Jamie’s blog

Jamie’s blogJamie's blog about software, startups, and swimming (and other things that don't alliterate).SaaS is dead; long live SaaS!https://jamie.ideasasylum.com/2026/06/20/saas-is-dead<div class="prose dark:prose-invert prose-slate mt-2 sm:px-8 font-serif" morss_own_score="3.0" morss_score="192.5">There is little doubt in my mind that AI coding models have fundamentally changed the nature of software development in 2026. But is the SaaS-pocalypse real? I think it kinda is, for some businesses. <h2>A history lesson from an elder developer</h2> I think it’s worth reflecting on what the original premise for Software-as-a-Service actually was. Back in the late 90’s/early-2000’s software was extremely difficult and expensive to produce. Businesses were ordering software for their staff from Microsoft in giant boxes of CDs and manuals, or consultants would fly in to install their database system on the machines in the server room, or you’d have custom software solutions developed by software houses, specifically tailored to your business. In 1998, I was one of those consultants. On one memorable occasion I flew over from Ireland to London with a CD of custom software for calculating bond yield curves in my jacket pocket. I arrived at Barings Bank by 9am only to discover they had a terminal version of Windows on which the installer wouldn’t run. With nothing to do, and no means of modifying the installer, I left the bank and by 10am I started exploring London for the day before my flight home. To me, nothing demonstrates the difficulty, expense, and waste of software delivery during that period than an enjoyable-but-unproductive day at London Zoo. And then the web happened. Specifically, businesses started to embrace the idea that we could deliver software over the Internet, through a web browser. That meant you could develop the software in one location, and deliver it to another location, all without shipping CDs, selling software in retail shops, or flying expensive people around the world to (not) install it. What’s more, you could sell this software to many more businesses which broadened the market for software because suddenly everyone could access it. And, as a software business, you could also take on the cost and responsibility for hosting the software because you were the experts after all. This also broadened the market for software because it could be sold to businesses without any IT expertise or in-house server room. This also brought about a new pricing structure: instead of selling the physical artefacts of the software in the form or disks or CDs for an expensive one-off purchase, you could now license the software for a small monthly payment to anyone in the world. The fundamental bargain at the heart of the SaaS revolution was this: <blockquote> Software is very expensive to produce, difficult to deliver, and out of reach for all but the largest businesses. What if, we instead of spending years developing a software product for a small sophisticated market at a high-price, we delivered software to a global market for a low monthly price. </blockquote> That, in a nutshell, is the SaaS business model which has dominated the software industry for a quarter of a century. The model had trade-offs on both sides. In exchange for the lower price—which wouldn’t cover a fraction of the development costs—the software would not be customised to each client. It was designed to address the needs of the masses and, if you needed something different, there was hopefully a competitor in the new global market that would accommodate your requirements. Or the customer would just adapt to the available software. From the developer’s side, continuous updates were the bargain for keeping customers paying their monthly or annual subscriptions. So the software continuously evolved and added new features; proving its value to existing customers and seeking out new markets. Those software businesses also had to take on all the liability of hosting the software and it’s data—for thousands or millions of customers—providing the servers, maintaining the software, applying security updates, and all manner of other regulatory work. I mention all this because the vast majority of people in the software industry today were not in the industry in 2000. They did not experience ordering a floppy disk of software from a classified ad in a computer magazine. Or licence codes on CD boxes. Or running a SparcStation server under the receptionists desk because that’s the only machine compatible with the business-critical software she used. In short, most developers were professionally born into the era of SaaS and have never considered an alternative model. They have not even conceived that software could, or should, be sold in another way. It’s my belief that AI fundamentally alters the bargain upon which the SaaS model is predicated. We still believe software is hard to develop. So we believe it must be delivered to a large market to justify the effort. So we believe we need venture capital to fund the initial production of it. And we need to charge a monthly subscription to make this palatable to customers. And then we need to constantly ship new features to justify that subscription. I think that sounds like the rantings of a madmen and just doesn’t make sense in a world where software becomes <del>trivial</del> vastly cheaper to produce. <h2>AI changes the SaaS equation</h2> Software is no longer (as) expensive to produce. We don’t need an army of developers to build it (I’d argue we never did) and the gap between a software idea and a working implementation of that idea is basically zero. We are entering a new golden age of software development, which promises a Cambrian explosion of software—and yes, that’s equally fascinating, empowering, and absolutely terrifying. We recently bought an electric car and had solar panels installed on the house and, as the Irish summer slowly got going, I started to wonder whether I should charge the car from the grid today or wait until tomorrow to charge it from the sun. It’s a trivial question, not much more than a customised weather forecast specifically looking at solar energy, and I could have tried to search for a suitable solution. I could, but I didn’t. All this occurred to me as I was going to bed one night and instead of using the App Store’s awful search, I just opened Claude Code on my phone and dictated three paragraphs of requirements to it. Then I brushed my teeth. By the time I was getting into bed, Claude had delivered a mobile-optimised web page that used open weather data for my current location, and calculated which days would be the best days to charge the car. I did not look at the code. I did not source the data. I have no intention of opening the app for anyone else. The only user is me and the only requirements are mine. There’s nothing to maintain, nothing to iterate on or improve. It was 10mins from “I wonder if…” to “I now have the solution”. And this story has been repeated over and over in my personal and professional life in 2026. I’ve always wanted an analytics solution for tracking my game stats on Star Realms but parsing the log files would be a bit of a pain in the neck. Sure, it would be a fun intellectual puzzle but I never found the energy for that outside of work. Instead I dropped a collection of the log files into Claude and told it to write a parser. Then I iterated with Claude to build a Rails app around that so I can paste in new logs and it would generate charts from each game to help me understand if my strategies were working. No logins. No multi-tenant hosting. No figuring out how to monetise it. I’ve built a workout tracker for rehab exercise because I couldn’t find anything simple enough for my needs. I even slapped passkeys on it just because I wanted to. At work—where we’re using AI but not in the 🙈 vibe-coding sense—we’ve seen a 65% increase in the number of PRs shipped to production. We’ve attempted large database migrations in a few days which we’d put off for years. I’ve been able to re-engage with the codebase on a more meaningful level because my AI-partner can fit around my schedule of commitments, never forgets what we were working on, and is now at the level of a senior- or even staff-developer. I’ve loathed the Rails server output for years as the requests, queries, and log lines scroll past with nothing to distinguish them. You’d click a link in the browser and then spend ages hunting through the console output to find the relevant query your were looking for. So I built a UI to view the server logs individually, broken down by requests. All in, it probably took me 30mins to solve a visceral problem I’d had for over a decade, and then for good measure I threw in a TUI. I’ve written an in-house AI bot to triage and investigate new issues (which I’ll expand on in a different post because it’s so good). This bot is the stuff I’ve dreamt about for years, which none of our vendors has delivered on, and which I coded up in a single 4hr session one morning. I’ve rewritten it a few times now, and iterated considerably on its features and tools, but it’s now an extremely important tool at Podia. It costs ~$250/mo to run and I’ve invested a few days of effort into building it over a few months. It’s a staggeringly useful tool—one which was literally priceless a few years ago because it couldn’t exist—and it’s now barely a blip on my effort register. I don’t want to pretend that AI makes software development completely free or trivial for anyone but it’s crystal clear to me that the cost and effort required has fallen dramatically. Solutions which were previously ruled out as too hard, too complex, or too niche are now possible. Perhaps you’re not quite ready to accept AI tools, and perhaps you don’t believe in the productivity gains I’ve seen. That’s ok but, just as a mental experiment, consider what it means for the SaaS model if my premise that software is now much cheaper to produce is correct. Just consider it as a theoretical possibility… <h2>Why would someone buy your software?</h2> This is really where the SaaS-pocalypse concerns are coming from. In a world where software is cheap to produce, why would someone buy yours instead of building their own? And there’s definitely a lot of truth in that. In previous years, there was an opportunity for indie developers and small teams to spend a few months developing a web or mobile app and then sell that to customers. It might not have been a rocketship opportunity for VCs to fight over but it was definitely a viable means for solo developers to make a great living. As an aside: If you can replicate the software that a $100m company sells, you have not created $100m of value in a weekend. You have, at best, destroyed $100m of value but mostly likely you haven’t understood the value of that business. That software you built is worth exactly zero to anyone else but you. That era of building a viable SaaS business in a few months is gone. I mean, it technically still exists today but only in the arbitrage sense that the rest of the world hasn’t yet caught on to how quickly and easily software can be built. It’ll be gone soon, I promise. If you could previously develop a new app in a few months, I can now build that by the end of the week—if not the end of the day. That’s especially because I don’t need to build any of the trappings of a multi-tenant app destined for the mass market. I can choose HTTP basic auth if it suits me. Or none at all. I might not worry about backups. I can host it alongside other internal apps with barely a glancing-thought towards scalability. I don’t need branding. Or marketing. Or billing. I can reuse internal design systems or let the AI run with whatever comes to its mind first. The sophistication of the software I’ll produce this way is much lower than what an indie dev might have written 2 years ago. It’s not the same product—mine isn’t even a product—but it’ll solve my problem equally well. I don’t have to build the same amount of software to solve my problem that you do to deliver a solution to everyone’s problems. After all, why would I rent your software for a monthly fee, which probably doesn’t do exactly what I want, in a way that I want it done, and inevitably does more than I actually need? Why would I adapt my processes to fit your software, when I can build my software to fit my processes? Why pay for a platform and only use 20% of its features? Why accept the constant changes and learning required because new features, or redesigns, are forced upon me? Why accept an infinitely flexible monster (go Jira), with all the inevitable trade-offs for usability and complexity, when I could have software that fits my team like a glove? Why should the local sign makers have to buy the same project management solution as a software company? Why should the small business selling electric gates use the same stock control system as a major retailer? I use those two examples because they are actual examples of conversations I had with local businesses about 15 years ago. At the time, both businesses had these problems and couldn’t find SaaS solutions to their business problems. They actually needed a custom software solution but couldn’t afford it. Perhaps they would have paid $5-10k for the solution but not the $100k+ it would have actually cost to build one. Today, I do think that equation has changed. I think if I talked to those businesses today we could engage in a successful project to build a custom solution for them—at a price they could afford, in weeks, not years. So it’s all gone, right? It’s all doom-and-gloom now in a post-AI world. Our jobs, businesses, and livelihoods are all inevitably going to fall to a single Anthropic subscription? The software industry is going to crash and burn? No, I don’t think so. Not all of it. Software itself has lost a lot of its value but not all software businesses have. <h2>What remains valuable</h2> For the past 25 years, SaaS has been all about the Software. The software was hard to build, so the software was valuable to sell, so the software was the moat against competitors. SaaS of the next 25 years will be all about the Service. It’s not about delivering software as a service, it’s about figuring out what service you can deliver through software. It’ll be how software enables you to deliver something else: relationships, access, data, trust, expertise, or community. Given two competing software solutions, the winner will not be based on the feature set. These days you could take a competitors feature announcement and a few help docs, and ship that feature in your product in days or weeks. Hell, if you were so AI-pilled you’d have an agent monitoring competitor announcements and automatically implementing them in your own app (don’t do that; it’s dumb because…). Features aren’t a defence. The hedonic treadmill of constant feature development, and a constant pressure to broaden the app, just doesn’t have much value any more. More software is not solving more of an individual user’s problems, it most likely just helping the app appeal to a broader market. And that’s a goal which serves the business, not any one customer. Since everyone needs a slightly different set of features, that inevitably broad set of features in SaaS apps comes at a cost: why is your customer paying $200/mo for an app but only using 20% of the features? At some point that’s going to look like bad value to them but, in a pre-AI world, they still couldn’t justify rebuilding it. In a post-AI world, the customer doesn’t need to replace your entire app, they only need to replace the 20% of the features they actually use and that might only take a few days of effort. More software has not saved you: it’s just encouraged your customer to seek a streamlined alternative that fits their needs, and that alternative is now easier than ever to build yourself. What matters much more—and always did—is what features you choose to ship, how you engage with your customers, and how you serve them. It’s all the non-software things which make a difference now. Here’s what I would bet on… Anything that’s a true system of record: If a business needs a single canonical place for their data, and you are that trusted party, you will remain valuable. Anything that provides access to a data source: When software is cheap, owning the data is more valuable. Think access to datasets, APIs, payments etc. Anything that connects people and communities: Relationships aren’t replaced by cheap software and, if anything, they become even more important. Anything with a high regulatory burden: payments, healthcare, personal data etc. The software might be cheap but the compliance isn’t and businesses will still pay to have those problems solved. Anything with a high complexity, low-risk tolerance, or niche domain knowledge: Again, businesses aren’t really paying for the software any more, they are paying for the trust and expertise. Anything coupled to hardware: software is cheap but hardware remains expensive so I’d expect this category to remain valuable. Consultants + AI: some think AI will kill consultancy but I think it’ll super-charge them: AI thrives in the hands of experts and withers in the hands of idiots. They are going to be transforming businesses, bringing vastly more software into the world than SaaS ever did. Consultingware 2.0? Internal product teams: I think there’s a huge opportunity for large Enterprises to develop internal tools for a fraction of the cost of the contracts they are currently paying. Small 4-person teams using AI can replace multi-million $ contracts and allow the business to operate the way it wants, free of annual contracts. Anything that agents use: UI components, libraries, APIs, payments, harnesses, models, etc. It doesn’t matter half as much what UI you put on your app, as what UI your customers can build on your API. Anything that gives trust in AI-generated code: tools for hosting, debugging, reviewing, observability, monitoring, auditing, security, etc. Consumer apps: whilst businesses will find software more within their reach to develop, I’m not convinced the average person will be coding their own accounting solution, email clients, games, fitness trackers, or reminders apps. That said, the consumer is usually high-investment and low-margin. Software without an agenda: I could joke than many software startups barely displayed a money-making intention in the past but now we can genuinely build software with no plans for commercialisation—without even a plan to share it as open-source. Just build it to solve your own problems without concerning yourself with anyone else. It won’t be a dramatic “SaaS-pocalyse” but I do think it will be a restructuring of the software market and SaaS won’t be the default business model. Consumers will still mostly buy the software they current do, typically from app stores & marketplaces, small subscriptions, or via ad-supported models. Small/medium businesses will still buy some software but increasingly will build internally or subcontract out for custom solutions. I would expect large enterprises, who already have internal resources to develop and maintain software, will be the biggest beneficiaries with a lot more internal building & subcontracting, and a lot less buying. </div> Sat, 20 Jun 2026 00:00:00 UTCSwimming review: 2025https://jamie.ideasasylum.com/2025/12/31/swimming-review-2025<div class="prose dark:prose-invert prose-slate mt-2 sm:px-8 font-serif" morss_own_score="2.9097410604192357" morss_score="297.26733296698">This post probably is not for you. It’s the indulgent reflections of a middling mid-life swimmer who really needs to remember to be proud of himself and, if he doesn’t write all this down, he’s likely to forget he did any of it. <h2>Highlights</h2> <ul> <li>First swim races/events in 10 years <ul> <li>Gaelforce Trilogy (3.5km, 3.9km, 3.9km)</li> <li>Garnish Island (5km)</li> </ul> </li> <li>Personal bests set this year: <ul> <li>800m</li> <li>1000m</li> <li>1500m</li> <li>Longest open water swim (5.6km)</li> <li>Longest swim (8km)</li> <li>Highest annual total (316,517m)</li> </ul> </li> <li>Shoulder injury: recovered</li> <li>Left ankle: recovered</li> <li>Right ankle: new injury!</li> </ul> <h2>Preamble</h2> The start of my 2025 swimming really started in late-2024 and with an unlikely joint. I’d been having ankle problems for about 18 months and working with a physio and rehab exercises in the gym to fix it. I finally got a specialist consultant appointment, and another MRI, and was diagosed with osteoarthritis in the ankle joint. That wasn’t a surprise since it has two screws in it from a car accident. Before Christmas I had an injection of hyaluronic acid into the ankle so I had to rest for a week, and only light swimming. But none of that mattered because I’d also buggered up my right shoulder. I started 2025 sitting on the couch, unable to swim, and figuring out how best to protect my ankle. Logically, I thought, this was the best time to sign up for The Gaelforce Trilogy—three swims in river, lake, and sea between May and September. The only swim events I’ve ever done <a href="https://jamie.ideasasylum.com/2015/05/scared-proud/">were 10 years ago</a> and that was a 1km “paddle in the park” and the 2km Lee river swim. With a few clicks I’d just entered a 3.4km, 3.9km, and 3.9km race and I wasn’t sure if or when I’d be able to swim again. <h2>January: Meh</h2> It was a miserable month with almost no swimming. In it’s place was lots of physio on the shoulder, particularly dry-needling on the infraspinatus and lats. Infraspinatus is a nasty muscle to have dry-needled as the pain radiates over the shoulder and down the arm. Did not enjoy. I started doing some cardio sessions on the bike at the gym just to replace the swimming and keep me active. Also meh. <h2>February: Slightly-less-but-still Meh</h2> I got back to some light swimming, mostly drills like sculling and fists to build up the rotator cuff without tightening everything up again. Still doing lots of rehab work in the gym and at home. The number of exercise bands in the house multiplied. <h2>March: I’m (slowly) back</h2> I was finally back swimming around 2-2.5km sessions, without pushing it, and continuing all the physio and rehab work in the gym. I did a CSS test to determine my threshold pace which came out as 1:42/100m <h2>April: Training hard</h2> Things are starting to get serious with The Great River Swim only 6 weeks away. I needed to get out of the pool and into the open water but, even with some unseasonably warm weather, the sea was still freezing cold. I got in two short (<1km) swims during the month and that’s about 8 weeks earlier than I’d ever been in the sea. Having the race was a huge factor forcing me into the water much earlier, and much less comfortable, than ever before. I was swimming between 3-4 times per week, 2 sessions in the gym, and rehab exercises each day it felt like I was seriously training for the first time in my life. There was rarely an “off” day. Apparently, and this still sounds ridiculous in my head, I need to think of myself as an “athlete” and prioritise fueling, training, and recovery accordingly. Got some new pool goggles: Arena Cobra Ultra Swipe. They really live up to their marketing hype of being anti-fog, which is a first for any goggle I’ve owned. Just wet the googles, swipe the inside with a finger, and hey-presto: no fog. That said, I tended to favour the Magic5 goggles for most swims as they are, by far, the most comfortable pairs I own. <h2>May: The Great River Swim</h2> The unseasonably warm weather had continued and we were treated to the best of Irish weather: blue skies, sunshine, and not too warm. We took the shuttle bus up to the starting point and the guy next to me asked how long I thought it would take me. I thought about 1:10 depending on the current. He replied, “oh, wow. You should have been in Wave 1 with the elite swimmers”. Um… what? The river was lower than usual and as we entered the water I looked around and started to wonder if maybe I was going to be faster than most of the wave. I moved up towards the front of the pack at the start line but there was still a fair number of people in front of me when the race started. I started swimming, filled with nervous adrenaline, but unfortunately some of the people in front of me hesitated or were waiting for more space. I kinda swam over one person and my elbow hit some part of another. Sorry. That was my first experience of the chaos of a swim start. I’d not swam in a river in a long time and I’d forgotten how easy it is: the current suited my long smooth stroke, the water was calm, and there was basically no navigation required. It was like swimming on rails, just following the meandering river from one corner to the next surrounded by reeds on either side. After a few minutes I’d cleared most of the other swimmers but there was maybe 3 ahead of me. One by one I started passing them out. And then I started passing out people from Wave 1 which had started 10 minutes before us. And some swimmers from, I think, the 12km event that had started hours earlier—they were looking pretty exhausted by that point. After maybe 1km, I noticed someone drafting on my feet but they weren’t touching my toes so, whatever, enjoy the free ride! Coming up to the bridge, which was the only real navigation required, I spotted the family on the bridge and gave them a quick wave then put my head down and lost the person on my toes. I managed a bit of a sprint for the finish and then struggled to actually get out! The random song for the day, by virtue of having these few lines stuck in my head for the entire swim: <blockquote> Gotta give a little something Gotta take a little pain Gotta dig a little deeper We gotta keep the faith Hell or high water We didn’t come to lose —Eclipse, We didn’t come to lose</blockquote> I finished the 3.5km course in 0:58:43: first in Wave 2, 10th place overall, and 2nd in my age group. A very good result for my first race! Though, upon reflection, most of the better swimmers were probably doing the 5.8km or 12km races. Something for next year, maybe. I was tired afterwards and my lats were pretty sore but otherwise I felt really good about it. There was a nice little festival atmosphere at the end and I enjoyed sitting on the grass afterwards having a picnic of Coke and fizzy jellies with the family. My daughter repeatedly told me how proud she was of me so that was good. After the swim I flew out for the company retreat in the US… and came back sick. Again. 😖 <h2>June: Mallorca</h2> June kicked off with the SwimSmooth course in Mallorca. There’s really nothing more I need to add to <a href="https://jamie.ideasasylum.com/2024/06/08/swim-smooth-mallorca-2024">my blog post about the experience in 2024</a>. It was a different group of people but the experience was just a good, the 50m pool at the BEST Centre was as glorious as ever, the hotel was nicer, and the coaching just as valuable. I set a new CSS of 1:35/100m showing that the training I’d done for the river swim had paid off, but I then spent the week trying—and failing—to live up to it! 🥵 I think I was still recovering from being sick but I completely died during one memorable pool session. I’d never experienced anything like that before: I was at the end of the pool, holding on, and I knew someone was talking to me but I couldn’t actually hear them or process the words. I ended that session early but it lives on in the SwimSmooth Guru app as <a href="https://www.swimsmooth.guru/sessions/direct/8o3faZxPyjjNwZpzDAEy/4200">“Jamie’s Downfall”</a>. I also remember leaving everything in the pool after some sprints and being destroyed afterwards. My shoulder injury also flaired up on another session but was mostly ok for the rest of the week. I left inspired after a conversation with Paul and I entered the only 5km swim left in the Irish 2025 swim calendar: Garnish Island. It was a swim I’d wanted to do for a while but it sells out quickly and I’d never had the confidence to enter the 5km distance. I had until mid-August to prepare. At the end of June I was back in Brighton for the Brighton Ruby conference and, of course, took advantage of the wonderful 50m Sea Lanes pool on the beachfront. I was a bit apprehensive about the Garnish Island swim so I decided to prove to myself that I could swim 5km by doing 10x 500m. Then I did it again the next day to prove to myself that I could do it when I was fatigued. <h2>July: The Great Lake Swim</h2> It was one of the hottest days of the year at Lough Derg for the Great Lake Swim, with the air temp around 30°C and the water 22°C. I swam the 3.9km distance in 1:06:46, placing 14th overall and 7th in my age group. I was pretty happy with that as it was at the faster end of my prediction. I also swam non-stop and, again, enjoyed passing lots of Wave 1 swimmers. After the first few hundred meters, I told myself that I wasn’t going to let anyone pass me—and I didn’t. I’m not really competitive but I do enjoy setting challenges for myself even if that means using other people to do it. Another mental game I played was thinking of passing other swimmers as “taking caps” (though, of course, I’m not actually assaulting them). Anything to pass the time and make it interesting. There isn’t much room in my brain for thinking during a race. I’m just constantly iterating between sighting for the next buoy, looking out for other swimmers/kayakers etc, enjoying the view of the weeds and fish, repeat a line from a song, counting my strokes (for no reason), reminding myself to stay calm, to relax, and to keep my form. After around 2km my thumbs and index, middle, and ring fingers felt like they were wrapped in elastoplast: they felt big, slight pins & needles, and I couldn’t feel them well (but could still move them). They weren’t white. I could keep swimming but I had less feel for the water. It went away after a few minutes on shore. The same thing had started creeping in during the river swim and the 5km in the pool but it was more obvious now. I felt a bit flat afterwards. I didn’t have the high of the first swim and I didn’t have anything to really compare myself to. Was this a good time? I didn’t really have any external reference for was “good” looked like. <h2>August: Garnish Island</h2> August was about preparing for the Garnish Island swim and I kicked off my birthday with my longest open water swim: 5,600m around Myrtleville. I rarely plan things out and I never tell anyone what I’m attempting but I thought a really good long swim in the sea what give me some more confidence going into the Garnish Island swim. The birthday swim was also an opportunity to try something I’d never done before: feeding. Packed into a new donut tow float I had a Rice Krispy Square, Precision Fuel energy gels, and some ‘Zero’ electrolytes. I made a point of stopping after the second and third laps to feed even if I felt like I didn’t need it. My birthday also brought a new watch as a present! Hilary bought me a Garmin Forerunner 570 to replace my beloved-but-aging Garmin Swim 2. I was apprehensive because of the “Forerunner” name but all the same swim functions were there and the buttons did the same thing so my muscle memory still worked. I also got lots more data like HRV, training readiness, training effect, better sleep data, ability to track strength workouts, water temperature, recovery time… and a lot more I don’t pay attention to. I highly recommend this watch for swimmers. And, most interestingly, I was finally able to install the SwimSmooth app on the watch to analyse my stroke. Sure enough, my top priority is to fix my straight-arm pull, particularly with the right arm. Glengarriff was packed on the day of the Garnish Island swim, as it is most weekends during the summer but now there were a few hundred swimmers & spectators there too. I left the family in town to get food and I took the shuttle bus to the starting point on my own. That felt kinda weird as they’d always been there to see me off in the previous races. The song on today’s mental jukebox in a 10 sec loop was <a href="https://youtu.be/I3Fn4s3SsHc?si=jmSDth5jnhaujKj_">Halestorm’s Everest</a>: <blockquote> It’s the shame, it’s the pain, something I can’t explain It’s a noise, it’s a voice, and it screams in my brain It’s a curse, and it hurts, but I just keep going</blockquote> It’s a fairly simple course: out from the slip to the island, two counter-clockwise loops, and a straight swim back. I stuck to my plan of executing one lap of the island (~2.5km) and then stopping for a feed. I think I only stopped for about a minute and just had a quick drink but it was a revelation! Feeding (or, as I told the kayaker, my picnic) made me feel like I could swim… forever?! It was also a fun and varied swim: calm heading out, then a bit of chop, then we were practically surfing with the waves behind us, and finally back to a serene bit of water. My goggles started fogging up after about 100m so I had to flip onto my back, take them off, spit into them, rinse them, and pop them back on. Garnish was an event (not a race) so there was no official timings or positions but my watch timed it at 1:30.31 for the 5km. For what it’s worth, Strava had me at 4th for those that recorded it, and 7th best time for that route. I don’t put a lot of stock in that but I felt really good. The 2km Myrtleville Swim, postponed from June, was cancelled again. I reckon the stars will never align for me to complete that one. <h2>September: The Great Fjord Swim</h2> After Garnish Island I stopped really worrying about the Killary Fjord swim. I’d just swum 5km and I’d done 3.9km back in July so this should be a walk in the park. Right? My preparation wasn’t great with a few weeks of low mood, low energy, and conference travel getting in the way of some long training swims. Mostly I managed 3km sessions. A week out from the swim, the weather forecast looked pretty wet & windy. With a westerly windy, it would mean the wind coming from behind us on the first leg, and into our faces on the return journey. My plan was just to start reasonably close to the front of the wave, use the first kilometre to get settled, then nice calm swimming with long strokes into the turn. Then I knew we’d be into the wind and the chop, and it would be a higher tempo fight on the way home. The forecast materialised as predicted with 25kph winds, gusting 45kph, funnelling straight down the fjord but, apart from a rain shower as we walked down to the start line, it stayed dry for the swim. We even got a lovely rainbow and some blinding sunshine on the swim out. Thankfully the organisers had changed the route from a looping rectangular course to a straight out-and-back, sticking close to the shore where it wasn’t quite as lumpy. Today’s song was <a href="https://www.youtube.com/watch?v=GWTNGMAH500">The Pretty Reckless’ For I am Death</a>: <blockquote> For I am death and I can feel I got my hands upon the wheel I am not lost For I have found the only one Who put me down For I am death and I won’t break I got a life I’ve got to take When will it end, this sufferin’ of late? It was nice to know you</blockquote> You might have wondered what thoughts go through a swimmer’s head as it’s being pounded by waves for an hour and a half, but you probably didn’t expect this man in his 40’s, wrapped in neoprene, to be saying “for I am death… for I am death… for I am death” on repeat. Things mostly went as I thought they would: I started well and got ahead of the pack. When my goggles started leaking after about 100m, I rolled onto my back to fix them and there was a few people ahead and no one on my feet. Yes, the same thing that happened in Garnish Island. I think I need a better system of getting the goggles race-ready. I continued for what felt like foreverrrrrr. The only notable events on the outbound leg was nearly missing a buoy because it was positioned closer to shore and I was sighting on the next one, and my tow float being constantly blown forward and banging into my elbows. The wind might have been with us but it was a pretty frustrating experience because on every single recovery the tow float would be blown into my arm. I decided that I’d take a quick break at the second turning buoy. My fingers were already numb and I knew it was going to be a slog on the way home. I took 40secs to have a pee (don’t judge!) and stretch out the arms then I headed into it. “It” was choppy waves that constantly slammed into my head, dug into my arm, or left me catching thin air. I channelled all the videos of Darragh swimming around Ireland or Ross battling through immense storms around Iceland. Conditions weren’t that bad, it wasn’t as far, and I wasn’t that good, but the visualisation helped a lot. I staggered out, leaning on a volunteer for support after 1:29:18; 31st overall and 9th in my age group. My watch measured the course at 4.5km though and this seemed to be corroborated by a few others. The time and placings might not look as god but I think it was the most competitive race I’d done and the conditions were easily the worst I’d swam through. <h2>October-December</h2> Bleugh. I wasn’t feeling in love with swimming for October and November and, even though I was on sabbatical for a month, I swam less than usual. Then I hurt my achilles. My other ankle. I started this year with a busted left ankle and finished it with a busted right ankle. Thank god I’m a swimmer and not a runner. I started the SwimSmooth 10 week CSS challenge for something to do. My CSS test came out a little slower (1:36/100m) than in Mallorca and I didn’t decrease the CSS pace each week as I was supposed to. I really enjoyed one of the sessions each week which had 800m of drills, 400m of 50m fast/slow/tempos, 8x100m at CSS pace, then some variation of 1000m at CSS (400/300/200/100, or 2x 500m etc). I became a lot more comfortable swimming at my threshold pace and by Christmas I was struggling to swim slow enough! Instead of 1:35/100 I’d be hitting 1:30, 1:32, 1:32, 1:32 (no, I was really trying to go slow on that one), 1:32 (wtf!), 1:34… though I still struggled with the pace on longer intervals and I definitely still struggled with the rest times. With the short fast/slows in that set, I started wondering how hard a 1:00/100m sprint is. It’s fucking hard and I’m waaaaaaayyyyy off (about 1:20). Still, it’s kind of an interesting medium-term goal. A 20m pool isn’t ideal because you hardly get going before you’re turning again, and I don’t usually tumble turn (I can—I just don’t like it and I’m not fast at it). Oh, and I can’t dive in this pool. Anyway, I played around with sprinting a bit for fun and managed to hit a 1:00 pace for a single length (12 secs for the 20m) and 1:07 for two. It’s a milestone. Maybe I’ll do something with it next year. The pool was closing early on the 23rd December and I got it into my head to attempt a long swim which might help me decide on some goals for next year. Like, a really long swim. I chose 8km as a potentially achievable distance though it would be my longest swim ever. 2x 400m, 12x 600m, and 2h40m later I got it done. Interestingly, I swam it at an average 1:42/100m pace which was the threshold pace I started off with in March and I’d have struggled to maintain for much more than 1000m back then. I spent the rest of the day on the couch! Whilst writing this I also discovered that I was just 2km shy of my highest annual total so, yeah, you bet I got in the pool and claimed that PB too. <h2>Inspiring swims of the year</h2> Instagram is my drug of choice for swim content and there was a ton of inspiring swims to enjoy this year. It was incredible to watch <a href="https://www.swimeire.ie/">Darragh Morgan’s swim around Ireland</a> and <a href="https://swimswam.com/ross-edgley-swims-1000-miles-around-iceland/">Ross Edgley’s swim around Iceland</a>. Absolutely insane overall distances, insane daily distances, insane sea conditions. Insane. Loved them. Brian Foster did <a href="https://www.southernstar.ie/news/kinsale-teen-brian-sets-trio-of-world-records-4337467">a triple Fastnet Swim</a>, breaking the record for the one-way swim, and then breaking the record for the two-way swim, and being the first to complete the triple. Also insane. We also had the Olympics! Watching Ireland’s Daniel Wiffen win the 800m was a source of national pride. Bobby Finke’s amazing 1500m final was stunning from the start. “Can he hold on? Can he? He fucking can!”—I loved it even as he was beating Wiffen. And then there’s Katie Ledecky: I can’t get enough of watching her dominate the 1500m. I saw <a href="https://www.instagram.com/catherine.breed/?hl=en">Catherine Breed’s</a> Faralon Islands swim, <a href="https://www.instagram.com/sarahswims04/">Sarah Thomas</a> just repping out daily 10km swims for practice, <a href="https://www.youtube.com/watch?v=R1z8aWxuRcQ">Hector Pardoe swimming the three longest lakes in the UK in under 24hrs</a>, and <a href="https://www.instagram.com/amy.swims/?hl=en">Amy’s</a> double Windamere. David from the SwimSmooth Mallorca camp did a double crossing of Lake Geneva. Paul was an inspiration again, winning the Oceanman 10km in Dubai and <a href="https://blog.swimsmooth.com/p/getting-better-with-age">showing what your 40’s can look like</a>. <h2>Next?</h2> I think before I got into the water in Lanesborough I realised I might have under-estimated myself; by the time I got out I knew I had. I was a bit apprehensive about the 5km swim in Garnish but I quickly proved to myself I could do it. And the 3.9km swim at Killary proved I swim in tough conditions—though it was a real slog. As soon as the Gaelforce Trilogy was over, I started looking ahead to swims I could complete in 2026. I’m inspired by these massive swims I see other people doing but I’m also resolutely uninterested in cold water—I just don’t enjoy it or find it rewarding, or invigorating, or feel the need to prove myself according to “channel rules”. I have no intention of swimming the English Channel. I’d like to do a 10km swim (a marathon distance for the runners still reading this indulgant rubbish). Or, more precisely, I think I’m capable of swimming 10km and I’d like to have done it. The actual doing of it isn’t likely to be pleasant. As of New Year’s Eve, I’m thinking of redo-ing the Gaelforce Trilogy but this time as 12km for the river swim, 3.9km for the lake (still the longest at that event), and the new 8km distance for the fjord. I think the river is doable though remembering the look of those 12km swimmers I passed adds a bit of reality to my optimism. The fjord makes me nervous if only because it’s 8km, straight West into the prevailing wind, which might be howling like this year. I’m not sure I want to travel for swim races. The UK is convenient, and has a lot of events, but the water conditions aren’t good and there’s a lot more expense and logistics. I’d also like to do an Ultraswim (33.3km over 4 days) but not sure I make the dates work or convince myself I have the endurance for it. Very tempting though. <h2>Random Reflections</h2> Wave 2 is quite nice. I enjoyed the less aggressive start line energy, relished the game of catching Wave 1 swimmers, and it was definitely easier having a constant stream of swimmers with their colourful tow floats to guide the way. That said, if I do it again next year I should be starting in Wave 1. Advice to paddlers: don’t bother talking to a swimmer when they are swimming. We can’t hear a thing with the sound of the water and some people need to use ear plugs. Pointing your arm or paddle in the direction is far more effective. You’re really more like a sheepdog that uses it’s body, not its bark, to control the sheep. Ideally position yourself to the left or right of the swim route and use your presence to shepherd swimmers in the right direction—we’ll try to steer away from you and the closer you get, the more urgently we’ll move. Mostly. I’ve taught myself to celebrate like an Olympian. It’s all too easy for me to complete a tough set, maybe set a new PB, and get out of the pool, sit in the sauna, get changed, leave, and mention it to no one. Instead, I’ve started celebrating like I’ve just won the Olympics: shouting “YES!!”, “Fuck, yeah”, punching the water etc. The pool isn’t usually busy so I’m not disrupting any one though I probably appear like a lunatic. I don’t care. No one else knows what I’m doing, no one else cares or understands what I might have achieved, but my body does and it deserves to physically celebrate that achievement. It deserves more than just a mental nod of approval and slinking off home. Songs for swims. A single line of lyrics on repeat can get me through a swim so I have a curated playlist now. Feeding is great. I need to do this more and it really opens up longer swims. There’s also a mindset change about not worrying about distance so much and thinking more about continuing for a certain duration. Numb fingers thing is annoying. I think it could be Thorasic Outlet Syndrome (nerve restriction around the collarbone) and I’m going to investigate more in the new year to see if there are exercises I can do to reduce it. Coaching and strength training. I’ve been swimming ~3 times a week for about 12 years but in-person coaching and strength training in the past 2 years has brought my swimming to a completely new level. Don’t sleep on it (I wish I didn’t). Comparison is the thief of joy. I’m not as good as many other swimmers and there’s plenty of opportunities to remind myself of that. On the other hand, I’m practically elite compared to the general population and I’m much much better than past-Jamie. </div> Wed, 31 Dec 2025 00:00:00 UTCWhat I can changehttps://jamie.ideasasylum.com/2025/10/29/what-i-can-change<article class="mx-auto sm:mx-2 max-w-prose" morss_own_score="6.978428351309708" morss_score="11.47378438846141"> <h1>What I can change</h1> <img src="https://jamie.ideasasylum.com/2025/10/29/what-i-can-change"> <div class="prose dark:prose-invert prose-slate mt-2 sm:px-8 font-serif" morss_own_score="2.9907120743034055" morss_score="49.91025230418846"><blockquote> God, grant me the serenity to accept the things I cannot change, the courage to change the things I can, and the wisdom to know the difference. </blockquote> I think everyone has heard that before. I was vaguely familiar with it but it’s not a particularly core part of my life. Or so I thought. I’ve been reflecting a bit on larger-scale changes which are happening around me: the major geopolitical changes we’re seeing, the rise of AI, the latest Ruby drama, climate change, etc. I realise now that my default response is to understand how to navigate this world as it is, not change it to how I believe it should be. Perhaps it’s the swimmer and the stoic in me: I’m swimming in the sea, adjusting my stroke, sighting for landmarks, and ploughing on because I can’t change the conditions. I can only change how I respond to them. I can’t wish for the wind to stop blowing waves into my face when I breathe but I can turn my head to the other side. I can’t change the water temperature but I can have predicted these conditions and trained for it (or worn a wetsuit). I can’t make the fog lift but I can remember the course, sight more frequently, and follow other swimmers or the directions of kayakers. I can’t do anything when a storm cancels a swim but I can be ready if it doesn’t. Others, I’ve noticed, are not like this. Perhaps stronger or braver, they rage against this tide. They scream at the injustice of it—as if life was supposed to be fair—and they rally others to the cause. They see a delta between how the world is and how they believe the world should be, and they believe they can change it. I guess? Or, at least, they seem to expend significant energy complaining about it. I don’t. That person is who they are. That company does what it does. That country is what it is. That government is exactly what they said they were. That situation happened. That technology does what it does. This industry is going that way. The climate is getting worse. Those incentives are aligned to inevitably produce this outcome. People are going to be who they are. I am who I am. I can’t change much of that. I’m so small in this world that all I can do is figure out what has happened or predict what’s going to happen, how that affects me (if at all), and how I react to it. I have control over myself, some influence over others around me, and a vanishingly small effect on the rest of the world. This is mostly how I’m thinking about AI: It’s inevitable. There is so much money aligned behind AI that the incentives for powerful people, and frankly the entire US economy, mean that AI is going to be pushed and pushed and pushed. And even with the simplest of tasks and the most naive implementation, the hint at value is there. This is not crypto. This is not a solution searching for a problem that normal people have (who aren’t money launderers)<a href="https://jamie.ideasasylum.com/2025/10/29/what-i-can-change#fn:1">1</a>. AI is going to be widely available, widely distributed, and widely integrated. Once you accept that AI is going to be a thing, you can start figuring out if it’s going to affect you (undoubtedly, yes). Then you can start plotting how you’ll react to it. The future is the sea, AI is the prevailing weather conditions, and your career is the course you need to plot. What are you going to do about it? How will you adapt your stroke? How will you change your mindset? What need equipment or skills do you need? Or are you withdrawing from the race? Putting your head in the sand, ignoring AI, protesting against its use or the crimes of its evangelical creators, or hoping it’ll go away is a losing strategy. It’s exhausting, it’s depressing, and it simply delays your actual reaction to the situation. It’s not getting you anywhere. Stop screaming into the storm and start learning how to swim in it. <ol> <li> Ironically, I actually think AI might provide a legitimate use-case for crypto if stablecoins are used for Agent-to-Agent commerce. We’ll see if that actually becomes a thing. It still feels like a solution searching for a problem. <a href="https://jamie.ideasasylum.com/2025/10/29/what-i-can-change#fnref:1">↩</a> </li> </ol> </div> </article> Wed, 29 Oct 2025 00:00:00 UTCWhy you need to adopt AI (as a software engineer)https://jamie.ideasasylum.com/2025/10/24/why-you-need-to-adopt-ai-(as-a-software-engineer)<div class="prose dark:prose-invert prose-slate mt-2 sm:px-8 font-serif" morss_own_score="2.9438567941415785" morss_score="206.72326855884745">Look, I get it, you’re sick to death of hearing about AI. It’s in the news, it’s all over social media, it’s in your favourite apps, and the slop is infecting every corner of your world. You just want to go outside, touch grass, and hope the whole thing goes away. I have devastating news for you: it’s not going away. I’m not a AI-evangalist, and I’m not an AI-luddite. I’m something much more nuanced because life, and humans, are messy and not easily categorised into a divisive binary system (hello, US politics!). So I’d like to explain my position with probably an excruciating amount of detail… <h2>What is AI?</h2> <blockquote> AI is a label that you apply to a product that’s worse than the alternative product without AI but it sells for more money </blockquote> j/k, kinda. I’m mean, that’s basically a true reflection of our frothy bubble-like times but it’s not particularly useful. AI is ‘Artificial Intelligence’, a field of academic research that’s almost 70 years old (the term was coined in 1956) and covers everything from speech recognition, robotics, computer vision, agent-based simulations, neural networks, machine learning, genetic algorithms, etc etc. It’s incredibly broad and has been widely taught in colleges for the past 30 years. It is not new. You will be familiar with many spin-offs of AI and they probably have a place in your home or pocket. Face recogition in your camera. Siri and Alexa. The entire personalised ad industry of the Internet. Book recommendations. Robot vacuums and lawnmowers. Grocery store vouchers. Fitness watches with their activity detection, analysis, and prediction. Gmail spam detection. Voicemail transcription. Google Translate. Non-playable characters in video games. AI is a broad field spanning all the way from cognitive science to computer science. It’s not all uncanny videos of 6-fingered/three-armed women and mediocre writing. That’s just how the current crop of AI companies, led by OpenAI, have claimed the word for themselves to specifically refer to systems based on Large Language Models (LLMs). I am more than little bit peeved the ‘AI’ name has been captured by such a narrow part of the field but that battle is lost now. The broader AI field will still be around, providing hugely diverse and interesting research, for decades to come. It’s this AI-as-LLMs that we need to discuss today. The last few years can be summarised as: <ul> <li>the latent potential of decades of research into neural networks +</li> <li>the invention of <a href="https://en.wikipedia.org/wiki/Attention_Is_All_You_Need">the transformer architecture</a> +</li> <li>the availability a massive training corpus in the Internet +</li> <li>rapidly increased computational power (particularly in GPUs that excel at vector arithmetic) +</li> <li>and the recognition by commercial companies that injecting large amounts of capital into this machine could now produce meaningfully-useful results.</li> </ul> That’s the basic ingredients for the AI boom we’re experiencing. <h2>Is it all just a stinking pile?</h2> All this enthusiasm has created a lot of noise. There’s a sudden gold rush to figure out what this technology can do and who will pay for it. The enthusiasm is exploding outwards in every direction with lots of energy, no direction, and no consideration for what’s in its way. Most of it, I think, is dross. It ranges from the pointless, to the awful, to the dangerous. Actually, there’s quite a lot of danger because <a href="https://jamie.ideasasylum.com/2025/10/15/you-are-the-scariest-monster-in-the-woods">the scariest monsters in the woods</a> (powerful people) just get scarier when you equip them with AI. For example, t’s never been easier or cheaper to <a href="https://www.thejournal.ie/catherine-connolly-ai-deepfake-condemns-6851912-Oct2025/">create fake videos of presidential candiates dropping out of the election</a> in an effort to subvert democracy. And there’s a lot more bad news to go around… Mass copyright-infringement. Forced adoption by companies. Life-altering decisions being made by AI. Incomprehensible amounts of energy usage during a time of a climate crisis. Hallucinations. Six-fingered women. Data centers. A pyramid scheme of investments holding up the entire AI industry (and US economy). Inhuman standards of beauty. Synthetic relationships. AGI. Goodbye searching Google; Hello asking ChatGPT. Life-threatening “medical” advice. AI-suicides. Misinformation. Disinformation. Lack of critical thinking. The mass-transformation of jobs for people into AI-revenue concentrated in just one or two companies. Did I mention the six-fingered women? Sometimes they have three arms too. Vibe-coding. AI-drones. AI-intelligence gathering. AI-enabled browsers snooping on your web activity. Adverts generated on-demand tailored just to you. Sycophantic responses. “I’m sorry, you’re right”. Slop. That’s a lot to be worried about. A lot to be pissed off about. But maybe not everything is bad; maybe there is some value to be found in this pile of shit. <h2>What you do need to care about</h2> Right now there is billions of dollars being poured into AI technology. It’s flowing into the few major players (OpenAI, Anthropic), it’s being reinvested inside the likes of Facebook and Google, and it’s trickling down into any company that can slap “AI” on the side of their product. The scale of the money involved is so large that there are convincing arguments that AI is the only thing propping up the US economy right now. With all this money on the line, there’s a lot of marketing going on and it’s broadly targeted at the decision makers in government and commerce. Every country seems to need an AI strategy now. We’ve seen Shopify, Coinbase, Intercom, and host of other companies go “AI-first”. Every department in our businesses is being hit with the AI wand, some more successfully than others. Support is moving increasingly to AI-chatbots. Emails are being summarised. Documents are being spruced up. Spreadsheet formulas are being calculated. Automation is being handed off to AI. And software development? Every single CEO is now asking, “if OpenAI can have AI write 80% of their code, why can’t we?”. So, what’s our answer? We’re engineers and inventing, building, using, understanding, and evaluating technologies is what we do. We’re supposed to be curious about this stuff! When someone hands us a puzzle, we figure out how it works. It takes only a cursory use of an AI-coding tool to realise that there’s definitely some value here as a developer. Perhaps you ask Claude to brainstorm a new approach for processing webhooks in your application. Or you give it a SQL query and ask it to suggest optimisations. Or you just experience an inline suggestion in your editor that autocompletes exactly the next line you were about to type. There is no denying that there’s the hint of value here. How much? What’s the extent of it’s usefulness? How can we make the most of it? How does it fit into the wider human and technical systems of software development? <h2>Why sitting this out is not an option</h2> I know there are many sceptical developers out there seeing AI either as a threat or a fad. As I mentioned previously, there’s much not to like about the AI industry right now and we should be rightly concerned about many of the ways it’s being used and the consequent effects on society. I also think we can separate out those troublesome uses of AI from the AI tools in our own domain. We can simultaneously hold the view that AI videos threaten democracy whilst also believing that AI coding assistants can improve our work—they are not mutually-exclusive views. <h3>Know your enemy</h3> You can’t figure out the value or limitations of AI by ignoring it. You need to use it and explore it to really understand it and, after giving it a genuine attempt, you are left with two outcomes. In the best case, the AI helps you do your job. Win! In the worst case, you can start to push back against the self-interested AI-crusaders that are red-pilling our entire industry—not with vague thoughts about your “craft” but actual data, knowledge, and experiences you’ve gained using AI. If you’re an AI-sceptic, then I think it’s imperative you start to learn this new tool because you cannot adequately question something you haven’t spent the time trying to understand. Your arguments will only get stronger as you become an expert in the use of AI. We’ve seen over and over again that the output of AI systems can look very authentic and convincing to the untrained eye. Ask ChatGPT for a swim set and it’ll give you back one but a real swim coach will poke holes in it. Likewise, a recipe from Claude is perfectly plausible until you try to bake it. As software engineers, we are both the best placed to evaluate AI’s ability to produce software and the only ones that should do so. If we leave it to the non-technical person that just vibe-coded their first web app, or the CEO’s listening to the latest AI sales pitch, then of course all we’ll hear is how amazing AI is for coding. If we really want to know how good or bad it is, we need engineers to explore it. <h3>Game theory-ing your career</h3> You might like to wish for a world without AI. You might like to believe that if you can just ignore it for a few more months then it will all blow over and next year Claude Code and Cursor will be the Bored Apes of 2025: embarrassing, laughable, forgotten. I think you’re playing a game akin to the Prisoner’s Dilemma with your career. You’re assuming that if you don’t embrace AI, and no other developer embraces it, then it will die on the vine. Except, if other developers do embrace it and learn how to use it well, they will be more productive and more valuable than you are. A developer using AI well, will be more productive than a developer without it. A developer using it badly is worse than both. Who are you going to be? Who is a business going to hire? In the Prisoner’s Dilemma, if one prisoner defects (i.e., embraces AI), then the other prisoner is subjected to an even longer sentence. That’s going to translate to a massive disadvantage in our careers from not getting promotions, to poor performance reviews, to being the first on the firing list, and missing out on job opportunities. There is no universe where your colleagues and competitors in the job market aren’t going to utilise a tool if it gave them even a 10-20% productivity boost. Wishing for broad non-compliance is a losing strategy. <h3>Bubbles won’t save you</h3> Perhaps you’ve seen the absurd amounts of money being invested in AI, noted that Nvidia appears to basically be paying companies to buy its own product, and you’ve seen the huge losses these companies are racking up with each AI query their users execute. You think this is a bubble. You’re probably right. You think that when the bubble bursts, all this AI stuff will go away. You’re wrong. After the dotcom bubble burst, e-commerce did not die. At the same time in Europe, the telecoms crash did not scupper 3G networks (or 4G, or 5G, or the iPhone, or Android). The housing crash of 2008 didn’t kill off houses, mortgages, or banks. Those events culled the dead and defective companies but it didn’t off the industry entirely. The bubble will (probably) burst but the useful AI tools will survive. <h3>It might actually improve your (work) life</h3> I was a Java developer for more than 10 years before I discovered Ruby on Rails. That jump from verbose, type-ridden code to method-missing magic and duck-typing was weird. It was very uncomfortable to leave behind the safety of types, compilation, and large comment blocks. But through that discomfort I discovered a language which I adore and one that fulfills my goals much better than Java did. Equally, adopting Java in the first place back in ’97 wasn’t an easy path: no one was hiring for this very new, and very slow language. The easy choice at the time was C++. And over the years, we’ve all had to learn new editors, new libraries, new languages, new processes, new version control. Or perhaps you’re still young enough that you haven’t but I’m here to tell you that you will need to learn new things. That new thing today is AI. It <del>might</del> will be weird and uncomfortable initially but it might also help improve your work and many of the stressors that come with it. There is the concern that all this “progress” will put ourselves out of a job once the AI gets good enough. I’m sorry to tell you this (if you have investment in AI companies) but the AI is not going to get good enough. Our teams of developers are not going to be replaced with agents running 24/7 because the fundamental architecture of the systems is statistical not logical. The LLMs have no conceptual model of a system, or its dependencies, its flows, and its constraints. They are “just” a very sophisticated language prediction engine. It’s incredibly innovative to solve problems by providing the context in natural language, appending the question, then getting the AI to predict the appropriate answer. That’s a stunning development in human technology—but I think it has its limits and we may already be seeing a slow-down in AI progress. Without any conceptual understanding, I think there will always be a role for human minds. I may be biased because I have always run small teams that punch above their weight, working on systems that our competitors have teams multiple times our size to manage. We achieved this, frankly, by having to be more ruthless about the work we take on. There are are so many good ideas, must-do’s and should-do’s that we don’t do because we’re small and so we have to focus only on the very highest priorities. I don’t see AI being a threat to our team size in any way. I don’t think we’ll ever need fewer developers but I am enthusiastic about being able to tackle more of those good ideas. I’m hopeful that AI can help us achieve things that were previously too high-effort and lower down our priority list. That we can file off the rough edges which we’ve worked around for years. I’m actually really hopeful that AI can lower the cognitive load for the team. I wrote recently about how I consider that to be one of our biggest scaling challenges and it’s something that concerns me a lot. A pair of developers might be working on a new feature for a few weeks/months and become familiar with it but it’s a massive workload to then jump to fixing a bug in another area of the app. Can AI, like Neo in The Matrix, load up that context into our brains faster? Can it help us summarise and map out the codebase? Can it remind us about a feature which is dependent on the current behaviour? I think there is a massive opportunity beyond “AI writes code”. I think constraining your thinking to AI-generated code is frankly, Level 1 thinking. It’s “mid”, as the kids might say. We never hired the programmers with the fastest typing skills (at least, not since the 1950’s). We have never felt like what has held our product back is the lack of a faster typist—because typing is not the bottleneck. And frankly, code is not the end goal either. We participate in a system of software development. Part of that system is the physical typing of the code into the machine but it’s not even the hard part or interesting part. The hardest bits are knowing what to type: whether that’s product decisions, or architecture decisions, or decisions around maintainability or performance. This is where the magic happens. In our heads, not our fingers. </div> Fri, 24 Oct 2025 00:00:00 UTCYou are the scariest monster in the woodshttps://jamie.ideasasylum.com/2025/10/15/you-are-the-scariest-monster-in-the-woods<article class="mx-auto sm:mx-2 max-w-prose" morss_own_score="7.0" morss_score="11.5"> <h1>You are the scariest monster in the woods</h1> <img src="https://jamie.ideasasylum.com/2025/10/15/you-are-the-scariest-monster-in-the-woods"> <div class="prose dark:prose-invert prose-slate mt-2 sm:px-8 font-serif" morss_own_score="3.0" morss_score="38.0">I don’t really believe in the threat of AGI (Artificial General Intelligence—human-level intelligence) partly because I don’t believe in the possibility of AGI and I’m highly skeptical that the current technology underpinning LLMs will provide a route to it. But I also think there’s something we should actually be afraid of long before AGI, if it ever comes. When talking about humans in any context, whether it’s us vs sharks, or us vs wolves, or us vs <insert anything that scares us>, and especially when it’s us vs <some technology>, I am reminded of The Gruffalo. In this children’s story, a timid mouse convinces a scary monster (The Gruffalo) that the mouse is the scariest animal in the woods. In reality though, the forest animals are frightened of the Gruffulo with it’s terrible tusks, and terrible claws, and terrible teeth. <img src="https://jamie.ideasasylum.com/images/gruffalo-ai.jpg"> We (humans) are the scariest animal in the woods. We’re the scariest animal anywhere. At any time, in any location, under any circumstances, if there’s a human present then that’s the scariest motherfucker in the woods. We’re a danger to every living thing, ourselves included. Our collective ability to survive, adapt, control, kill, or wipe out any other species is unmatched. Anyone trying to tell you otherwise is trying to distract you. AI is not the monster to be afraid of; we are. Just like a hammer, sword, or a rifle lying on a ground is nothing to be feared, so too is AI. It’s just an inanimate object; a tool, potentially. Now, if you equip humans with a hammer, or sword, or rifle, or AI then you’ve just made the scariest monster in the woods (that’s you) even more terrifying. To bring this around to more concrete thoughts: I do not believe that AI will enslave us, destroy our democracies, or our environment, or rob us of our skills, our purpose, or our jobs. That’s what humans will do. What we’ve always tried to do. We don’t need to worry about AI itself, we need to be concerned about what “humans + AI” will do. Humans will do what they’ve always tried to do—gain power, enslave, kill, control, exploit, cheat, or just be lazy and avoid the hard work—but now with new abilities that we couldn’t have dreamed of. I’m not saying “don’t worry”. I’m saying don’t worry about the technology of AI and continue to recognise that humans, and how they will use a technology, has always been our biggest threat. The means accepting that AI isn’t something passively happening to us. It’s not a meteor heading towards Earth that we play no part in it or have no control over. It’s a thing we’re building, for other humans to use. We’re not building AI for gerbils here. How would we like other humans to use AI? How would we not like them to use it? We’re building and using this technology but pretending that unlike cars, and guns, and knives, and nuclear weapons, that we are powerless to understand, control, or regulate it for the betterment of humanity. The scariest monster in the woods just got scarier and we can’t ignore that. </div> </article> Wed, 15 Oct 2025 00:00:00 UTC