Links and listens

The narrative around sustainability has too often been dominated by compliance costs and regulatory burden. But this framing misses the bigger picture. The most successful companies aren’t just ticking regulatory boxes; they’re using sustainability as a lens to unlock operational efficiencies, reduce costs, and create entirely new revenue streams.

As always, a big difference between AI systems and a human is that an AI system cannot be held accountable for its actions. I’ll let my trusted friend use my logged-in browser only because there are social consequences if they abuse that trust!

We all knew that Stripe’s founder Patrick Collison preferred interfaces that were densely packed and full of efficient shortcuts and keyboard commands; teams would bias their work towards Patrick’s preferences to ensure a smoother review and approval process. We weren’t solving user problems. We were solving for Patrick’s aesthetic sensibility. At one point, Patrick sent out a memo imploring teams to stop this preemptive bias, but it only added another recursive layer to the feedback process: it had to meet Patrick’s high expectations, but steer clear of sicophancy.

There's a striking lack of historical perspective in how we talk about tech today. Let community voices lead instead of a tiny group of tycoons, and you'd get much more interesting, accurate stories.

Thinking of AI as an instrument recenters the focus on practice. Instruments require a performance that relies on technique—the horn makes the sound, but how and what you blow into it matters; the drum machine keeps time and plays the samples, but what you sample and how you swing on top of it becomes your signature.

This is the first rule of strategy: strategy is contextual. A crucial insight, because often when leaders fail, it’s because they tried to apply a strategy that worked in one context, to a different one, without considering the difference.

The risk there is that the public — or indeed anyone not deeply versed in these nuances — will not understand the difference. That’s why Brander, the Edinburgh professor, argues that regardless of how it all shakes out, the Greenhouse Gas Protocol itself needs to provide more explicit guidance on what these numbers mean and how companies are allowed to talk about them.

The artificial intelligence companies have already violated the expectations of even a public web. Regardless of the benefits they have created — and I do believe there are benefits to these technologies — they have behaved unethically. Defensive action is the only control a publisher can assume right now.

At some point, you realize that simplicity is not the absence of sophistication as Da Vinci said. It’s the evidence of mastery. Writing clear code, building transparent systems, communicating intent precisely. Because clarity compounds like everything else. One clear decision enables ten more.

The person worried about their community or country changing too quickly and the immigrant seeking a better life are both responding to forces larger than themselves.

Not many people talk about how or why RSS won the content syndication war because few people are aware that a war ever took place. Everyone was so fixated on the drama over RSS’s competing standards (Atom vs RSS 2.0) that they barely registered the rise and fall of the Information and Content Exchange (ICE) specification, which had been created, funded, and eventually abandoned by Microsoft, Adobe, CNET, and other household names.

Open source projects exist under a social contract between the maintainers who produce the software and the users who consume and who may contribute back to it. While maintainers do not owe anything to their users — to quote the popular MIT license “THE SOFTWARE IS PROVIDED ‘AS IS,’ WITHOUT WARRANTY OF ANY KIND” — most maintainers do try to meet an implied social contract of the minimum expectations of an open source software project.

That’s because there’s one key feature that remains unique to human staff: accountability. A human can take responsibility for its action and learn from its mistakes. Putting an AI agent on a performance improvement plan makes no sense at all!

Without shared standards and life cycle-inclusive reporting, efficiency gains risk being cosmetic. With them, environmental performance could become as visible and competitive a metric as accuracy or speed — aligning AI’s development with global climate goals.

It is not easy to do the simplest thing that could possibly work. When you’re looking at a problem, the first few solutions that come to mind are unlikely to be the simplest ones. Figuring out the simplest solution requires considering many different approaches. In other words, it requires doing engineering.

We have learned to scan UIs like robots because the folks who designed them don’t care for the words as much as they cared for the fonts or the colors.

But in the last year, the balance of presumption has shifted. The signal-to-noise ratio has degraded, and the unsolicited PR is now more likely to be a high-effort review of a low-effort contribution.

In summary:
• Use “your” when communicating to the user
• Use “my” when the user is communicating to us

I really enjoyed this Freakonomics Radio episode which started out exploring why candles still exist, but then side tracked into explorations of obsolence, energy demand, and human nature.

The biggest example that people discuss when they discuss obsolescence in energy infrastructure is whale oil. People say, you know, fossil fuels came around and they saved the whales. This has become such a strong talking point for thinking about energy transition to say that renewable energies could come around and replace fossil fuels. But when you actually look at the history of whale oil, you see that we start using oil in the mid-19th century. Fossil fuels become more and more pervasive into the 20th, 21st century. And whale oil, the killing of the whales actually intensified the most in the 1960s. The introduction of fossil fuels actually led to the killing of more whales. Why was it? Because fossil fuel powered ships could go around the world and kill whales in a much more effective way.

When a person runs into a problem, they are able to temporarily stash the full context, focus on resolving the issue, and then pop their mental stack to get back to the problem in hand. They are also able to zoom out and focus on the big picture, allowing the details to temporarily disappear, diving into small pieces as necessary. We don't just keep adding more words to our context window, because it would drive us mad.

My argument: If so many developers are so extraordinarily productive using these tools, where is the flood of shovelware? We should be seeing apps of all shapes and sizes, video games, new websites, mobile apps, software-as-a-service apps — we should be drowning in choice. We should be in the middle of an indie software revolution. We should be seeing 10,000 Tetris clones on Steam.

The consensus from leaders is that vibe coding puts critical software qualities at risk: security, clarity, maintainability, and team knowledge.

The participants in this study showed significant agreement on the importance of carbon-aware websites and a willingness to incorporate carbon emissions into their quality of experience. However, it also showed little agreement on the specific website changes they would like to see to implement lower emissions websites.

Reflect on the state of mind of a player who is said to be “hot” or “playing in the zone.” Is he thinking about how he should hit each shot? Is he thinking at all? Listen to the phrases commonly used to describe a player at his best: “He’s out of his mind”; “He’s playing over his head”; “He’s unconscious”; “He doesn’t know what he’s doing.” The common factor in each of these descriptions is that some part of the mind is not so active. Athletes in most sports use similar phrases, and the best of them know that their peak performance never comes when they’re thinking about it.

Clearly, to play unconsciously does not mean to play without consciousness. That would be quite difficult! In fact, someone playing “out of his mind” is more aware of the ball, the court and, when necessary, his opponent. But he is not aware of giving himself a lot of instructions, thinking about how to hit the ball, how to correct past mistakes or how to repeat what he just did. He is conscious, but not thinking, not over-trying.

Our findings demonstrated that the length of the response generated by the LLM is a major driver in energy consumption. Experiments suggest that response correctness does not significantly influence the energy consumption.

Open source, the thing that drives the world, the thing Harvard says has an economic value of 8.8 trillion dollars (also a big number). Most of it is one person. And I can promise you not one of those single person projects have the proper amount of resources they need. If you want to talk about possible risks to your supply chain, a single maintainer that’s grossly underpaid and overworked. That’s the risk. The country they are from is irrelevant.

Our world is a collaborative endeavor. (Competition is Darwin-inspired and, together with property and scarcity, capitalism-affirming BS.) Collaboration requires trust. Trust requires truth—and truth requires trust. Just making and aggressively pushing on claims, like lobbyists and political extremists do, is untruthful and therefore undermines trust. Without trust and without truth, there’s ultimately no collaboration—and, ultimately, no world.

Use Energy Intensity (kWh/GB) as a starting point to raise awareness, prioritize efficiency, and track progress. Use the Power Model if you’re doing advanced research or want to model real-world network behavior more precisely.

Think of energy intensity as a compass — not a precise GPS, but enough to steer in the right direction.

Environmental impacts from “AI” are often derided as being based off improbable forecasts of the future. But you can find any number of material, real and immediate impacts from the data centre boom happening right now. This is tactile stuff, and again, not what you would expect if these companies were delivering inert, energy-efficient services that ‘don’t harm the environment’.

Tech companies have recognised how they need to evolve control of the public narrative around the products they provide, and so they’re successfully shifting the conversation towards consumer decisions rather than their own decisions.

“I’m sure as the tech matures and we adopt specific tools organisation-wide, those discussions may happen. Right now we are still piloting different tools and figuring out what does and doesn’t work for the organisation.”

Right now, LLMs are an extremely immature technology. I personally believe they’re not going to get much better than this, but a breakthrough innovation could change that. Either way, it doesn’t matter to me. If the technology is a fad and completely implodes, I couldn’t care less. If it’s not a bubble and LLMs actually turn out to be the new best thing, I can easily adopt them into my own business model. ... So when I see people jumping on the latest hype, telling me I’m going to get left behind, I can only chuckle. If LLMs as a technology are viable, they’ll still be around when and if I decide they’re useful for me. If not, I’ll have missed out on losing my life’s savings in Beanie Babies, The DotCom bubble, or NFTs.

It’s not even necessarily that corporate executives are being stupid. Sometimes they are, which can result in things like sinking more money that it costs the US government to put the sun in a bomb into the worse VR game ever. But usually it’s just greed and shortsightedness.

AI right now feels much the same. An industry fueled by the gluttony of myopic visionaries. An industry grasping at every straw to find a use case for their technology. An industry built on the premise of hyperbole and empty promises. But in this specific case, I actually don’t think big tech companies are making the wrong decision, at least, considering the choices they have available.

Even when the law explicitly says sharing this data is mandatory, we saw around a 36% participation rate from companies that own and operate these data centres in Europe. We know of no consequences for not reporting at present.

We bring tools in to reduce cognitive load so we can “focus on higher-level work.” Good intention. Wrong model. Abilities do not sit in cold storage while a tool handles them. They weaken with disuse. The less you do a thing, the harder it becomes to do that thing, and the faster your confidence plummets.

The fear is not that AI becomes sentient. The fear is that humans become inert.

Feminising job titles, in my opinion, belittles success and achievement in the context of a patriarchal default. Just as “Women in Tech” awards and closed spaces for “non-men” segregate underrepresented groups of people from the spaces where people with power make decisions, women-coded labels inherently classify people on a separate, non-default scale of success and achievement.

"More powerful" never seems to mean "does more," and "more powerful" oftentimes means "more expensive," meaning that you've just made something that doesn't do more but does cost more to run.

Deep Research has the same problem as every other generative AI product. These models don't know anything, and thus everything they do — even "reading" and "browsing" the web — is limited by their training data and probabilistic models that can say "this is an article about a subject" and posit their relevance, but not truly understand their contents. Deep Research repeatedly citing SEO-bait as a primary source proves that these models, even when grinding their gears as hard as humanely possible, are exceedingly mediocre, deeply untrustworthy, and ultimately useless.

These men are constantly lying as a means of sustaining hype, never actually discussing the products they sell in the year 2025, because then they'd have to say "what if a chatbot, a thing you already have, was more expensive?"

If there's one type of mind I most desperately want that view into, it's that of an AI. So, it's in this spirit that I ask: what does the Earth look like to a large language model?

Some new participants in standards arrive with the expectation that the de jure nature of a formal standard creates a requirement for implementation, but nothing could be further from fact. This sometimes leads to great frustration; enshrining a design in a ratified standard does not obligate anyone to do anything, and many volumes of pure specifiction have been issued over the decades to little effect.

The voluntary nature of web standards is based on the autonomy of browsers, servers, and web developers to implement whatever they please under own brand.

Imagine a world where your library card includes 100GB of encrypted file storage, photo-sharing and document collaboration tools, and media streaming services — all for free. Your data is encrypted end-to-end, but is shareable to anyone on any other service through standardized protocols.

The investors were done with us, i.e. they weren’t interested in helping to climb out of the hole. That’s normal for them. They expect some investments to fail and when that happens they walk away. We were one of their fails.

The team, very impractically, wanted to climb out of the hole. I think there is something about the Medium ethos that motivates the staff here even when things are grim. And I’m not even done saying how grim.

Bottom line: Feedback should be thoughtful, specific, grounded, and mutual. If you’re not doing it with intention, you’re not doing it right.

Because the truth is, feedback can land hard. Even when it’s accurate. Even when it’s kind. Even when it’s necessary.

And yet, it’s still one of the clearest forms of care we have. Especially when it’s offered with intention, received with curiosity, and held in an environment of trust. Feedback is not an attack. It’s a mirror. And the point isn’t to feel ashamed of what you see, it’s to adjust the lighting, stay open, and keep building.

I was thinking this morning about how once you understand that your technology choices have security, performance, and accessibility considerations you become a much more boring developer. Acknowledging those obligations can sort of strips the fun out of programming, but we’re better for it.

In the Additive model, new energy resources greatly slow and occasionally halt the growth of legacy energy sources, eventually coming to dominate all growth. Yet the legacy sources—because they are embedded or aligned with legacy infrastructure—maintain their toehold long enough to eventually grow again, if even just a little.

For the longest time, writing was more expensive than reading. If you encountered a body of written text, you could be sure that at the very least, a human spent some time writing it down. The text used to have an innate proof-of-thought, a basic token of humanity.

Now, AI has made text very, very, very cheap. Not only text, in fact. Code, images, video. All kinds of media. We can't rely on proof-of-thought anymore. Any text can be AI slop. If you read it, you're injured in this war. You engaged and replied – you're as good as dead. The dead internet is not just dead it's poisoned. So what do we do?

More subtly, it feels like our own aliveness is what’s at stake when we’re urged to get better at prompting LLMs to provide the most useful responses. Maybe that’s a necessary modern skill; but still, the fact is that we’re being asked to think less like ourselves and more like our tools.

Shifting away from measuring reality towards estimating ‘how well you avoided a worse possibility’ could end up with massive underreporting of emissions - and therefore, less impetus to act to reduce them.

Plenty of vendors will sell you “guardrail” products that claim to be able to detect and prevent these attacks. I am deeply suspicious of these: If you look closely they’ll almost always carry confident claims that they capture “95% of attacks” or similar... but in web application security 95% is very much a failing grade.

If you’re not already a web developer though, it is a relatively expensive intervention – hiring a web developer to make your site more sustainable through front OR back end performance changes relies on you being able to:

1. dedicate loads of time to learning all the techniques to become one, or
2. hire a web developer with the required set of skills to apply them to your site, who are hard to find, and generally more expensive than developers who haven’t specialised like this

Both of these are hard to justify with their short term returns vs the other things you’re expected to do when sustainability isn’t your main job.

In a saturated market, where perspectives of growth are know quite weak, new AI features become very often an opportunity to increase here and there the subscriptions prices of several digital services. We, users, have then to pay the price of the financial risk taken by companies in this AI race. And then more or less visible changes that overwhelmed our interfaces are the direct consequence.

How would your design choices differ if you refused to display more than two user actions within the same view at the same time? How would you group and differentiate content if you disallowed the use of borders and shadows?

Far from being exercises in austerity, these limitations invite resourcefulness. They force a deeper reckoning with the fundamentals of structure, rhythm, and intention. They encourage inventiveness over rote repetition of existing patterns and standard practices.

We need to regain familiarity with our inner world. To develop interoception, or an ability to slice and dice apart the feelings we all experience, instead of trying to push them away or avoid them. Pick something you do regularly and do it without external input. It doesn’t have to be full-blown mindfulness meditation. It could be every time you wash the dishes. Or leaving your phone at home when you go on a walk or run. Or putting your phone in the glove box when you commute to work. Or not pulling out your phone when at a restaurant and your dining partner gets up to use the restroom. These bite-size moments are great training, and very important. They remind your brain that you don’t have to fill every second of nothingness with stimulation. You don’t need to outsource your brain’s attention and entertainment.

Search Engine is fast becoming one of my favourite podcasts. In this episode, PJ Vogt explores the use of AI in the education system and the attempts that have been made to regulate it.

I've seen people use image generators in amazing ways, but the reason they are able to use them in amazing ways is because they can talk to the image generator like an art director. Not like a prompt engineer, but like an art director. And these are the things that we need to be teaching people to use technology that can generate images or syntax or video or music or what have you. Not, how do you interact with this thing on a moment-to-moment basis to get an output? Because that just puts ... everything behind a veil. It's like the great Oz is back there doing stuff. We should know exactly what Oz is up to when we are asking it to do things.

Greg Storey recaps his time at IBM, and how messy data was/is such a big blocker for many companies ability to automate processes. The commentary on the Google Superbowl commercial made me lol.

If Google’s AI can’t even fact-check the popularity of a cheese, how the hell is it supposed to take over someone’s job? What is the value of having an assistant that does an amazing job 60% of the time—every time? And how is it going to do good work if it can’t find the data because it’s still on spreadsheets on a laptop somewhere?

Line doesn't always have to go up, and growth doesn't always have to mean "more".

If you build a nice, sustainable fire, it’ll keep you warm, cook food and sustain life. And if the only thing you care about is how big your fire is, then it’ll set fire to everything around it, and the more you throw into it, the more it’ll burn. Eventually, you’ll have nothing left, but if you desperately desire that fire, you will constantly have to find new things to burn at any cost.

An article earlier this month by Robin Sloan has prompted a few bloggers to voice their opinions about the merits of LLMs, and the merits of the arguments Robin made in their article too. Michelle makes the point that LLMs have limited utility, mostly for processing texts and do not create new knowledge.

Both of these applications are interesting and potentially useful. But they are not the same. An LLM as described above, while useful, shouldn’t invent new information. It processes the text that already exists, not the science behind it, and if it appears to offer up something new then that should be met with the utmost scrutiny. And it remains to be seen whether they (and others like them) will be worth the extraordinary amount of energy and resources that AI demands.

I came across this piece as a side effect of some internet hot drama that I was passively watching from the sidelines. I like distinctions made for when we might choose to call something/someone out versus when it could make more sense to call them in.

According to Thom (2015), calling out is when we publicly name the harm, often responding with strong emotions like anger, drawing the attention of others to the problem. Calling in is when we privately respond to the person and gently explain why their behaviour needs to change. Both of these strategies can be effective in holding people accountable, but it is important to consider what is happening in the situation and what result you would like to achieve.

Honeypot make damn good developer documentaries. This one about Angular is no exception.

A really enjoyable interview pod from The Changelog, where Jerod and Adam chatted with Bert Hubert about building software that lasts. This is super important in todays age of "npm installing away our problems". Bert covers the expected stuff, like keeping projects simple and reducing complication in code. But he also mentions other things, like walking a mile in the shoes of your support team so you can feel the pain of the things you might put out into the world.

... we all know that we should not write software that is as complicated as we can make it because that's not going to end well. But everyone said, look, we had very bad experience. Like we need to figure out seven years ago why this clever code, what it actually does and then often you find that this clever code there was no need to make it clever. ... even if you have 25 [political parties], there is no need to set up a complicated data structure to hold 25 political affiliation names. But one day you might sit there and say, hey, wouldn't it be useful? Wouldn't it be fun if we had a complicated red-black tree so that we could make really rapid searches of our 25 political parties. And, and then you sit there and maybe in 2032 trying to debug why it doesn't work. And, and that's because you try to be clever.

Bert Hubert

Not a long read, but rather a handy tool shared by Chris over on the Frontend Masters blog.

Om Malik writes about the growth of fixed connection broadband availability around the globe, which now reaches 1.5 billion people. He also looks at what lies ahead with caution as the next billion people connect with fast fixed broadband connections.

Looking ahead, the challenge isn’t simply about reaching the next billion connections. The world faces a more complex problem: preventing a new form of digital bifurcation where advanced markets pull even further ahead. If artificial intelligence (AI) becomes a core part of all digital experiences, the lack of connectivity will leave a large swath of humanity behind, simply because they lack access to better digital tools.

Found myself nodding along to this short post about the uneven distribution of AI's benefits and "value creation". It's timely, especially with the AI Summit happening in Paris this week, and the release of a Joint Statement on Limiting AI’s Environmental Impact which we (Green Web Foundation) and 100+ other organisations are signatories to.

The two groups of actors that truly benefit from generative AI so far are 1) a dozen or so of big tech companies and 2) to a smaller degree, employers who might get some small efficiency gain out of their employees.

It’s the opposite of the (long debunked) Trickle Down Economics of yesteryear: The financial value bubbles to the very top, and only trace amounts of the value capture happens anywhere below.

This Freakonomics episode features a conversation with Ben Zhao, whose research lab has been doing some pretty pioneering work to prevent AI misuse and protect the rights of content creators from having their work eaten up by AI training models. One approach that's covered in this episode is the idea of "poisoning" the material on which the AI is trained, so that when it's asked to reproduce something similar the AI application consistently ends up generating something that's totally incorrect.

Art is interesting when it has intention, when there's meaning and context. So when AI tries to replace that, it has no context and meaning. Art replicated by AI, generally speaking, loses the point. It is not about automation. I think that it is a mistaken analogy that people will oftentimes bring up. They say, well you know what about the horse and buggy and the automobile? No, this is actually no about that at all. AI does not reproduce human art at a faster rate. What AI does is takes past samples of human art, shakes it in a kaleidoscope, and gives you a mixture of what has already existed before.

Here is Simon Willison recaps a conversation on the Real Python podcast, in which he extols the virtues of open source code.

I realized that one of the best things about open source software is that you can solve a problem once and then you can slap an open source license on that solution and you will never have to solve that problem ever again, no matter who’s employing you in the future.

On a related note, I also read this during the week which answered a question I've also asked many a time - Is “Open Source” ever hyphenated?

As always, Harry Roberts nails it. Rather than reaching for a framework for each new project, redesign, or update, he urges development teams to use the platform. Not only with a view to better web performance, but with a view to their longer term sanity as well.

If I was only able to give one bit of advice to any company: iterate quickly on a slow-moving platform.

Shayle speaks with Mike Grunwald about the challenges of industrial agriculture, how we can increase yields per acre to continue feeding the world without using up more land for agriculture, and other potential solutions. This was a fascinating conversation about a topic that I'm not too familiar with, but which touches each and every one of us.

If you can grow beef without growing the cow ... you can imagine how it could be a really efficient process. Now, right now, they're basically using pharma-grade equipment to make food. And, you know, it's the economics of growing somebody a new heart are a lot, you know, easier than the economics of growing lunch.

Max Böck has a dig at the different AI "enhancements" users are having aggressively shoved down their throats. Often they're being added to products and services that the users are already used to using, and the users are being forced to opt-out (and sometimes that's even hard to do). More often than not, the quality of these AI "enhancements" detract from what once was a useful product.

I get that it’s not as fun to build “a faster horse”. To just make the thing you already have better, more reliable, more helpful. It doesn’t get your shareholders excited, and it doesn’t make you look like a visionary genius.

Jeremy Keith explains how his views towards LMs are shifting especially in relation to their environmental impacts. He also floats an idea that I've seen a few others talking about too, the idea of charging a fair price for LM useage based on their actual energy and/or water usage.

Personally, I’d just like to see these tools charge a fair price for their usage. Right now they’re being subsidised by venture capital. If people actually had to pay out of pocket for the energy used per query, we’d get a much better idea of how valuable these tools actually are to people.

Laurie Voss reflects on his experiences with AI, where it shines, and where it falls short of expectations. I vibe with his sentiments that Language Models (LMs) at the moment are suited to tasks are based on summarising text. I've found that's increasingly all I'm using them for.

There is no way to get an LLM to perform the thought necessary to write something for you. You have to do the thinking. To get an LLM to write something good you have to give it a prompt so long you might as well have just written the thing yourself.

I really enjoyed listening to Adam and Jerod chat with Elecia White on the bus to my part-time Touch coaching gig this week. I've tinkered with home automation before, but never got my hands dirty in embedded systems. It's one of those things that I feel I'd like to try out if I have a spare week of nothingness on my hands. I love how she talks about falling in love with software and embedded systems in this clip

... the first time I moved a motor, because I told it to, not just because I powered on something, but the first time it was under my control, it was magical. It was like software can touch the world. Software can change physical things. It's not just zeros and ones.

My colleague, Chris, shared this idea of "git scraping" to me around a small side-project I'm building at the Green Web Foundation. I vibe with the idea, and reckon it'd be a pretty neat way to track changes in a dataset over time. It's stored away in the back of my mind & hopefully I'll have something to use it on later this year.

Thibaud wrote this reflection on his involvement in our Grid-aware Websites project. I'm so glad to be able to lean on the guidance of our advisory group for this project. As as developer, I'm just like write some code, make some cool demos, and success! But having the advisory group to bounce ideas off, and get feedback from, opens our eyes to other things we should be considering - like the business case for making grid-aware changes to a website. Also, I really like this line from Thibaud, and am hoping to work it as a way to frame the Grid-aware Websites project.

Websites, and the web platform generally, are excellently suited to adapt to user needs (color themes, font sizes, responsive, etc).

I'd never heard of pace layering until reading this post by Chris Coyer. It makes a lot of sense, thinking about it, and it aligns with the way I've seen things work in a lot of spaces.

I find this a helpful framework to think in sometimes. For instance, if you feel frustration at how quickly or slowly a particular technology moves, are you considering its place within the layers? Perhaps that speed is because it is part of a system that pressures it to be that way or it being that way is beneficial to the system as a whole. Even zoomed into browser technology, HTML not moving as fast as JavaScript feels like it could be mapped onto Pace Layers.

Alex Russell with a banger of a conference talk in which he explores the advantages of the web, and the challenges it has faced from other platforms, especially mobile, and how we can/should be taking action to create a future where the free and open web wins.

This article about the history behind the concept of progressive enhancement on the web has really stuck with me. I've been thinking about progressive enhancement a fair bit in the context of making this website grid aware, and our overall Grid-aware Websites project. At the moment, I "degrade" functionality on the site by removing certain bits of code whenever a user is visiting from a region where the electricity generated is coming from mostly fossil-fuels. But this post, and some of the feedback I've received about this site has got me thinking *how might I go about "progressively enhancing" the site instead for visitors who are on cleaner power grids and so can afford to take on the additional compute of some features?"

In the last few weeks, I've read a lot of year in review posts. I even wrote on myself. This one by Simon Willison is pretty much a must read for anyone who wants to be updated on what's happening in the realm of AI and LLMs. He covers a lot of ground, including making several points throughout about the nuance with which we should be thinking about AI, LLMs, and their environmental impacts.

I think telling people that this whole field is environmentally catastrophic plagiarism machines that constantly make things up is doing those people a disservice, no matter how much truth that represents. There is genuine value to be had here, but getting to that value is unintuitive and needs guidance.

Simon Willison

This post by Hannah Smith, my colleague at the Green Web Foundation, captures our sentiments about this week's drama in the Kingdom of WordPress Matt Mullenweg. The WordPress sustainability group was suddenly dissolved by Matt Mullenweg, apparently shortly after he learnt WordPress even had a group of volunteers working on sustainability in the WordPress community.