert Kingett was in a writing group when he realised what LLM vendors do is to gentrify our imagination and our language:

So the Tech Bros, in their infinite mediocrity, decided to bypass the human element entirely. They built a machine that scrapes our work—our pain, our joy, our very souls—without consent, grinds it into a mathematical slurry, and extrudes it as a flavorless, inoffensive paste that can be sold by the bucket.

They built a machine to gentrify the English language.

And the horror of watching my friend lose his soul almost eats me alive.

(From: The Colonization of Confidence., Sightless Scribbles)

(via Matthias Ott)

Originally posted on 2 January 2026 on Hidde's blog.

Reply via email

Opinions about Garamond

2026-01-03T00:00:00Z

When you're not into ITC Garamond:

The most distinctive element of [ITC Garamond] is its enormous lower-case x-height. In theory this improves its legibilty, but only in the same way that dog poop’s creamy consistency in theory should make it more edible.

(From: I Hate ITC Garamond - DesignObserver)

(via Eric Bailey)

Originally posted on 3 January 2026 on Hidde's blog.

Reply via email

Scale of energy and water use

2026-01-21T00:00:00Z

Simon P. Couch wrote about energy consumption of LLM usage and shares some of his napkin calculations (in lieu of data being made available my large AI vendors).

He concludes that it isn't too bad:

Personally, I don’t know that this scale of energy (and, ostensibly, water) use is significant enough to make me decrease my use of coding agents

(From: Electricity use of AI coding agents | Simon P. Couch – Simon P. Couch)

Originally posted on 21 January 2026 on Hidde's blog.

Reply via email

Friction and creativity

2026-03-21T00:00:00Z

Dave wrote a great piece about friction, and Matthias pointed out something rather important that followed from this:

The frictionless version of creative work isn’t faster creative work. It’s no creative work at all.

(From: The Shape of Friction · Matthias Ott)

I could not agree more. Bringing in experience and judgment (and, imo, intentions) is what makes works meaningfully creative, removing that leaves us with very little left.

Originally posted on 21 March 2026 on Hidde's blog.

Reply via email

LLM thoughts

2026-03-31T00:00:00Z

My opinion on "Artificial Intelligence" | ./axel.leroy.sh) gives an overview of developer's Axel Leroy's thought process on the question whether to use LLMs at work.

Originally posted on 31 March 2026 on Hidde's blog.

Reply via email

LLMs in the standards process

2026-03-31T00:00:00Z

As part of my work for the W3C's Advisory Board (AB), I co-wrote a short overview of reasons why or why not to use LLMs in the standards process.

My own personal opinions are a bit more extreme than what is captured here, and that's fine, the goal of this document is to capture what the AB mostly agrees on when it comes to usage of these tools. We wanted to be balanced, too, which is why a sceptic (me) and an enthusiast (Elena) decided to collaborate on it.

As we wrote:

we want to highlight considerations around different ways in which LLMs can be useful or problematic when it comes to leveraging them in standards work at W3C.

(From: Use of Large Language Models in Standards Work)

Originally posted on 31 March 2026 on Hidde's blog.

Reply via email

Not a substitute

2026-04-16T00:00:00Z

In another great piece about what LLM usage actually means in the workplace beyond promise of ‘efficiency’ and ‘profit’, Dave Rupert notes:

I’d rather have a human-to-human conversation with you, not a chat with Claude by proxy. What Claude said is an okay chunk of “anecdata”, but it’s not a substitute for our working relationship.

(From: I don't want a screenshot of your Claude conversation - daverupert.com)

Originally posted on 16 April 2026 on Hidde's blog.

Reply via email