blog

The Slack DevEx team reduced execution time of their end-to-end testing pipeline by reusing frontend assets when no changes where made. A seemingly simple change that's undoubtedly complex with a large monorepo being worked on by hundreds, maybe thousands, of developers simultaneously and an ecosystem of integrated tools.

Identifying changes to frontend files was done efficiently by git diff and finding the last frontend build was done using "straightforward S3 storage concepts". The latter was never elaborated on.

With hundreds of PRs merged into this repository daily, identifying a prebuilt version that was fresh enough required robust asset management. By using straightforward S3 storage concepts, we were able to balance recency, coherent file naming, and performance to manage our assets.

This leaves me to speculate. - Is the latest build always stored with the same prefix? - Is the prefix of the last frontend build recorded in a key-value store that can later be retrieved? - Is the last frontend build information retrieved from git and the prefix pieced together on retrieval? - Or is it something more complex? Probably.

Given the quote from the blog, it leads me to believe the prefix is made up of the date, with the commit hash in front for better prefix performance. It's all just a guess for now.

Microsoft used their fundamental advantage as an incumbent software leviathan to bypass the terms on which we wanted to fight them. With very few exceptions, they didn’t steal our customers. Our customers loved Slack. But they tapped the massive market they already had and greatly reduced our chance to ever reach them. A company that was already paying a steep premium for the essential tools of Office 365 got Teams “for free” and it seemed fine to them.

That's sounds like 🎩 to me. Winning an antitrust case in the EU in 2024—filed in 2020—seemed to be too little, too late.

Filed in 2020 and resolved in 2024, the suit alleged that Microsoft had used its monopoly power to bundle Teams into the Office suite and give it away for free, thereby undercutting competition and unfairly skewing the market in their favour. In response, Microsoft has agreed to unbundle and charge for Teams.

Jeff Bridgforth shares two pieces of advice that could greatly help your career. The first is "networking", something that doesn't come as naturally to me as described in Jeff Bridgforth's post. The second is maintaining a career management document that records your work over time.

While I felt the first point was over-indexed on at my university, I never came across the second until I worked for an employer that recommended logging accomplishments to highlight when it comes time for promotions. I wish I'd started this sooner in my career.

I've posed a similar question before which is often met with hesitation, or dismissal, but today, I came across this post circulating online, and it's great to see some discussion around it.

This is my favourite line.

Ad companies are never going to regulate themselves—it's like hoping for heroin dealers to write drug laws.

Ploum talks about a component of enshitification, Androidification.

Androidification is not about degrading the user experience. It’s about closing doors, removing special use cases, being less and less transparent. It’s about taking open source software and frog boiling it to a full closed proprietary state while killing all the competition in the process.

Brian on why 88x31 buttons are precisely 88x31. Discovered via misty.

Seth Larson talks about how about the irony that an AI busting services, like CloudFlare's AI Labyrinth, use generative AI to combate the web crawlers used to create generative AI models.

Seth then goes on to talk about how heavily subsidized the AI industry is and draws an interesting comparisons, one that most of us have lived through first have.

Today this subsidization is mostly done by venture capital who want to see the technology integrated into as many verticals as possible. The same strategy was used for Uber and WeWork where venture capital allowed those companies to undercut competition to have wider adoption and put competitors out of business.

Writing-as-sales-product is not what a people-focused Web is or ever will be about.

Discovered via Roy Tang.

Steve Jobs is often quoted about computers being the bicycle for the mind or something similar, as computers further enhance our present capabilities just like a bicycle uses our energy to propel us further. In a blog post that was shared on lobste.rs, Xe uses this quotes to introduce the breakthrough that Apple Intelligence would be, provided it works it course.

It's a quote that's been used over and over again in the last two decades but what I'm more interested in comes from the comments. Lobste.rs user minimax extends the bicycle analogy:

You can build a bicycle from spare parts, because the parts are standardized

You can learn to build and maintain many bicycles by simply observing the mechanisms

Cheap bikes work nearly as well as expensive ones for non-competitive riding

Bicycles don’t attempt to do your work for you, rather they channel and direct your energy

You need to learn to ride a bike; most able-bodied people can do it, but it’s not trivial

You need to learn additional skills, including some discretion, to ride safely in traffic

My takeaway? Computers need to be even more like bicycles.

Discovered via makeworld on Mastodon.

Rakhim's comic graphs the number of blog posts published against the number of posts about the blogs setup. I think I can see myself in there.

Discovered via bluelander.

Tom Usher provides some steps to curate your feeds and break away from algorithmic complacency.

Manuel Morale recently wrote about how personal should a personal site be. I for one don't get very personal on here or other forms of social media intentionally. On occasion I'll share an opinionated post tagged as status in my blog. Going through them, however, I can see they're still not very personal.

Manuel writes:

Every time I share something that’s a bit more personal, a few emails from strangers inevitably trickle in and they all share this sense of relief in knowing other people out there are struggling with the same problems they’re facing. And so I’m wondering if I should be doing it more. Maybe there’s value in being vulnerable on the web sometimes.

Reading that paragraph reminded me of an academic article which talked about the effects social media has on mental health, in particular Instagram. This is more of an extreme example as people tend to only share the best or best fabricated versions of themselves.

In particular, Instagram users who engage in digital status seeking (looking for popularity online) and social comparison (evaluating oneself in relation to others) tend to experience negative psychological outcomes. Such behaviors have been linked to increases in depressive symptoms, social anxiety, and body image concerns across age groups, as well as decreases in self-esteem (Sherlock, M., & Wagstaff, D. L., Psychology of Popular Media, Vol. 8, No. 4, 2019; Cohen, R., et al., Body Image, Vol. 23, 2017).

Working towards a curated feed can be beneficial or at the very least less harmful.

As those findings imply, Instagram users can attempt to curate their feeds to be less harmful, for instance by muting or unfollowing accounts that post idealized content and following those that promote diversity.

In one experiment led by Fardouly and Rachel Cohen, PhD, a clinical psychologist based in Australia, women who viewed “body positive” posts—which promote acceptance of diverse body types—reported improved mood, body satisfaction, and body appreciation (New Media & Society, Vol. 21, No. 7, 2019).

With that being said I don't believe person websites cause such harm, they're not designed to suck as much time with doom scrolling as engaging with them is more intentional. I do see the benefit of being able to relate or feel included when reading personal entries and there are many bloggers who do a good job in sharing just that.

Paweł Grzybek celebrates 10 years of blogging, congratulations! In his post, he addresses why writing is still important in the current AI-hype cycle. This related to a quote I shared from Michał's blog about writing only providing fuel to LLMs without much benefit to the writer. Paweł writes about the benefit provider to the writer using research supported by Microsoft.

At first glance, the situation does not look like the slow process of blogging is an idea worth pursuing. Precisely the opposite is true! Critical thinking required for writing (and other acts of creation) is the only thing that can save us from becoming idiots. Microsoft, the same one that made a pretty close partnership with OpenAI, funded the interesting research about “The Impact of Generative AI on Critical Thinking.”

The study mentioned surveys 319 knowledge workers. On the topic of writing in related works, relying on LLMs fo write can hinder self development, however using the tool to request feedback can help the writer improve. The overall conclusion to the study says

Analysing 936 real-world GenAI tool use examples our participants shared, we find that knowledge workers engage in critical thinking primarily to ensure the quality of their work, e.g. by verifying outputs against external sources. Moreover, while GenAI can improve worker efficiency, it can inhibit critical engagement with work and can potentially lead to long-term overreliance on the tool and diminished skill for independent problem-solving. Higher confidence in GenAI’s ability to perform a task is related to less critical thinking effort.

In August 1993, the project was canceled. A year of my work evaporated, my contract ended, and I was unemployed.

I was frustrated by all the wasted effort, so I decided to uncancel my small part of the project. I had been paid to do a job, and I wanted to finish it. My electronic badge still opened Apple's doors, so I just kept showing up.

An incredible story of how Ron Avitzur and Greg Robbins were able to develop and ship software that would be included in every computer Apple would produce while not actually being employed by them.

Discovered via Simon Willison.

I start to wonder if it is even ethical to publish on the web anymore. I won’t benefit from it, people won’t find it so they will also not benefit. But who will it help to become multi-billioner? Samuel fucking Altman. We are creating an all-he-can-eat buffet for him, while we won’t even get to eat the leftovers. And this makes me very sad.

We are actively partipaiting in our own destruction, because it’s fun. And while we never agreed to it, the result is still the same.

This quote is from Michał's blog and it's a bit grim but it did make me stop and think. Now I do believe there's personal and public benefit from a blog and I think the author believes this as well but still...

Recently I came across some discussion on Mastodon about archival services and physical mediums to use for long term storage. Seeing that M-Disc was mentioned I couldn't help but bring up Dr. Gough Lui's website. They have written high quality and in-depth reviews on mainly optical data storage mediums, like M-Disc, for many years now. I had linked reviews for the Verbatim Lifetime Archival (Millenniata/M-Disc) BD-R and the Ritek M-Disc DVD as examples.

Michael Miszczak writes about the effects of Google search over the years on smaller bloggers and publishers. First through snippets of information directly on the search page and now through AI summaries and down ranking smaller sites.

Google is often called a tech company, but that’s a misnomer. It might have been true a decade ago, but that label no longer applies to the Alphabet of today. What Google has actually become is the largest advertising company in the world. They feed you ads and make money that none of us can dream of making.

Discovered via Ana Rodrigues.

Clara and I have an AppleTV, a curious device made by a company that seemingly cares little for it. This is fantastic news from a large IT company in 2025, because it means it has mostly escaped from being stuffed with new “features” nobody wants.

Funny to think my favourite piece of tech is my old Kindle Paperwhite. The battery still works great, the power button hasn't hammed, I can load books from various stores, I can download books directly from the Kindle store if I choose, and it's not plagued with pop ups or other obnoxious attention grabbing features. The same can't be said with the Kindle store itself, but that may be a story for another time.

Juhis writes about how long it takes him to publish a blog post. In his answer he excludes time spent on activities such as consuming additional material like articles, podcasts and videos, taking notes, developing code, discussing the topic with others and more.

I don’t consider most of the activities listed above as something specific to blogging. Blogging is just a public outcome of things I do.

Whenever thinking about time I spend on post I always consider the preparation as well, in essence, all the activities Juhis lists. However, these are activities I would have done regardless of whether I'm writing about it publicly or not.

Maybe I need to reconsider how I think about time spent per post.

No matter how much I enjoy the work I do, or how great my team is, I remind myself that the company I work for is my employer and I'm nothing more than an employee. I signed a contract to do work in exchange for a salary and that's all I am within work hours.

Mert Bulan wrote a piece on employer trust, job security and what to do if you haven't been laid off.

When I looked back on my time at the company and all the things I had accomplished, I was surprised to be impacted by the layoffs. It wasn’t because I thought I was better than others—it was because I believed I was doing more than what was expected of me. However, during a layoff, it seems that who you are and what you do doesn’t matter. In most cases, the decision is made by people who don’t even know you.

Discovered via Arne's weekly post.

Earlier this month I linked to a post on Kottke where they talked about using hypertexts in blogs. Arun extends this in his blog with tasteful visuals that would not otherwise be possible in traditional media.

Blogs are native to the web, which affords a very different form of consumption than traditional print media. I love weaving in video, illustration, and animation into my otherwise static posts. For example, my post on the Apple Card extensively uses video clips that illustrate ideas much better than static images could.

Riccardo Mori surveyed a 106 participants on their backup practices using Windows and Macs, and includes some takeaways on changes from spinning disk to SSD.

I received 106 replies, 75 from Mac users. Of these Mac users, only 11 are still actively, routinely backing up their data. Of the remaining 64, 21 told me they’ve never backed up anything. In the remaining group of 43 users, a few of them relied solely on Time Machine backups (without even verifying them), but the majority was simply using some cloud service (Dropbox, iCloud, OneDrive, Google Drive) to save selected critical data and nothing else. [...] every person in my sample who was using an Apple Silicon Mac didn’t bother with any particular backup solution, and a lot of them specifically told me that they had stopped bothering with backups since Apple stopped including spinning hard drives in their computers, and especially since transitioning to the Apple Silicon architecture.

Interesting observation of moving to Apple Silicon and SSDs made users feel safer in their data to not bother with backups. I'd have expected most Mac users to use backup with Time Machine given how easy it is to use.

[...] of the 31 Windows users who submitted their replies, the vast majority used OneDrive as main backup solution, while 5 people told me they relied on local NAS solutions to preserve their data. Even among them, SSDs inside their main computers meant a general sense of increased reliability and security.

A similar observation on HDD and SSD with Windows users as well.

This is a post on Kottke, discovered through People & Blogs, about writing on the web and the ability to load a paragraph with links enabling information density. The example given quotes a paragraph from NY Times which contains only 110 words but has 27 links to other NY Times opinion pieces adding "receipts" to their point of view. This is a style of writing that is only possible on the web, something I hadn't fully realised before.

We saw all the memes about cubicle farms being soulless mazes, but frankly I’d love to be back in one. I was more productive in my first corporate job cubicle (and my home office) than any open-plan office since, and we’re all starting to understand the psychology why.

I knew I wasn't alone in liking cubicles over open office floors!

Professor Matt Might wrote about improving technical writing and created bash scripts to catch passive voice, weasel words and lexical illusions (repeated words). Simon Willison converted those scripts to a web tool. Even if you don't use the script or the tool the advice in Matt Might's post is great.

A write-up looking at the statistics on why (or why don't) hosts of the Olympics do well based on the athletes participating and medals won.

This post from Candost reflects on traits of a good software engineer. It's very well thought out.

Not sure why I'm surprised but there are many web formats in which to write recipes. There's a microformat, why wouldn't there be? There's another format that Google prefers and Robb takes stab at designing his own.

Alex writes about creating HTML files in folders to browse files in various ways.

This allowed me to radically simplify the folder structure, and stop chasing the perfect hierarchy. In these mini-websites, I use very basic folders – files are either grouped by year or by first letter of their filename. I only look at the folders when I’m adding new files, and never for browsing. When I’m looking for files, I always use the website. The website can use keyword tags to let me find files in multiple ways, and abstract away the details of the underlying folders.

Why HTML?

I’m deliberately going low-scale, low-tech. There’s no web server, no build system, no dependencies, and no JavaScript frameworks. I’m writing everything by hand, which is very manageable for small projects. Each website is a few hundred lines of code at most.

It reminded me of a time when I had to present a project I'd worked on as part of a job interview. I created a number of linked HTML files that walked through various parts of the code. It had the benefit that if a question was asked I could directly open up the relevant file and dive into the code. The interviewers didn't seem impressed but I ended up getting the job, so I guess it worked.

Mark Wagenbuur, from Bicycle Dutch, writes about forever ongoing discussions about building bridges over the IJ, a river that now has seven ferry lines taking passengers back and forth. A quote from the blog:

In the 2015 version, I reported that there were six ferry lines, and now there are seven. The most important one of those, behind Central Station, is the F3 Buiksloterweg ferry. The earliest mention of this ferry dates back to the year 1308, but it is probably much older. Two ships operate on this line during the day, which means a ferry crosses the river every six minutes. During rush hour, a third ferry is added, reducing the waiting time to four minutes. At the busiest times, there is even a fourth ferry in operation.

I was reminded for an incident in January this year when the Metro M52 was closed towards Amsterdam Noord because of a flooding in a tunnel at Sixhavenweg. This was paired with the F3 Buiksloterweg ferries being taken out of service because the platform used to cross when docked had collapsed, the ferries ran to IJplein instead. To make matters works, a number of ferries from the 60 series were taken out of service because of technical problems resulted in cancelling the F9 route between Sporenburg and Zeeburgereiland.

Ruben writes about keeping himself hydrated throughout the day and the ambiguity about how much water is actually required by your body. The thought about tracking water consumption through a database had just entered by mind only to be replaced instantaneously by the simplicity of a spreadsheet. In the very next paragraph Ruben writes:

So I did what every self-respecting computer engineer did: I built a database table to track my water intake! Then realised this was a bit of a pain, and made a spreadsheet instead.

Simon Willison shares his experience serving on the board of the Python Software Foundation over the last two years and some of the responsibilities that entails.

The Python Software Foundation supports the development of Python and the community by allocating their donations towards running infrastructure other activities. However, they are not directly related to developing Python which is handled by the core team ran by the Python Steering Council. Infrastructure includes running PyPi and Python.org and activities most notably include organising PyCon. Simon also mentions an activity I hadn't considered before and that's acting as a fiscal sponsor to other python-related communities.

Simon's write-up is dense with information and definitely worth the read if this is interesting to you. This has also prompted a write-up by Makoto Nozaki on serving on the board of The Perl Foundation.

Dom Corriveau runs his blog of an old Pixel 5, they also mention kaimac.org who I previously read about running their website of an ESP32, and compose.party which was new to be me and is also a website that is run on solar power and off a phone. Lots of interesting links to read through from the blog post along with instructions to replicate.

Discovered via 82MHz' Linkdump.

As annoying as cookie banners are I like seeing them because they give me the choice to deny being tracked but also because I get to see all the vendors a company would have otherwise sold my data too. The longer the list the further I tend to stay away from site unless absolutely necessary. The linked write-up on Bite Code! is a neat summary on why the banners don't have to be as bothersome as they typically are especially because we could have had a standard Do Not Track HTTP header!

There has been for years a proposal for a standard, designed in 2009 (!), still available in all the popular web browsers (except safari) that can make for a seamless experience: the DNT header.

Almost no website have implemented it, because companies WANT to nag you, hopping to trick you into being tracked. They know nobody would click yes on those settings.

So now it's deprecated.

Companies are making your life hard by choice. They got told by the EU they could not be secret abusers anymore, so now they decided to be irritating on top.

I'm not sure any re-org I've been subjected to has led to anything useful but wasted time, confusion and the de-prioritisation of work already done which leads to more wasted time. One aspect of the wasted time comes from the need to explain or maybe even justify the work done by the team having active documentation on the team’s tenants, work and roadmap is one way that helps. Logan's recent blog titled Mean Time To Reorg; Writing as Resilience provides some good examples of this in action.

As a tech lead of a project, these reorgs led to frequent changes in management reporting lines. The twice-annual reporting line change drew my time away from the team and project, and was spent on understanding the positions and intent of the new management structure, and briefing them on the work we did and why it was valuable. [...] Folks new to the project who only did surface level due diligence would misunderstand the details rather than fail to grok them in the first place. When a new hire joins the company there’s an expectation of some number of months before they develop expertise in all of the in-house systems. When there’s a reorg there’s an expectation that the new manager of an area is effective overnight.

I learned that in order to preserve my time for the team and project I needed to improve the speed at which new people (particularly management) onboarded [...] I did this through writing and documentation. We had in-depth docs for users of our product. [...] Over time I learned that essentially writing context and documentation about the roadmap greatly benefited the onboarding of new managers, engineers in the area, and curious customers.

Our car-centric cities and towns rob people of their independence. People who would otherwise be perfectly capable of going out on their own to meet friends, or grocery shop, or go to the library are prevented from doing so because they can’t drive. Maybe they’re too young to drive, as I was at 14; or maybe they’ve gotten old enough that it’s no longer safe for them to drive; or perhaps they have a disability that prevents them from driving. Sometimes people can’t drive simply because they can’t afford a car, because these things are really expensive.

From Evan Sheehan.

John Eldredge, creator of the Winamp Skin Museum, goes about digging in some skins and makes a few interesting discoveries.

Discovered via Tom Scott's Newsletter.

CommonMark is the Markdown specification created by John MacFarlane, Jeff Atwood and others, to encompass the various flavours of Markdown that was adopted by different software over the years. GitHub adopted CommonMark along with it's extension for Markdown called GitHub Flavored Markdown (GFM) sometimes around 2016-2017. I've complained about how different platforms deveate from the standard. This GitHub Engineering post shows how good of a job the CommonMark contributors did to represent common usage of Markdown with only 1% deveating from GitHub's previous Markdown parser.

We [GitHub] actually enabled CommonMark for all new user comments in the website several months ago, with barely anybody noticing — this is a testament to the CommonMark team’s fantastic job at formally specifying the Markdown language in a way that is representative of its real world usage.

All in all, less than 1% of the input documents were modified by the normalization process, matching our [GitHub's] expectations and again proving that the CommonMark spec really represents the real-world usage of the language.

From Matthew Graybosch via People and Blogs

I still have a website on matthewgraybosch.com, but its mainly a substitute for having a LinkedIn account because LinkedIn has always been the Ashley Madison of job hunting, only more cultish. I mean, have you seen the people posting there? Its like the Stepford Wives got equally robotic husbands and they all got corporate jobs.

Yes, yes and a 100 times yes. No social media outlet is more artificial than LinkedIn.

"Simplicity is a great virtue but it requires hard work to achieve it and education to appreciate it. And to make matters worse: complexity sells better."— Edsger Dijkstra

Derek Kedziora made a some good comments on Eugene Yans post titled "Simplicity is An Advantage but Sadly Complexity Sells Better".

Removing unnecessary complexity is a thankless job.

By using satellite images from Google in QGIS and marking areas of potentially good Christmas trees then exporting that layer and joining it with the National Forest Motor Vehicle User Map in Wherobots Cloud using Spatial SQL @lyonwj was able to map the areas with accessible roads.

Theres a lot of interesting bits to unpack there, but my main takeaway was the use of spatial SQL, it certainly beats relying on osmium tags-filters and manually editing GeoJSON files.

Stephen Wolfram, mathematician, computer scientists and CEO of Wolfram Research, describes all the incremental improvements made in different parts of his life in the pursuit of productivity. It's an article I've read and come back to three different times now, and with each I've taken some new bits and pieces that I could use myself.

The reasoning behind pull out shelves:

One of my theories of personal organization is that any flat surface represents a potential "stagnation point" that will tend to accumulate piles of stuff—and the best way to avoid such piles is just to avoid having permanent flat surfaces.

Collecting personal analytics of physical and digital text:

I have systems that keep all sorts of data, including every keystroke I type, every step I take and what my computer screen looks like every minute (sadly, the movie of this is very dull). I also have a whole variety of medical and environmental sensors, as well as data from devices and systems that I interact with.

Archival and searchability:

At the top of my personal homepage is a search box. Type in something like "rhinoceros elephant" and I'll immediately find every email I've sent or received in the past 30 years in which that's appeared, as well as every file on my machine, and every paper document in my archives