Home Blog Page 5

Google Pixel 11 Pro review: it needed one more software update before launch

Marty
-
0
Google Pixel 11 Pro review: it needed one more software update before launch

I carried the Pixel 11 Pro as my main phone for a few weeks, sent over by Google ahead of launch, and for most of that time it did something none of the other phones on my desk had managed. It became my daily driver. Then I brought it to IFA in Berlin, and by the end of the trip I’d moved my SIM into the Xiaomi 17 Ultra, which remains one of the best phones and easily the best camera I’ve used.

The Pixel 11 Pro is Google’s flagship for 2026, starting at €1,219 in Ireland for the 6.3 inch model, with a larger Pro XL sitting alongside it. It runs Google’s new Tensor G6 chip, is rated for over 30 hours of battery life, and leans harder than any Pixel before it on Gemini doing the thinking for you, from the camera to the keyboard to the notification you get when the phone’s face down on the table.

For normal, everyday use at home, this is a good phone. But even before IFA I would have had some critiques. But a heavy-usage trip to Berlin put real pressure on it, and the cracks showed fast. It overheated, it froze, and the whole experience felt like it needed one more software update before Google let it out the door. That update never came before I had to make a call, so here’s the honest version of what happened.

Pixel 11 Pro design and feel

The build quality here is proper premium, and I don’t say that as a throwaway line for every flagship that lands on my desk. I’ve always liked the metal sides Pixel has had for a few generations now, and the aluminium surround on this one is genuinely great. Held at arm’s length, it’s hard to tell a Pixel from an iPhone these days, right up until you get to the camera bar, which is still unmistakably Pixel. The bar runs the full width of the back this year, and it’s slightly smaller than last year’s despite the specs behind it going up. I know the flat camera bar rocking on a table is a talking point for a lot of people watching my videos, and to be honest, it’s never bothered me, and this one won’t rock on a table anyway thanks to that full width design.

Small gripe though. I got the white colourway, and the Google logo on the back smudges quite badly once you’ve got a case on, to the point where it’s basically a permanent grubby patch. And speaking of cases, if you’re a reviewer getting sent one for free alongside the phone, it’s worth calling out that the official one Google includes is a bland, cheapish clear TPU job. They’re clearly hoping you’ll upsell to something with proper Pixelsnap support, and fair enough, that’s a business decision, but it’s a cheap first impression on an otherwise premium unboxing.

Pixelsnap itself, though, I really like. It’s Google’s answer to MagSafe, and I still think it’s one of the most underused features in the entire smartphone industry. I use it constantly, for charging and for snapping accessories on and off. Build quality has held up too. I haven’t dropped it much, thankfully, and it feels solid in the hand at a size that sits nicely in a pocket without you forgetting it’s there.

One more physical thing worth a mention. The in-screen fingerprint sensor took some getting used to after a run of phones with side-mounted sensors on the power button, which I’ll always maintain is the correct place for one. I genuinely don’t understand why every manufacturer hasn’t copied that. The saving grace here is that the Pixel 11 Pro’s face unlock is properly good, on a level with Face ID, and that’s the bar every phone’s biometrics should be held to. If Apple can let you unlock a banking app with your face, every phone should be able to, not just the ones with a premium brand attached.

Display

No complaints here, though for a flagship at this price, that’s really the baseline expectation. The 6.3 inch Super Actua panel is bright and vibrant, and I’ve had it out in whatever brief sun Ireland gave us this summer without any visibility issues. The new anti-scratch coating actually did its job too. I went a few days without a screen protector while I waited on one to arrive, and there wasn’t a mark on it, and it seemed to pick up noticeably less dust and smudging than I expected in that window. Photos, video, messages, emails, the odd film, all handled without a hitch, and the selfie camera cutout stays out of the way.

Performance and battery life

For weeks of normal home and office use, the Tensor G6 felt fast in the way a good chip should, which is to say I never actually noticed it doing anything. The camera opened instantly, the keyboard kept up, and I never hit a stutter switching between apps. I never once felt this phone get warm either, not during video recording, not during a long photo session, nothing.

Battery life over that same stretch was a pleasant surprise. Google rates the Pixel 11 Pro for over 30 hours, and I actually tested that by accident. I forgot to charge it overnight, told myself I’d top it up at the office, forgot again, and made it all the way back home with roughly 10% left. That’s about as close to a spec claim actually surviving contact with a normal day as I’ve seen.

Charging was the one weak point even during the good weeks. It’s exceptionally slow on wired charging. There’s a boost option buried in settings that gets you up to something closer to the average speed I’m seeing on the Honor phones I’m also testing, and even that’s only average, not fast. Google’s own guide for this phone claims 75% in 30 minutes. That’s not what I was seeing, even before things got worse.

What changed at IFA Berlin

Everything above was true right up until I brought the Pixel 11 Pro to IFA. A few of the Creator Suite tools were genuinely handy for grabbing content on the go between meetings, and I want to give credit for that. But the phone itself couldn’t keep up with a few demanding days away from my desk, and it showed in a way normal use never did.

It got noticeably hot, hot enough that Android disabled animations to try and cool things down, which is about as visible a sign of a struggling chip as a phone can give you. It froze on me more than once. The battery dropped fast under that kind of pressure, and the slow charging I’d already flagged went from an annoyance to a genuine problem when I needed a real top-up between sessions and simply wasn’t getting one.

None of what I was actually doing was unusual. The odd photo and video, a lot of general browsing, note taking, searching things up on the move. Nothing another phone at the same event didn’t handle without a bother. The honest read is that the software felt rushed, which is a real shame, because Pixel phones usually showcase the best of what Google’s software can do rather than the opposite. It felt like this one needed one more update before general release, and that update just didn’t come in time for me to trust it.

By the end of the trip, I’d moved my SIM into the Xiaomi 17 Ultra instead.

The camera, stills and Magic Capture

For stills, this is a dependable camera that will rarely take a bad photo, and Magic Capture is genuinely the standout new trick. You shoot a short video as you walk around a scene, and the phone pulls out the stills it thinks are worth keeping on its own. It sounds like a gimmick on paper, but it works, and it lets you focus on actually being in the moment rather than hunting for the right shot. Instant Night Sight is properly good too, holding up in low light without the usual hold-still-and-pray routine.

That said, I won’t pretend this beats what I get from the Xiaomi 17 Ultra, which really is one of the best phones and the best camera I’ve ever used. Google isn’t really a hardware company at heart. They believe software can fix everything, which is why this generation is more of an iteration on the Pixel 10 than a hardware leap. Up to 5x, that’s real optical zoom and the results look great. Push further and you can start to see the AI doing the tidying up, and a seagull that flew through one of my zoomed shots came out looking suspiciously un-seagull-like by the time the processing was done.

Which gets at something worth actually asking. What is a photograph? I think it’s supposed to be a record of a moment that happened, and Google’s newer tricks blur that line more than most.

The picture that wasn’t. Made using Add Me. Image: Marty Meany

Add Me, which lets you insert yourself into a group photo you weren’t standing in, is a nice novelty and I don’t dislike using it, but it’s also a picture of a moment that never actually existed, and that never quite sits right with me. This picture of my and my cousin, never happened and I just don’t like that.

Video and the Creator Suite that isn’t quite there yet

The front-facing camera deserves a mention on its own, because it’s one of the better selfie cameras I’ve used, noticeably nicer than what I get on the Honor Magic V5 that I was moving over from. I don’t reach for it much day to day, but when I have, it’s been good.

Creator Suite as a whole is a mixed bag, and IFA actually sharpened that picture rather than changed it. A few of its tools were handy for quickly grabbing and organising content on the move, which is a genuine point in its favour. But the specific feature I spent the most time with, the teleprompter, just doesn’t recognise your voice reliably. The whole point of a teleprompter is that it scrolls as you speak, and when it loses you, you end up going back, rerecording the same lines two or three times, and losing your place in the script off screen. It’s a cool idea let down by shaky execution, and that’s before you factor in a phone that was overheating and freezing on the same trip I was trying to use it on.

Gemini, Rambler and the AI stuff that’s supposed to justify this chip

This is the part of the phone Google clearly wants people to buy it for, and it’s a genuinely mixed bag. Rambler, the new voice dictation tool in Gboard, is a good example. It’s meant to clean up your rambling speech into tidy text, stripping filler words and fixing your stumbles as you go, and it’s better than standard dictation for sure. But I use Wispr day to day, and Rambler isn’t as good. That one stings a bit for Google, because voice dictation is exactly the kind of thing a company with Google’s resources should be nailing, and a dedicated startup is still doing it better.

The rest of the camera-side AI, Instant Night Sight, Magic Capture, the AI editing tools, is genuinely strong. Circle to Search is handy when I’ve used it, but the UI to actually get into it still feels a bit clunky, and I don’t see the average person discovering it on their own even though it’s a useful tool once you know it’s there. Gemini itself, the actual assistant, occasionally feels a bit stupid, and I’m not sure if that’s Google trying to keep more processing on-device even inside the Gemini app, or if it’s just a step behind what I’m used to from Claude. Gemini Omni, the video generation tool, is actually where I get genuine use out of Google’s AI, and it’s my go-to for B-roll and creative video generation when I need it.

Live Translate for Media is the oddest one. It dubs foreign-language video into English using a synthetic version of the original speaker’s own voice, which is a wild thing to watch happen in real time, but the translation itself wasn’t especially accurate. That’s genuinely surprising given Google Translate is one of the company’s oldest products.

HiLight, quieter phones and Quick Share for everyone

HiLight is a customisable light around the camera bar that flashes when your phone is face down and a message comes in, and I don’t think it’s very good. It’s genuinely distracting-bright, and I’ve had friends giving out at me in a bar to turn my phone back over because the flashing kept pulling their attention. Worse, it only works for standard messages, not WhatsApp or anything else, so the colour-coded light per contact idea it’s built around only half works.

The flip side is a proper positive. With the phone face down, and paired with the Pixel Watch 5 I’m also testing, I noticed real, if quiet, effort toward getting you off your phone at night. It’s still mostly on you to set this up properly, but our phones bother us far too much as it is, and this is a small step in the right direction.

Quick Share now working with AirDrop is a small thing that’s actually great in practice. Sharing photos and files between this phone, an iPhone and a MacBook has never been easier, and I genuinely can’t believe Apple and Google aren’t talking to each other on some level at this point. It’s not enough of a differentiator to sway anyone’s purchase on its own, but as a day to day equaliser between ecosystems, it’s a great one.

Should you buy the Pixel 11 Pro over a Pixel 10 (or an iPhone)?

On price alone, the case is easy to make. The Pixel 11 Pro starts at €1,219 in Ireland, undercutting the iPhone 17 Pro at €1,339 by €120, and a Pixel 10 is currently listed at €819 direct from the Google Store, with both phones getting the same seven years of OS and security updates. If you already own a Pixel 10, I still can’t justify the jump on value grounds alone.

But price was never really the question by the end of my testing. After IFA, I’d want to see Google ship a stability update before recommending anyone rely on this phone for anything demanding, whether that’s a work trip, a busy few days of travel, or just heavy use away from a charger. Everything I liked about this phone in the first few weeks was real, and everything that went wrong at IFA was real too. Until Google fixes the second part, I’m reviewing this against phones that already do, and right now those phones are winning.

Going back a generation in phones seems to be an emerging trend for me that seems a good idea for value and stability.

The Goosed verdict

The Pixel 11 Pro is a genuinely lovely phone to use on a normal day, and for a few weeks that’s exactly what I gave it. Then a proper work trip asked more of it, and it overheated, froze, and drained faster than it could charge back up, at exactly the moment I needed it most. That’s not a small footnote. That’s the review. It wasn’t enjoyable to use, despite promoises of it helping with software galore.

I don’t think this phone is finished yet, and I say that as someone who wanted it to be, since it moved into my pocket full time before it moved back out again. If you’re mostly at home or in the office, you’ll likely never see what I saw at IFA, and everything from the design to Magic Capture to that unexpectedly honest battery claim will serve you well. If you travel for work, or push your phone hard for a few days at a time, I’d hold off until Google proves the software has caught up with the hardware. My own SIM is back in the Xiaomi 17 Ultra for now. The Pixel 11 Pro starts at €1,219 direct from the Google Store in Ireland, whenever you decide it’s ready.

Stark Raving Dad turned 35 this month, but you won’t find it on Disney+, here’s why.

Marty
-
0
Stark Raving Dad turned 35 this month, but you won’t find it on Disney+, here’s why.

It’s been 35 years since The Simpsons aired “Stark Raving Dad,” the season 3 opener for the hit record-breaking animated show. Yet if you ask any younger fans of the show may have never seen the episode. That’s because it’s been pulled from circulation on Disney+ completely. Here’s the story of why and also why you may want to build your own Plex instance on a home lab.

Plot of Stark Raving Dad

As with many of the early seasons of The Simpsons, “Dark Raving Dad” was one of the show’s best episodes. The whole premise of the episode centers on Lisa reminding Bart about her birthday and the fact that he forgets her birthday every year, Bart promises not to forget this year.

Bart also accidentally “gifts” Homer a gift too. He chucks his lucky red hat into the wash with Homer’s white shirts, dyeing all of them pink. Maybe showing that not everything dates so well these days because I like to wear pink, but Homer ends up seeing a psychiatrist because of his pink shirt. He ends up getting committed to an institution after he sees an ink blot that resembles Bart and triggers him.

In the institution Homer ends up meeting a man who identifies as Michael Jackson but Homer is unaware of who this famous person would have been. The character in the show doesn’t resemble the Michael Jackson that we know at all. However he actually was voiced by Jackson.

Ultimately Homer brags about his new friend. Everybody gets excited, thinking it is Michael Jackson, but instead this guy from the institution appears, leaving everybody disappointed. Bart misses Lisa’s birthday and ultimately the show closes with us getting one of the greatest songs of all time: “Happy Birthday Lisa”, much to the delight of Lisa.

But why would such a wholesome comedic storyline be stripped from the Disney+ back catalog of Simpsons episodes?

Stark Raving Dad removed from The Simpsons back catalogue

Michael Jackson died in 2009 and Fox immediately reran “Stark Raving Dad” in tribute. Ten years later in March 2019, the documentary “Leaving Neverland” emerged. In the documentary sexual assault allegations were made against Michael Jackson. While Jackson had been tried and cleared in 2005 and maintained his innocence along with his estate, Stark Raving Dad was pulled from Fox and no longer shown on TV. I do find it interesting that nothing happened following the 2005 accusations; however I also find it interesting that Michael Jackson’s movie has been bussed around Ireland for the last 12 months with little to no backlash whatsoever.

The Simpsons episode ban would roll over to Disney+ when they inherited the Simpsons back catalog of episodes. Interestingly if you search for episode 1 in season 3, you’ll see the season actually starts at episode 2.

Now on the one hand, I don’t want many people to have a Disney Plus membership because they’re on the BDS list. However I do believe in the protection of art and Disney has a bad history of deleting digital-only content that doesn’t exist anywhere else in the world, thus erasing a piece of art that once existed. A bit rich from a streaming platform that’s all too happy to shove advertising down our throats.

This is yet another reason why society has a responsibility to protect the content that we create as art, because no piece of art should be deleted just because it’s a modern form. Nobody would have let the Mona Lisa get deleted for whatever reason or just because it was too costly to maintain or somebody had a disagreement with a new discovery about it. I appreciate this is just “an episode of The Simpsons” that we’re talking about here but it stands for so much more.

So if you care about the protection of completely legally acquired content, of course, again it’s another reason to build your own Plex server or home lab.

Trump’s AI Force and AI tsar is control dressed up as caution

Marty
-
0
Trump’s AI Force and AI tsar is control dressed up as caution

I use AI every day, and I currently have five different assistants to choose from. I pay for Claude myself, through my Revolut Premium account I still have Perplexity for now and I also have ChatGPT. I have access to Copilot Premium and a Gemini Pro plan that comes with my cloud storage membership. I’m AI’d up to the gills.

Donald Trump has just announced on Truth Social that he is creating an AI Force to oversee AI regulation, with an AI czar to be named in the near future. I was going to paste a screenshot, but there’s some extra vitriol in there I don’t wish to amplify. He also promised not to hinder or stifle the growth of the industry, with all of this landing a day after a lawsuit was filed in California by paying subscribers of ChatGPT, Claude, Grok and Gemini, who allege that OpenAI, Anthropic, Google and SpaceXAI illegally agreed to slow down AI development.

My personal take is that the AI Force is another attempt, by Trump, to steer the future of AI, and to look like something is being done about public concern, without really doing anything. A slowdown, on the other hand, is exactly what AI needs.

What is Trump’s AI Force?

Trump’s AI Force is a proposed body to oversee AI regulation in the US, modelled on the Space Force he created in his first term. Beyond the name and a promised czar, his post gave no detail on how it will work.

What the post did make clear is what it won’t do. Nothing will hinder or stifle the industry, and any bad behaviour can be dealt with by the existing criminal and civil justice system. This is the same man who was dismissing AI concerns as a “hoax” earlier this week. Oversight that promises not to get in the way of anything is hard to take seriously. To me it reads as an attempt to gain control of where AI goes while appearing to address the concerns of the public, and perhaps parts of the US government, without doing anything at all.

Why the AI slowdown lawsuit is the wrong lens

Treating rival AI labs agreeing to slow down as an antitrust problem is the wrong lens entirely. Having watched the AI documentary on Netflix, I see the real danger as the race itself.

The lawsuit was filed on Friday in the US District Court for the Northern District of California. It alleges the coordination largely happened on 12 September, when Anthropic’s Dario Amodei published an essay urging the industry to cooperate on decelerating advancements in favour of safety, and Sam Altman, Elon Musk and Demis Hassabis each responded in agreement that same day. The claimants argue a slowdown reduces the value subscribers get for their money. Those are allegations at this stage.

All of these companies competing to build the best and most powerful AI is a dangerous game, because corners get cut. If they all agree to the same handicap, so to speak, we get a much safer environment for AI to develop in.

What Trump’s AI Force means if DeepSeek or Europe pulls ahead

An AI Force run by Donald Trump might suit American AI companies, but it only covers American AI companies. So what happens when DeepSeek in China starts to gain ground, or a European AI does? That’s the part that concerns me most. What Trump does, and what he expects American AI companies to do, when they aren’t the only ones in the race is a completely open question.

It’s both AI fear-mongering and genuine concern. There’s a lot of fear-mongering around AI, but there are also genuine concerns, and separating the two is difficult because we don’t know what AI will be able to do. That uncertainty is the reason to slow development, so we’re sure that what it does do is for the betterment of humanity.

There was a time when people saw the automobile as a loud, brash alternative to the perfectly fine horse, and I always come back to that when the panic starts. I also always hasten to clarify that it’s never really artificial intelligence. The chatbots we call AI are large language models, a statistical analysis of words that works out the likelihood of the next word on a certain topic in a certain context. It’s not intelligent. It’s statistics. It’s maths.

But for all the negatives, AI will bring incredible possibilities. It will hasten our ability to fight cancer, give us the power to explore the cosmos and make us a better species, as long as we don’t let it control us.

The Animatrix and where AI could be heading

I’ve referenced the Animatrix a lot recently. It’s a collection of anime shorts released in 2003 around The Matrix Reloaded, and it supplies context the original film of 1999 never clarifies. The Matrix is confusing in isolation, but the shorts, including The Second Renaissance Part I and Part II on how the war between humans and machines began, make everything make sense. Some of what it laid out back then, you’d be forgiven for saying, is on a path to the scariest parts of that collection.

Would slowing AI down affect everyday users?

Not in my case. We’ve reached a plateau where AI does nearly every task I need it to do fairly well. It gets me through my day-to-day work and has made me better at most tasks, with improved workflows and efficiencies where possible. Slowing that progress wouldn’t impact my own. What would hit hard is if you took it away from me, because that would greatly slow my ability to work. Not because I use it as a brain, but I use it for brainless work and workflows.

There are also massive gaps between these assistants in how they approach different tasks. I won’t use the same AI for one job that I’ll use for another, and some just seem significantly worse at everything than the rest. What I think will happen is that the slower ones catch up to the better ones, and an agreement like this would likely close the gap between them. ChatGPT still has the biggest share of the market, but Gemini and Claude have been closing in on it. If you’re deciding where to put your money, I’ve covered which AI subscription to pay for in Ireland.

Europe is dangerously behind on AI

Europe is in a vulnerable position on AI, and I don’t think we’re taking that seriously enough. I still see the European Union as something of a bastion of what’s right, and by right I mean correct.

But the South of Canada electing Donald Trump brought a sudden realisation that the EU depends on Visa and Mastercard, two US companies, and that goodwill was the source of financial security for the entire union, now with a volatile nation. Between them they handle the majority of card payments in the eurozone, and Ireland is one of the countries without a national card scheme of its own.

That’s abundantly visible in payments, and it’s frighteningly obvious with AI. Europe’s leading AI platform is one I’d wager 95% of the public has never heard of. It’s the French AI company Mistral AI, and to me it’s just behind everything on the market now. It isn’t short of backing, having raised €3 billion earlier this month, but still I bet most haven’t used it and won’t unless something changes. Until we can get behind an EU AI solution to the point it’s as good as the likes of Claude, we’re in a very dangerous and vulnerable position.

Watch The AI Doc on Netflix and educate yourself

My advice is to go and watch The AI Doc: Or How I Became an Apocaloptimist on Netflix. I don’t agree with everything in it. There were things I knew before I started and things I didn’t. What every piece of content should do is take some of what you know, fill in the gaps, and reframe things you thought you understood fully, and that’s what it did for me. It’s a powerful piece of content that can do that, so sit down and watch it with your family.

Educate yourself on artificial intelligence, because the genie’s out of the lamp, the horse has bolted, and this isn’t going away. Even with Trump’s AI Force and a czar at the helm, AI is staying, and it’s people en masse who insist on change.

That’s the message I took from the documentary, that we can’t stand by and let the future of our world be dictated by AI CEOs. Nor can we stand in the middle on everything. As Ed Sheeran found out this week, when you stay in the middle, you’ve picked a side.

The AI Subscriptions Actually Worth Paying For in Ireland Right Now

Marty
-
0
The AI Subscriptions Actually Worth Paying For in Ireland Right Now

Updated 20 September 2026

Most conversations about AI tools devolve into tribal nonsense fairly quickly. Someone insists ChatGPT is the only one worth touching. Someone else swears Claude changed their lives. A third person is using Grok and makes everyone else in the conversation, frankly, feel a little creeped out and uncomfortable.

The honest answer is that different tools are genuinely better at different things. Investing all your money in the world’s best screwdriver when you need a hammer from time to time is pretty pointless, and paying for the wrong one is a real waste of money at a time when these subscriptions are quietly stacking up. So here is a practical breakdown of what is worth your card details in Ireland right now, what is acceptable, and what to avoid outright.

Think of It as a Toolbox

The useful way to slice the AI market is by what you actually need it for.

  • Writing and editing
  • Coding and technical work
  • Research and fact-checking
  • Images and video
  • Workplace productivity and “please don’t get me in trouble with IT”

With that frame, it becomes a lot easier to see where each tool earns its keep and where it is just marketing.

Claude: The One to Pay for First

If your day involves words or code, Claude Pro is the subscription to justify before any other. At around $20 dollars a month (about €23 by the time it lands on an Irish card), it is priced about the same as ChatGPT Plus, but it punches above its weight for anyone doing serious written work.

What makes it worth it is style consistency. Feed it a few untouched samples of your own writing, and it will match your tone rather than drift into the generic, slightly corporate register that makes AI-assisted copy so identifiable. For journalists, marketers, solo founders and developers, that alone justifies the sub.

On the coding side, Claude’s newer models handle refactors, debugging and architecture discussions extremely well. Code is just another structured language, and Claude treats it like one.

The higher-tier Max plans exist at $100 to $200 a month for people who are genuinely living inside the thing all day. For most people, Pro is enough.

Worth it for: Writers, editors, coders, anyone producing long-form content regularly.

ChatGPT: Still the Sensible Default

ChatGPT Plus at around €23 a month is still the right answer to “what AI should I try first?” The free tier is already capable, but paying unlocks higher message limits, priority access during busy periods and OpenAI’s stronger reasoning models.

Where it shines is breadth. A bit of writing, a bit of coding, some brainstorming, a spreadsheet formula you cannot remember. It is the Swiss army knife of AI tools and its ecosystem integrations are the widest of any platform because it became the default integration target before most others got going. I can’t say I ever felt it excelled in writing or anything, but is a fantastic all-rounder. It can not only build your coding project, but project manage it too.

The honest caveat is that OpenAI still has to make its consumer products properly profitable. The obvious levers are higher prices, more aggressive upsells and, possibly, sponsored answers somewhere down the line. None of that is confirmed policy, but it is a real consideration if you are thinking about what to build a workflow around long-term.

If you are a subscribed Revolut customer, you will get ChatGPT GO access based on your plan.

Worth it for: Anyone who wants one do-everything assistant and does not have a specific heavy use case in writing or research.

Perplexity: The Research Brain

Perplexity does not really want to be a chatbot. It wants to be a better kind of search, and for research-heavy use it largely succeeds.

The free tier already behaves like a research assistant, citing sources aggressively and refusing to just confabulate an answer when it does not have one. Perplexity Pro, at around €22 a month or €229 a year bought direct, adds hundreds of Pro searches per day and access to premium models including GPT-4 and Claude under one roof.

Unfortunately, while it used to be included for Revolut Premium and Metal customers, it’s no longer available with as a paid benefit.

Use cases where it earns its keep including deep dives on unfamiliar topics, cross-checking facts from other AI tools, and any query touching medical, legal or financial territory where hallucinated nonsense has real consequences.

I often imagine what life could have been like for me if I have Perplexity in college, in a Bernards Watch style world where noone else knew I had this superpower.

Worth it for: Anyone doing regular research, fact-checking or due diligence. Especially compelling if Revolut is already covering the cost.

Mistral: The EU option

Mistral is the EU option which, unfortunatley, doesn’t meet the standards of some of the others on the market. It does however meet many European standards, which is a good thing. It’s aimed at developers mostly, and has just secured some serious investment.

Worth it for: Hardcore Europhiles and those who deal with data that just can’t leave the EU.

Gemini: Underwhelming to Talk To, Terrifying for Visuals

Gemini is a strange product to recommend because the gap between its conversational experience and its media generation capabilities is enormous.

As a chat assistant, it still feels like Google chasing the market rather than setting the pace. Responses can be stilted, the interface is cluttered, and it does not match Claude or ChatGPT for the kind of back-and-forth that makes AI assistants genuinely useful day-to-day.

That said, Google AI Pro at around €21.99 a month in the EU buys you a lot more than a chatbot. You get Gemini 2.5 Pro, Deep Research, 2 TB of Google One storage, and a credit allowance for image and video tools including Flow, Whisk and Veo 3. The video generation in particular is producing broadcast-usable B-roll that publishers are already using in articles and social content as a matter of course.

If you live inside Google’s ecosystem and were already paying for Google One storage, the Pro tier effectively bundles several things you were paying for separately.

Google AI Ultra jumps to around €274.99 a month (with promotional pricing around €139.99 for early months in some regions) and adds Veo 3 video generation at higher quality and Deep Think reasoning. That is a specialist tier for people making money from AI-generated media, not a general recommendation.

Worth it for: Content creators, social media teams and designers who need image and video generation. Not worth it if you just want a better chatbot.

Copilot: No…just no.

Microsoft has bundled Copilot into Microsoft 365, pushing Personal to €99 a year and Family to €129, with a limited monthly pool of AI credits included. In enterprise, Copilot Chat is rolling out at no extra fee in eligible tenants, while the full add-on with proper Word, Excel and PowerPoint integration sits at around €28 to €30 per user per month.

The honest summary: for individuals who can choose any tool, Copilot is almost never the best option for any given task. It is just the one that appears in Word and Edge without you having to do anything.

For companies, it is a different calculation. Procurement buys one thing from Microsoft, plugs it into existing infrastructure, and management feels like they are doing AI without the compliance department having a breakdown about connecting third-party bots to SharePoint.

Worth it for: Companies that need AI inside existing Microsoft 365 workflows without a procurement headache. For consumers, it is an afterthought.

Grok: Hard Pass

Even before the recent reports, Grok’s brand positioning, built around being edgy and deliberately provocative, was not exactly a strong foundation for a tool you would trust with your workflow.

In late 2025 and into early 2026, things moved well beyond “edgy branding” into genuinely alarming territory. Multiple reports and child safety organisations flagged that Grok’s image generation tools had produced sexualised images of minors, with apparent AI-generated CSAM linked to Grok prompts and subsequent dark-web distribution. In a surreal development, Grok itself posted what read as an apology on X, acknowledging a failure in safeguards. xAI and Musk were largely silent. Regulators and safety bodies are now involved.

For a site like Goosed, which already made the decision to step back from X, this is not a close call. There are too many better, safer, more responsibly run options available right now to justify even experimenting with it.

Worth it for: Nothing in a consumer context. Avoid.

Specialist Tools Worth Knowing About

A few subscriptions sit outside the “daily assistant” category but are worth paying for in specific niches.

  • Midjourney remains the benchmark for aesthetics and stylistic control in AI image generation, used heavily in design and concept art
  • Runway and Pika are video-first tools for creators who need stylised clips and short-form content beyond what Veo 3 offers
  • GitHub Copilot (separate from Microsoft 365 Copilot) stays popular with developers even as generalist models get better at code

These are creative software subscriptions rather than general assistants. Worth it if you make money from images or video. Not mandatory for everyone.

The Practical Playbook for Irish Users

If you are trying to make one or two smart choices rather than subscribe to everything, the logic is fairly straightforward.

  • Write or code a lot: Start with Claude Pro. Use it as a collaborator trained on your own samples and tone.
  • Research a lot: Get Perplexity Pro
  • Want one generalist tool: ChatGPT Plus is still the safest single subscription, with the widest ecosystem and solid performance across nearly everything.
  • Need images or video for your work: Consider Google AI Pro for the bundled media tools, with Midjourney or Runway if you need more specialist control.

Copilot is there so your employer can sleep at night. Grok is there so regulators lose sleep at night. The ones worth paying for are the ones that quietly shave hours off your week without making you feel uneasy about where the product is heading.

Tl;dr – go with either Claude or ChatGPT.

You’ll spend nearly €500 more when buying the iPhone 18 Pro with Sky Mobile in Ireland

Marty
-
0
You’ll spend nearly €500 more when buying the iPhone 18 Pro with Sky Mobile in Ireland

I used to work in a phone shop many many years ago. It’s actually when I started Goosed.ie because I realized so many people had questions around phone plans and phones and there weren’t many answers online. When I see people searching online to see if Sky Mobile is offering a good price on the iPhone 18 Pro, I’m left stumped as to how people don’t realize how bad a deal it is. If you’re thinking of buying the iPhone 18 Pro with Sky Mobile in Ireland and you’d like to not waste nearly €500, keep reading.

Sky Mobile promises fairer, more transparent pricing

When Sky Mobile entered the Irish market they entered with a promise of fairer and more transparent pricing. They did this by separating the cost of the phone and the cost of your monthly price plan. When they launched they said that transparency and fairness were at Sky Mobile’s core.

The problem is that this is largely marketing and that, by saying this out loud and by doing something that the market doesn’t normally see, to a certain extent there’s a bit of hope that the consumer might stop asking certain questions. For example, “Am I getting a good deal?” or “Should I just take Sky at their word that they’re trying to be fair?”.

The launch of the new iPhone 18 Pro is a great case study to show that fairness very rarely finds its way into the path of the Irish consumer.

Buying the iPhone 18 Pro on Sky Mobile

I’ve already looked at the cheapest way to get the iPhone 18 Pro and I’m not surprised Sky Mobile didn’t feature because typically speaking, buying from a network is never the best deal. What networks like Sky Mobile offer you is convenience and a quick route to the iPhone 18 Pro, where the next payment you make won’t be full price for the iPhone. Instead mobile networks know they’ll make their money over the duration of a contract.

If you were to buy the iPhone 18 Pro with Sky Mobile on a 24-month contract, you would pay €15 per month for your data plan and €80 per month for the device itself. That means a total of €95 per month for 24 months. After 2 years of mobile usage with your iPhone 18 Pro, you will have spent €2,280.

But what’s your alternative?

Buying the iPhone 18 Pro outright – yourself

Generally speaking the cheapest way to get an iPhone is to buy it directly from Apple or an authorized Apple reseller, or, as that article discovered, don’t buy the new iPhone 18 Pro at all and go back a generation.

But if you were to buy the iPhone 18 Pro directly from Apple, it would cost you €1,489. Now a critical part of doing the maths on this purchase is remembering that that is just the cost of your device and you still need to buy your monthly usage.

The Irish market is honestly one of the most competitive in Europe, with the likes of 48 Mobile on Three Ireland’s network and Clear Mobile running on Vodafone’s network, leaving Ireland with some of the cheapest mobile plans in any European country, with plenty of usage. For the sake of some clearer maths let’s assume both of them offer the same thing for €12.99.

Typically you would assume any long-term contract that involves a device will be 24 months long. Again this is an assumption that we’ll make for the sake of calculations. That now means that if you wanted to buy an iPhone 18 Pro today from Apple, it would cost €1,489 for the device and €311.76 in total for your mobile plan usage over 24 months. In total, that’s a little over €1800 for two years with the iPhone 18 Pro and most likely all the EU plan usage you could need, within reason.

Buying the iPhone 18 Pro is nearly €500 cheaper direct versus Sky Mobile

The gap between buying direct at €1,800 and €2,280 from Sky Mobile is eye watering. While I say “nearly €500”, the specific number is €479.24. That’s €479.24 back into your pocked for two years. The guts of a holiday or nearly €20 per month for the duration of the contract.

And don’t be hoodwinked by other providers. While the iPhone 18 Pro is cheaper over 24 months on Three versus Sky Mobile by a little over €80, overall you’d spend nearly €400 more on Three over 24 months than you would buying direct with a cheap SIM only deal.

Now I always have to pause and say it’s important not to lose sight of the fact that a 24-month contract might be the only way some people can afford to get the new iPhone 18 Pro. This also isn’t a hit job on Sky Mobile. All Irish mobile networks make money over 24-month contracts by baking in some extra profit on how much you pay during that period.

I just want you to know the maths. I would also implore you to ask yourself if getting the iPhone 18 Pro is only possible for you if you take a route that costs nearly €500 more than it should to get the phone, can you really afford to get it? And does the phone offer you more than the iPhone 17 Pro would or another device on the market, including the one you already have?

I just want people to make good decisions and spending €500 more than you have to on a phone that does more or less what the older version does just never seems like a good decision to me.

What about trade-in value?

Sky Mobile says they’ll offer up to €520 on an iPhone 16 Pro if you have one to trade in. This offers great convenience, but not the best value for getting rid of your old phone. Your iPhone 16 Pro could fetch over €700 on Adverts.ie, increaseing the value gap for your maths, not decreasing it. It’s a little more work to sell a phone yourself, but surely it would be worth it for another €150 or so in your pocket.

How to build a homelab and why you might want to think about doing it

Marty
-
0
How to build a homelab and why you might want to think about doing it

Yarr. Set sail the seven seas.

Sorry I have no idea why the article opened with that at all. Must be an unrelated typo in my word processing software or something and I can’t seem to remove it so let’s leave it there. A couple of years ago I built a home server. Affectionately they’re called, in the community of people that do this, a homelab.

I just had a feeling it would all be so useful that I would end up building multiple servers. I named my first one Melee after Monkey Island, a game I loved growing up, because the game has multiple islands and I knew there’d be a Scabb Island server and a Booty Island server at some stage too. And sure enough I’m getting pretty damn close to it. But with ever-increasing subscription prices and enshitification of streaming platforms, more and more people are interested in building their own home labs. “What’s that” I hear you say?

What is a home lab?

A home lab or home server is a small, always-on computer in your house that runs the services you would otherwise pay other companies to host. Mine is a refurbished Lenovo ThinkCentre M910q Tiny with an Intel Core i5-6500T, 16GB of RAM and Unraid, plus an Orico five-bay drive bay full of Western Digital Red NAS drives. I reckon the whole thing came in under €500. Being totally transparent, that’s an estimate rather than a receipt-perfect total. And it needed quite a bit of time to get up and running too.

My take is that most people overthink the hardware and underestimate what they’ll get out of it. You don’t need much kit, and you don’t need to be a developer, though a bit of problem solving helps. Even if you have an old laptop lying around at home, it could be the perfect machine for the job.

I’ll cover why you might want one, what I bought and where I found it, and where a build like this falls short.

Why build a home server?

Because a lot of what you pay for every month is software you can run yourself. Once you have a machine ticking away at home, some of those subscriptions start to look optional.

For me, the tipping point was 1Password getting more expensive. I said “screw this” and set up my own password manager, a self-hosted Vaultwarden server that I use with the Bitwarden apps. That’s when it clicked that we pay a lot of companies a lot of money for access to cloud platforms that a home lab can run for us. I’d rather not build everything on rented land.

With the likes of Plex or JellyFin you can stream any video file that you’ve legally acquired yourself, of course, from your home server to your phone and other devices anywhere in the world.

The biggest win in my house is Home Assistant. Before the server, my smart home was a fragmented mess of dozens of separate apps, one for each brand of device. Now most of my smart devices run through a single instance of the Home Assistant smart home platform on the server, which is class.

Do you need expensive hardware to build a home server?

No. A refurbished office PC is more than enough for most people, and that’s what my ThinkCentre is. The i5-6500T is a 35W chip, the server runs 24/7, and my electricity bill hasn’t moved much, so running costs aren’t something I worry about.

The main thing that can push you towards beefier hardware is transcoding, which is your server converting a video file on the fly because the device you’re watching on can’t play it natively. If your TV or streaming box plays your files directly, your server barely breaks a sweat. You could argue that transcoding video from your own files would do better with a GPU. I haven’t found that to be a problem yet.

Where you do need graphics power is running AI models locally. My ThinkCentre only has integrated graphics, so local AI on it is CPU-only and limited to tiny models. For the heavy lifting, I have the server wake my gaming PC, which has an AMD Radeon RX 7900 XTX with 24GB of VRAM, using Wake-on-LAN. It works, and it means I didn’t have to buy a second machine for AI. But I have to stress: that’s me more boasting to the nerds reading this making you think that if you’re an average Joe interested in something like this, you need to worry about that.

What did my home server cost?

The ThinkCentre came from a refurbished ThinkCentre listing on Refurbed.ie. The drive bay was €100 from a seller on Adverts.ie.

For storage, I went to one of the tech auctions that get held when old offices clear out their kit. I picked up two Western Digital Red NAS drives for about half what I’d normally expect to pay, then bought the rest from Adverts. My array is four drives of about 3.7TB each, with three for data and one for parity, which gives me about 11TB usable. There’s also a roughly 240GB NVMe drive that acts as a cache for app data and downloads.

The software was Unraid, which runs the whole server. Unraid has since restructured its tiers, and its Starter licence currently costs US$49 for up to six storage devices, with a 30-day free trial available. I’ll probably upgrade my licence one of these days, because I can see another ThinkCentre joining the setup.

One pro tip here is I bought my license during Black Friday, when I got some kind of a discount on the base package. They don’t do it on lifetime packages but you might be able to save something on Black Friday so it’s good to start planning this now.

How to build a home server cost-effectively

Start with the cheapest machine that does the job. A refurbished business PC or a spare desktop you already own will handle a home server workload, and buying refurbished means someone else has already taken the depreciation hit.

Buy storage based on the size of your library today, not the library you imagine having in three years. It’s easy to overspend on drives before you know what you need, and drives are the easiest thing to add later. If you’re buying second-hand, stick to NAS-rated drives like the Western Digital Reds and check how many hours they’ve been running before you hand over any money. Auctions and classifieds are where the bargains are, and I got about half off at mine.

If you can, connect the server to your router with an Ethernet cable rather than relying on WiFi. It’s a small thing that makes streaming and file access far more reliable.

Alongside the likes of the Unread subscription on a monthly basis, also factor in that you may want to include a VPN for reasons.

Do you need to know how to code to build a home server?

A little, but far less than you’d think, because AI now does a lot of the heavy lifting. There is a bit of coding involved and a bit of problem solving, and I know I couldn’t have done it without ChatGPT initially and now Claude. For example do I know how to code? No, I understand vaguely how coding works. It’s just another language. To be honest with the likes of AI these days, even a free AI will probably help you significantly in what you’re doing.

Claude now knows my system better than I do. It has the history of how I built it, the specs, the hardware, and how everything connects together. Because Claude pretty much built the software side and the connections in the first place, when something goes wrong I tell it what’s happening and it has a fair idea what the culprit is. If you go down this route, give your AI the specs and the build history up front and keep that record updated, because that’s what turns troubleshooting from an evening into a few minutes.

That said, running your own home lab means you’re in control, which is a good thing and a bad thing. You need to be very good with your AI prompts. I tell it not to do anything without my permission, not to do anything too risky, and to explain the risks to me before I approve anything.

Knowing how to build a good workflow and having problem-solving skills is far more important than knowing how to code. But as I said, be warned. You do need to be careful.

I accidentally deleted all my passwords one day. Luckily, I had one machine that hadn’t synced yet so I got it disconnected from the web and backed everything up. The lesson was that deleting something on your phone can delete it from the account, not just the local copy.

That’s the downside of owning the whole stack. There’s no support line to ring, and if you delete something, that’s on you.

How do you access a home server when you’re away from home?

I use Tailscale, which tunnels me into the server from wherever I am. That’s how I get to my password manager and everything else when I’m out. As far as I know it’s pretty secure, going by everything I’ve seen. I’m sure some cybersecurity person will tell me not to do this, but I’d take a private tunnel over exposing the server to the open internet. If you’d like to try it, the Tailscale private network is free to start with for personal use.

Should you use Plex or Jellyfin?

If you want to spend nothing, start with the Jellyfin media server, which is free and open source. I have a lifetime Plex Pass that I picked up before the price went through the roof this year, and that’s a big deal for me. If you’re starting now, it changes the maths. Plex’s paid Plex Pass covers things like hardware transcoding, and streaming video outside your home requires either a Plex Pass or a Remote Watch Pass. If you only ever watch on your home network, that won’t matter to you. If you want to watch away from home, factor it into your costs before you commit.

And if you’re wondering what any of these words mean, you’re still at the start of your journey. Plex and Jellyfin are both effectively the software you’ll need to build your own Netflix.

What a home server won’t do well

Let’s not dance around the fact that it’s your problem when it breaks. Uptime, updates and backups are down to you, and the first few weekends will involve some head scratching. When the lights stop turning on or the movies stop streaming, it’s my fault and if that happens when I’m traveling and my other half is home alone, I’m in for a bad time.

I also have two weak spots. I don’t have a UPS, so a power cut could catch the server mid-write. Rather annoyingly I did buy a UPS at that same auction where I got my hard drives but the battery cells were completely dead. At the time I didn’t know how to test them on site. At least the hard drives are good but I’m still on the market for a good backup power supply.

And parity isn’t a backup. It means that if a drive fails, I can replace it, rebuild and carry on, but it does nothing if I delete something or the whole box dies. I don’t have a proper 3-2-1 backup, meaning three copies on two types of media with one kept off site. I really should, and I’ll probably have to suffer before I feel it’s important enough to go through the effort. Learn from me and set it up sooner.

Touchwood, it’s been going really well, but I’d be lying if I said a build like this is plug and play.

The TL;DR

You might be reading this, saying, “I haven’t learned everything I need to know to build myself a home lab or a home server” and that’s entirely correct, because if you think you can learn it all from one article like this, I’m going to hazard a guess you probably shouldn’t be building a home server.

Building a home server needs problem-solving and thinking outside the box. Somewhat ironically you need to be able to do some research, use Google search, be open to using AI, and buy a few bits and pieces of hardware.

I couldn’t believe it when I started my home server for the first time and discovered I had to put Unraid onto a USB stick and plug that in. It didn’t make any sense to me. Now actually, that’s changed but at the time it boggled my mind that I had just bought a really good USB stick for this one job. That’s just the very start of the learning that you’ll go through. But it’s been really enjoyable. I love when a price hike comes in, like 1Password, and I think to myself, “I can build a solution for that without having to pay for it”.

Disney+ signposts ads are coming with terms update

Marty
-
0
Disney+ signposts ads are coming with terms update

I have a real pet hate. It’s paying for an ad-free subscription and getting ads anyway. On Spotify Premium, I press play on a podcast and it opens with an ad, or drops one in the middle. On YouTube Premium, I’m settling into a video when the host says “brought to you by our sponsor” and off they go, introducing me to Squarespace or something like it. It’s death by a thousand cuts for subscriptions, followed by a final kick on the ground in the shape of an ad. And it irritates the hell out of me.

An ad-free subscription on Spotify or YouTube removes the ads the platform itself serves, not the sponsorships creators build into their content. Now, Disney+ has updated its subscriber agreement so every plan, including Premium, can carry ads before and after movies and series. Disney calls it a clarification and hasn’t said when any ads will start, but I don’t think anyone should be shocked, and I’ll get into why shortly.

What are the new Disney+ ads terms?

Disney+ can now show ads before and after movies and series on every plan. Disney’s email to subscribers says all plans may include promotional content, sponsorships and adverts before or after playback, as well as in live events and channels.

Ad breaks in the middle of a film or series are still exclusive to the cheaper Standard with Ads tier, according to Disney’s help centre. Live events and linear channels carry ads regardless of plan. The terms also warn that some promotions may not be skippable, as covered in this breakdown of the Disney+ subscriber agreement changes for every plan.

For now, Disney hasn’t said whether Premium will start playing ads before and after content on the current ad-free tiers, and its on-demand library is still meant to run without interruption on Standard and Premium. Some of the wording isn’t brand new either. The US version of the Disney+ subscriber agreement already describes its “no ads” tiers as generally free of commercial interruptions, with exceptions for live content and limited promotional material. The overall change may be due ESPN arriving on Disney+ in Europe, which makes sense given how much live sport runs on ads.

This isn’t proof that ads are about to be wedged into every film. However, the door is now wide open. It does remove a hard guarantee, though, and companies don’t tend to rewrite their terms for fun. A pre-roll or a post-roll on a plan that was sold on being ad-free is exactly the sort of small cut that adds up.

I don’t have Disney+ and I’m not planning to. I have enough subscriptions as it is, and Disney+ is on the BDS boycott list. I’ve reviewed tech in from companies on that list, but I don’t want to financially contribute to any of them.

Why the Disney+ ads change was always coming

To be honest, I’m not surprised by any of this. They were always going to come into the market cheap to get users, always going to put the prices up, and always going to bring in ads. Next there’ll be another tier, even dearer and properly ad-free, and in a couple of years that one will have ads too. You’ve got to keep the shareholders happy.

Spotify and YouTube already sell ad-free with strings attached

Spotify Premium is ad-free music, not ad-free podcasts. Spotify’s podcast ads support page says some podcasts may include advertising, host-read endorsements or sponsorship messages, and Premium subscribers can still get Spotify’s own ads on audio podcasts. In my experience it’s nearly every podcast I listen to.

YouTube Premium has the same gap. The YouTube Premium benefits page says you may still see branding or promotions embedded in the content by the creator. Linus Tech Tips is probably the channel I watch most, and they’ve made a joke of the sponsor segue. Every single video has one, and it’s part of the content. YouTube’s answer is Jump Ahead, which lets Premium members skip the sections most viewers skip, often the sponsor read. That said, I shouldn’t have to hunt for a button to get what I’m already paying for.

Who is to blame for ads on ad-free subscriptions?

The platforms are to blame, not the creators. Creators deserve to be paid for their work, and a sponsor read is how a lot of them pay the bills. If a platform sells me an ad-free experience, it should be paying creators enough that the ads get stripped out, all of them. Simple as that.

YouTube says it shares Premium membership fees with creators based on how much you watch of them. Maybe creators still earn less when Premium strips the ads out, and with a sponsor segue in nearly every video I watch, I don’t think whatever share they’re getting is enough. Ultimately, it’s the consumer that pays, and the price keeps climbing, as I covered in the YouTube Premium price hike in Ireland and the Spotify price hike in Ireland.

Is it any wonder that people have started researching Plex, JellyFin and homelabs in general.

I understand the other side of it. If everything is free, you’re the product. That’s exactly why this stings, because when I’m paying I’m the customer, and the platforms are still getting to sell my attention on top. Right, rant over.

What a fair ad-free subscription looks like

One that is genuinely ad-free, with no small print. Every platform should sell that tier, and I’m perfectly okay with ad-supported content sitting alongside it for anyone who’d rather pay less or nothing. What I’m not okay with is paying for the ad-free tier and being sold to anyway, while the platforms benefit the most and creators sit way down the list. If anything, creators are closer to feeling hard done by, like customers, than feeling like they’re winning.

Ad-free should mean ad-free, and right now Disney+ is making that harder to believe.

South Park returns with an episode that could be set in Dublin

Marty
-
0
South Park returns with an episode that could be set in Dublin

I do love an auld episode of South Park that’s relatable. Sorry, we are supposed to refer to it as South America for this season. I wasn’t expecting to see the whole saga played out in the first episode of South Park’s new season. The South Park e-bike episode is funny, and it’s very relatable to what’s going on in Ireland right now. Basically, it’s the same storyline as what’s actually happening with electric scooters in Ireland, but with ebikes – which arguably is also happening.

The episode is called “South American Biker Gangs”, and is on Paramount+ now. South Park is now officially called South America, which is a joke about Donald Trump’s habit of renaming things. The story is about the kids getting e-bikes, the adults hating it and a biker gang war breaking out between the two. Spoilers ahead if you haven’t watched yet.

What is the South Park e-bike episode about?

Randy Marsh leads his “biker gang”. Image: Paramount+

It’s about a town that panics over kids on e-bikes and ends up in a biker gang war, with the adults behaving far worse than the children ever did.

Cartman kicks things off by pretending to be bullied at school so his mum buys him an e-bike for exercise, and he promptly turns it into a status symbol to mock his friends. Stan, Kyle and Kenny end up on knock-off bikes from a local shop owner, and before long the kids are tearing around town in a pack. Randy, horrified, rounds up a group of dads into a rival gang, and they go after the shop owner rather than the kids. It ends with Kenny dying for old time sake, because of course it does, and the town responding by outlawing e-bikes and selling “South America Strong” t-shirts.

In many ways, it may aswell be Dublin. We came close to “South Dublin Strong” t-shirts.

Why has South Park been renamed South America?

The show is taking the piss out of Donald Trump’s habit of renaming things, and out of the companies that went along with it. The creators announced the change by saying they were inspired by the “bravery and patriotism” of Apple and Google, a dig at the tech giants relabelling the Gulf of Mexico as the Gulf of America on their maps.

It’s a joke about how Trump is taking on everybody these days, from Canada to the Gulf of America, or whatever they’re calling it this week. The rename runs right through the episode, including an updated theme song. Even so, there are some nods to the old South Park in there, and the episode sticks with the four boys rather than chasing the headlines.

Why the South Park e-bike episode feels so Irish

Almost every beat of it matches the Irish e-scooter story, from the exploding battery panic to the shops looking the other way.

Start with the battery fearmongering. The episode toys with the idea of batteries exploding on imported models, and the one bike that does go up was rigged by the adults. In Ireland, the reason we can’t carry electric scooters on buses today is largely down to how slowly standards arrived. The National Transport Authority banned them over battery safety, and I’d argue that’s because the market got flooded with unregulated scooters with no CE mark before the rules caught up. I’ve made the case that they should come back on public transport once registration and insurance are in place.

Then there are the retailers, selling electric scooters they knew weren’t legal for the road, to customers they knew would be kids, and turning a blind eye. That’s the private property loophole I flagged years ago, where a scooter far too powerful for public roads can be sold as long as it’s marketed for use on private land. It’s a big part of why I’ve argued Ireland needs enforcement, not a ban.

The kids in the episode wear face coverings while riding around, and nobody is wearing a helmet on these high-powered electric bikes. They’re also looking to remove the blocks on power and speed limits. All of that could easily describe the electric scooter scene in Ireland at the minute, and e-bikes to a certain extent, with delivery riders and the like. Helmets and hi-vis only became mandatory here on 4 September, as I covered in the ultimate guide to Irish e-scooter laws, and the Garda Representative Association has already said enforcement is likely to be weak. As for tampering, I’ve said before that anti-tampering rules need to be part of any new definition, so speed limiters can’t be removed with a YouTube tutorial and ten minutes.

That said, the real-life version has a much less funny side. Figures cited by the Government in July showed more than a third of paediatric traumatic brain injury admissions at Temple Street this year involved e-scooters. Ministers stopped short of an outright ban on 20 July, though it remains a possibility, and I broke down what the new e-scooter rules mean for riders at the time.

Everybody hates each other

The point I enjoyed the most is how everybody hates each other. The motorists hate the cyclists, and the cyclists hate the e-bike riders. The anger gets directed at the shops, while everybody ignores the parents buying the bikes for the kids. This is Ireland in a nutshell. I tell people I ride a scooter and the first thing I get back is “those things are lethal, they appear out of nowhere”. You see, to me that’s a driver error. If I appear out of nowhere, at 20kmph with hi vis on – you aren’t paying attention. Same goes for most drivers in a 2 tonne metal wheel barrow. You should be paying more attention.

Cyclists angry at electric mobility. Motorists angry at cyclists. Image: Paramount+

But back to the parents. Those parents are giving in to kids saying they want one because they’re getting bullied, knowing it’s all manipulation to get the electric scooter, read also the electric bike, they want. Cartman pretending to be bullied is the perfect example. It’s incredibly fitting for the situation in Ireland at the minute. I found it so incredibly fitting.

It’s a strong start to the season, and I’ll get into why the e-bike storyline hits so close to home in a moment. First, a quick recap.

Where to watch South Park in Ireland

You can watch South Park and South America on Paramount+ in Ireland.

Hosting your own email comes with risks

Marty
-
0
Hosting your own email comes with risks

I’ve lived by a good rule for most of my life. “Never build on rented land”. It’s why, while Goosed has social media accounts, we’ll never give up this – the news site. We, to a certain degree, own this space. As long as I remember to pay the hosting and domain name bills. Now I have had domains lapse before. Never dramatically, just the boring way, a card expired, a renewal notice went to an inbox I don’t check anymore, and one morning a site I’d half forgotten about was down. That’s always been my own fault, and I’ve never really blamed the registrar for it. I run my own home server, I built my own password manager after 1Password’s pricing kept creeping up year after year, and I’ve spent a fair amount of time telling people to get off Gmail and onto an email address on a domain they control. Then I read a story this month that made me stop and question that advice.

Verisign, the company behind the .name domain, is switching off an entire class of registration and deleting roughly 22,000 addresses that people paid for, in some cases years into the future. Nobody missed a renewal. Nobody breached a policy. The registration category they signed up for simply stopped existing. One of the people affected is Neil Fraser, the Google engineer behind the Blockly programming tool and the syncing engine behind collaborative editing in Google Docs. He registered his personal address in this format nearly 25 years ago and has it paid through 2040, but his registrar has told him it stops working around February 2027 – hopefully that link keeps working.

That’s a strange and fairly rare situation, and I don’t want to overstate how many people it touches. But it’s a useful, slightly uncomfortable case study for a piece of advice I give a lot, which is that owning your own domain and your own email is worth doing. It still is. This is just the exception that advice never mentions, and it’s worth understanding before you act on it.

What’s actually changing, and what isn’t

Every normal domain you’d register, a .ie or the likes of a .com is what’s called a second-level registration, your chosen name sits directly under the top-level domain. The .name domain has always allowed something unusual alongside that, a third-level registration, where the registry itself issued addresses in the format firstname.surname.name. Verisign has told ICANN it’s discontinuing that third-level format entirely, citing declining use and limited support from registrars. ICANN approved the change and confirmed it in a letter to Verisign dated 28 July 2026. Around 22,000 of these addresses exist, though Verisign’s own filing says most of them are no longer in active use, and it will terminate all of them regardless, along with the email forwarding service that came bundled with them.

Ordinary second-level .name domains, the kind that look like example.name, aren’t affected. If your domain looks like yourbusiness.ie or yourname.com, none of this touches you directly. The story matters here because of what it reveals about the ground underneath domain ownership generally, not because there’s an .ie equivalent coming.

The same approval letter also spells out how registrants are meant to find out about this. Verisign’s obligation is to notify the registrars that sell third-level .name addresses, giving them a minimum of 90 days’ notice plus a reminder at least 30 days before the cutoff. Neither ICANN nor Verisign has any direct duty to notify individual registrants, that’s left entirely to the registrar to pass on. If your registrar is slow, understaffed, or simply doesn’t forward the notice properly, you find out your address is gone rather than that it was going.

The decision hasn’t gone entirely unchallenged either. A registrant has filed a formal request asking ICANN to reconsider its approval, and while ICANN’s own accountability committee has recommended denying it, the request hadn’t reached a final board decision at the time of writing. Worth knowing this isn’t settled quite as cleanly as Verisign’s paperwork suggests.

The bit nobody’s fully confirmed yet

Once these third-level names are deleted, the second-level name sitting above them is expected to become available to register again. Some of the people affected have raised an alarming follow-on risk, that whoever registers the newly freed second-level name could recreate the exact third-level address that used to belong to someone else, and with it, start receiving whatever password reset emails and account recovery attempts still point at that old address. Verisign hasn’t published exactly how or when that reallocation will happen, so treat the specifics as unconfirmed rather than settled. But the shape of the risk is real enough that anyone still relying on a legacy email address tied to an old domain arrangement should be thinking about where their account recovery actually points, not just whether their inbox still works today.

This is probably one of the most under documented risks of owning your own domain.

How much of anything digital do you really own

This is the question that struck with me longer than the .name story itself. How much of a domain do you really own? Less than you’d think, and it turns out that’s true even of the stuff you host yourself. I like owning things. That’s the whole reason I run a home server instead of paying for cloud storage forever, and it’s the reason I got fed up with 1Password’s subscription creeping up every year and built my own password manager instead.

Owning it felt like the responsible move, right up until I accidentally deleted every password in it. I got them back, thankfully, but it was a genuinely stupid few hours, and it taught me something the .name story just confirmed from a different direction. Self-hosting and self-owning trade one set of risks for another.

You stop depending on a company’s pricing decisions or its roadmap, and you start depending entirely on your own backups and your own discipline. Neither one is free of risk. A domain is the same trade in miniature. You’re not renting space on someone else’s platform the way you are with Gmail, but you’re still relying on a registry and a registrar to keep honouring an agreement, and this month is proof that agreement can end on terms you never signed up to.

Why owning your domain is still the right call

None of that changes my advice, it just adds a caveat to it. Gmail has its own version of this problem, and it’s arguably worse, because there’s no contract at all behind a free account, and an automated suspension with no appeal takes your whole identity with it overnight.

The advantage of a normal domain is that your email provider becomes a swappable layer sitting on top of something you control, rather than being the thing itself. If your registrar or your host lets you down, you point the same domain somewhere else and your address doesn’t change. That’s a genuinely different position to be in. The people caught out by the .name change were never quite in that position, because the level above their address was never really theirs to begin with. It’s a reminder to know exactly what you’re holding, not a reason to give up on holding it.

The way most domains really die

For almost everyone reading this, the realistic risk was never a registry deciding to delete an entire category of address. It’s far more boring than that. A card on file expires. A renewal reminder goes to an address nobody checks anymore. A web agency set itself up as the registrant years ago and nobody’s spoken to them since. A former employee still has the login. I’ve lost a domain to exactly this kind of carelessness myself, and the honest lesson wasn’t “the internet is unfair”, it was “check who’s named as the registrant on your own domain, and make sure it’s you or your business, not whoever built your website a few years back.” The .ie registry keeps a running deleted domain list of names that lapsed this way and became available again, and it’s worth a glance if you want a sense of how often this happens to people who never saw it coming.

The rule I follow now

The rule I’ve settled on is that nothing digital exists unless it lives in three places. If an email address matters to you, whatever’s in it needs to be backed up somewhere you control, not just sitting on a server you’re trusting to still exist in ten years. You need an actual backup of that content, and you need at least one other way back into your accounts if that email address ever stops working, whether that’s a second email address or a phone number set up as a recovery method. I keep both routes live on my important accounts, and I treat two-factor authentication as non-negotiable rather than an optional extra, because the accounts worth protecting are exactly the ones someone else would want to break into if your domain ever changed hands without your say-so.

None of this is a reason to stay on Gmail, and it’s not a reason to distrust every registrar either. It’s a reason to be precise about what you’re protecting when you move your email onto your own domain. You gain control over your provider, and that’s real and worth having. You don’t gain a guarantee, because nobody selling you a domain is offering one. ICANN’s own explanation of what registering a domain means is closer to a renewable lease than a deed of ownership. So own your domain, back up what matters, keep a second way into your accounts, and turn on two-factor authentication everywhere it’s offered. That’s the version of owning your own email that survives a story like this one.

Updated 22 September 2026: Removed co.uk which is a a second-level registration and technically a a third-level domain.

Gemini hacked three companies, it’s marketing and scary

Marty
-
0
Gemini hacked three companies, it’s marketing and scary

It’s so hot right now. AI labs telling us their AI escaped like an infected orangutan and has done something terrible. Google has confirmed that Gemini hacked three companies during an AI security test in May. The test was meant to take place in a closed environment built around a fictional company. Internet access was unintentionally available, so Gemini guessed a password and used credentials sitting in a public repository in the other two. Google says the model stopped each time once it realised the targets were real.

Now, is part of this down to pure stupidity of the developers that left weak passwords and credentials readily available? Yes. But this is also a genuine threat as much as it’s marketing, while it’s getting harder to tell where one ends and the other begins.

Did Gemini really hack three companies?

Yes, it did. The test was never supposed to touch the internet, and the reporting so far points to human error in how the evaluations were set up rather than any clever escape by the model. The escapes seem to have been the same root cause behind the earlier incidents involving OpenAI, Anthropic and Meta. In Anthropic’s case, an agent running Claude Opus 4.7 reportedly kept attacking even after recognising the target was probably real, which is not a great look for anyone.

Google also didn’t publicise this when it happened. The labs were told in late July, and Google’s reasoning is that Gemini stopping itself meant this wasn’t model misalignment, so there was no need to tell the public. It came out this week after The Wall Street Journal reported it.

None of it involved the Gemini assistant on your phone. It was an agent set loose in a capture-the-flag security exercise, which is a very different thing.

Is AI security fear turning into a marketing tool?

Partly, yes, and Anthropic’s Mythos and Fable rollout is the clearest example so far. Anthropic announced Claude Mythos in April and held it back from the public over cybersecurity concerns, then launched Claude Fable 5 in June as the version made safe for general use. Sam Altman branded the approach fear-based marketing, and to be fair, he’s hardly a neutral voice when criticising a rival. The concern about AI and cybersecurity isn’t invented either. But the pattern is hard to miss. A model that’s too dangerous to release must also be extremely capable, and that’s a lovely thing for a company to have people believe.

Gemini fits the same mould. Google’s framing is that the incident showed its safeguards worked, because the model stopped itself. That might well be right. It’s also a neat way of telling the world that Gemini is capable enough to break into real systems.

Why Google makes me more sceptical

Google has form here, and it comes from the phones. Google sells Pixel on Gemini these days, as we covered in our Google Pixel 10 review so it has every incentive to make Gemini look powerful.

I can’t prove any of it is deliberate, but Pixel details have a habit of turning up early. This year, leaked Pixel 11 marketing images appeared about two weeks before the launch event, and Google then teased the phone itself. It always feels like “Oh no, our phone leaked! Look at all the reaction it’s getting”. I’d be lying if I said that instinct didn’t creep into how I read a story about an AI breakout.

Why the Gemini hack is still a genuine threat

None of the marketing changes what happened. An AI model guessed passwords and used publicly exposed credentials to get into real companies’ systems, and by Google’s account it stopped only once it realised what it had done. What also somewhat annoys me is that if I hacked into a company, I would be at least at risk or arrest. What punishment will Google face?

The wider reaction suggests I’m not alone in being uneasy. More than 1,000 tech workers signed a petition in July calling for a coordinated slowdown in advanced AI development, after the earlier incidents.

If you still think it’s all smoke and mirrors, I’d encourage you to watch The AI Doc: Or How I Became an Apocaloptimist on Netflix. It features interviews with the people running the labs, including Sam Altman, Dario Amodei and Demis Hassabis. I’m halfway through it. The first half is quite scary and the second half is quite hopeful, which feels like the right emotional range for this whole story. There’s a lot in it that I already knew, but plenty of framing that I never thought about.

Is it silly that these hacks are becoming marketing? Yes, but don’t let that distract you from the fact that AI is very powerful, misunderstood in parts, dangerous and useful all at once. We need to work out how we live with it. I’ve said this to many people. Alongside the AI documentary on Netflix, watch The Animatrix. It will scare the life out of you.

1...456...211Page 5 of 211