2025-12-17 21:00:25
I was watching Demis Hassabis talk in The Thinking Game recently and something he said stuck with me. The scientific method is one of the greatest mental models you can adopt. Being able to say "here's everything I've considered, this is what I currently believe" is just being human. What makes it powerful is the willingness to shift that position when new evidence shows up, and being excited about that prospect instead of digging in to defend what you thought before.
That second part is where most people struggle. We get attached to our positions and defending them becomes part of our identity rather than a working hypothesis we're testing. I've written before about questioning yourself as a way to build strength in your convictions, but there's a difference between examining your reasoning and clinging to conclusions that no longer hold up. The scientific method doesn't work that way - you state your position clearly, outline what you've taken into account, run your experiments, and when the results come back contradicting your hypothesis you're supposed to get excited rather than defensive.
I have a science degree that I've done absolutely nothing with professionally, but it gave me one thing that's proven valuable: a keener eye for bullshit. Not because I can cite studies or run complex analyses, but because it taught me to question everything and look at the method behind claims rather than just accepting them at face value. The problem is we've built a world that rewards certainty over curiosity, and social media amplifies the loudest voices rather than the most thoughtful ones.
Changing your mind gets framed as weakness or flip-flopping when it should be celebrated as learning. Admitting you were wrong becomes this huge thing instead of just being part of the process of understanding the world better. But being wrong is information - it tells you something about reality you didn't know before, and the people I respect most are the ones who can say "I thought X, but I was wrong, here's what changed my mind" without making it a whole thing.
I'm guilty of not doing this well too. I've published blog posts about systems or tools or ideas that didn't survive contact with reality for more than a few weeks. The temptation is to delete them or pretend they never happened, but they're part of the record and part of testing things out in the open to see what actually holds up. The best part about treating your beliefs like hypotheses is it removes the ego from the equation - you're not defending yourself, you're defending an idea that can be replaced with a better one at any time.
2025-12-17 15:44:47
I've been thinking about the Galaxy Z Fold 7 and cameras in general. Don't get me wrong, Samsung cameras are fine, I just don't think they're good enough for a £1,700 phone. I don't think they're good enough for a £1,000 phone. Samsung cameras have never hit the right quality mark for me, and some of the issue here is comparison.
Apple realised a long time ago that the cameras in their devices are so important to their users. They worked hard to make sure you get a good standard photo every time you click the button. Think about the number of people who buy a phone purely for it to be their reference point for the world. These devices we carry in our pockets are memory makers. For so many people, millions of people, it's the only thing they rely on to capture the world around them.
I take a load of photos and can be overly critical of smartphone cameras. Most modern smartphones are fine and you can't really pick fault with them. But when I think about trusting the Z Fold 7 sensors, or even the S25 Ultra sensors, to pull the phone out of my pocket and take a good enough photo, I don't have that level of trust. Eight or nine times out of ten it would get a photo that's fine, but I don't want to risk those memories being ruined by a smudgy, blurry, not-focused image.
I used an Oppo Find X9 Pro recently and it has an absolutely stunning camera that I would trust to capture the world around me. I know people rely on their smartphone cameras for all their memories, so it's not just the iPhone. I think it's Samsung's attitude towards their cameras.
Some people will swear by Samsung cameras and I just don't see it. They say comparison is the thief of all joy and I think there's something at play here. If I'd solely used Samsung phones, perhaps I would be happier with the cameras. If you're moving between Samsung phone to Samsung phone, just buying the newly upgraded one, you'll be perfectly happy with each iteration. The cameras are consistently okay and fall down in the same areas every time — low light, focus speeds and capturing detail.
Could I take good photos with them? Yes. Could I take brilliant photos? No. It loses too much detail and isn't reliable enough without messing around with too many things. Expert RAW, different apps or whatever, all to get a great photo out of it. And even then I don't think it would be great. I spend too much of my time thinking my iPhone would have captured this miles better And maybe that's the problem.
2025-12-16 18:46:40
I've been noticing some traffic on my blog from LLM based search. Small little drips of visitors from ChatGPT, Perplexity, and similar search tools. Not huge amounts, but enough to make me curious about what's actually happening.
Someone out there is asking these LLMs questions, and my lowly little blog is ending up in the answers. The problem is I have absolutely no idea what questions led them here or what context my writing appeared in. Did the AI misrepresent what I said? What were people even searching for that made my blog seem relevant?
At least with Google, you could sometimes reverse engineer what people were looking for. Search Console would tell you the queries that brought people to your site. You'd see "iPhone review" or "best camera for street photography" and think "right, that makes sense." With ChatGPT and Perplexity, there's nothing. Just a referral from their domain and complete darkness about why.
I wrote about how AI summaries stop people clicking through to sources back in July. The data from Pew Research showed only 8% of people click links when they see an AI overview, compared to 15% when they don't. So the fact that anyone's reaching my site at all from these tools feels almost accidental, and also means that I am showing up much more than these small numbers that I can see.
Part of me is glad someone's reading. I don't obsess over stats, mainly because I'd drive myself mad if I started worrying about page views. But I'd genuinely love to know what searches are bringing people here. Are they asking about Perplexity's advertising plans? My thoughts on Siri versus Gemini? Something I wrote years ago and forgot about?
I'm not precious about my writing. If I was, I wouldn't put it on the internet where anyone can read it or feed it to their large language model. I'd at least like to understand how it's being used and whether the AI is representing my actual position or just cherry-picking sentences that fit whatever point it's trying to make.
The reality is this is probably just what publishing on the web looks like now. Your work gets ground up, fed through an LLM, and spat out as part of someone else's answer. You get a referral if you're lucky. Most of the time you probably don't even get that.
2025-12-16 15:41:34
Nick Heer writing about a high-profile Apple account lockout:
Customers need a level of protection from any corporation with which they are required to have an ongoing relationship. This single high-profile incident should raise alarm bells within Apple about its presumably-automatic account security mechanisms and its support procedures.
An Apple account lockout isn't like getting locked out of Twitter or losing access to Netflix. It's losing every photo you've taken in years, every password you own, access to everything else those passwords unlock. An algorithm flags something and suddenly you can't get into your bank, your email, your work systems. Not because those services locked you out, but because Apple did and that's where all your credentials live.
The photos hurt most because they're genuinely irreplaceable. You can reset passwords and rebuild access to other services, painfully, over time. You can't get back photos of your kids that only existed in iCloud. Everyone knows you should back them up locally, but iCloud is supposed to be the backup. That's the entire pitch. Your phone uploads automatically and you never have to think about it. Turns out you do need to think about it, constantly, because trusting the system completely means one algorithm decision away from losing everything.
Nick's right that this needs actual laws. Not Apple promising to do better, not policies they can change next quarter. Legal requirements that say if you're going to lock someone's account, they get a grace period to export their data first. Make it mandatory and enforceable. Right now these companies can hold everything hostage and there's zero recourse because they own the infrastructure and make the rules.
The scale of dependency is what makes this different from older tech problems. Losing your email account twenty years ago was bad. Losing your iCloud account now means losing your photos, your passwords, your ability to access anything else. We've built these single points of failure into our lives and handed them to corporations who can cut us off for reasons they won't explain. That's not a sustainable system.
2025-12-12 17:28:08
I've been an iOS user for years. The halo of products that kept me locked into the ecosystem has fallen away one by one. The iPad stopped being my computer ages ago, then the Apple Watch got replaced by a Garmin for running. What's left is a phone running an operating system I don't find useful or particularly user-friendly any more.
The design language has become genuinely user-hostile. Everything blurs into everything else, you can't tell what's interactive and what isn't, and the lack of clear boundaries makes the whole interface feel like it's melting. It's not just bad design, it's actively inaccessible. Android phones feel faster and more fluid by comparison. I think that's linked to iOS becoming increasingly weighed down by these visual effects that serve no purpose beyond looking pretty in marketing materials. Animations are smoother on Android, transitions make sense, and the interface actually tells you what you can do with it.
When I pick up an Android phone now, it does what I expect from a modern device without making me work for it. The AI integration actually functions properly, the customisation isn't just cosmetic nonsense, and the whole experience feels like it's designed around getting things done rather than protecting Apple's ecosystem. Android is miles ahead in the ways that matter for something that's with me all day, every day.
Apps make the operating system, and Apple has become increasingly hostile towards developers. The constant rule changes, the arbitrary rejections, the way they squeeze every penny out of people trying to make a living building software for their platform. You can feel it in the apps themselves - developers are tired of fighting Apple, and that shows up in what gets built and what doesn't. There are still some gaps in my workflow on Android, mainly writing apps, but Android is good enough now to make me want to overcome these things rather than stick with iOS out of habit.
The privacy conversation is where this gets interesting. Do I give up privacy things by using Android? Of course I do. But let's be honest about what's actually happening here. Google already knows everything about me through search, email, and maps - has done for years. Pretending iOS is protecting me from that feels like theatre at this point, a performance we're all supposed to participate in where we pretend Apple's walled garden is keeping us safe from the big bad internet.
The trade-off used to weigh differently for me. Privacy mattered more when it felt like there was actually something concrete to protect, when the principle wasn't just abstract but tied to real things in my life. Now it's just something I'm supposed to care about in exchange for a service that doesn't work properly. You give up some level of privacy and in return you get something genuinely helpful. With Android you get Gemini that's built into everything Google does - it's fast, it's reliable, it actually answers questions instead of punting them to ChatGPT and waiting forever for a response that's probably wrong anyway. Siri is useless by comparison.
Apple keeps talking about on-device intelligence, which sounds fantastic until you realise the result is an assistant that can't actually assist anyone with anything. The privacy is great, I'm sure, but what's the point if the thing doesn't work? I'm not saying privacy doesn't matter, it does, and I'm not comfortable with how much data these companies hoover up. But the balance has shifted for me and the thing I'm supposed to be protecting doesn't feel worth protecting any more when the cost is a phone that barely functions as a modern device should.
Don’t get me wrong, I still have an iPhone, it's my job to use both, but the practical reality is I want a device that does what I need it to do. Right now that's Android, and I've made my peace with that compromise.
2025-12-10 18:34:38
Casey Newton wrote about Australia's new social media ban for under-16s, touching on Jonathan Haidt's "The Anxious Generation":
The book has been criticized for its lack of empirical data on the central claim that social networks are largely responsible for the mental health crisis in young people. But it was also a runaway bestseller that spoke to parents' mounting anxieties over what they can see with their own eyes: a generation of children that seems more distracted, depressed, and homebound than any in recent memory.
I'm critical of social media, particularly the algorithmic kind. Wrote a whole post about this back when Haidt's book came out. There's real harm happening. Kids spending hours on platforms built to hack their attention. Parents can see it. The research is mounting.
But there's confirmation bias at play too. We know something's wrong, so it's easy to point at patterns that fit our perspective. Same reason people fall into conspiracy theories or why humanity invented thousands of gods to explain what we couldn't understand. Our brains hate open loops. They'll work overtime to fill in gaps, even if it means forcing square pegs into round holes.
I wrote before about how we all have this gut feeling that social media isn't good for us. Charlie Warzel nailed it when he said we know this connectivity is working on us, but it's hard to pin down exactly how. That difficulty doesn't mean we're imagining things. It just means the mechanisms are complex.
The truth sits somewhere uncomfortable. Yes, modern social platforms are fundamentally different from anything that came before. MTV didn't have an algorithm designed to keep you watching at the expense of everything else in your life. Your Super Nintendo wasn't constantly A/B testing ways to hijack your dopamine system.
But we're also pattern-seeking creatures desperate for simple explanations. Mental health issues in young people are real and rising. Social media usage is real and rising. The correlation is obvious. The causation? That's where things get messy. What bugs me is when people dismiss all concerns as moral panic, when they trot out the "kids these days" line as if nothing's changed. That's lazy thinking. The platforms themselves admit they're designed to maximise engagement. Meta's own research showed Instagram made teen girls feel worse about their bodies. This isn't speculation.
At the same time, treating social media as the singular cause lets everyone else off the hook. The underfunding of mental health services. The housing crisis making young people feel hopeless about their futures. The climate crisis hanging over everything. The pandemic that robbed them of crucial developmental years. All of these things matter too. Australia's ban feels like politicians doing something visible instead of something effective. It's the legislative equivalent of doom scrolling – responding to anxiety with action that makes you feel better but doesn't actually address the root problem.
The science is still developing. Our brains are filling in gaps with assumptions. And while we figure it out, kids are growing up in this weird digital landscape we've created for them.
I still think the algorithm is the problem, not social connection itself. I still think these platforms were built to exploit us. I still think most of the harm traces back to the attention economy and advertising revenue. But I'm trying to hold space for uncertainty too. Trying not to let my own biases fill in gaps the research hasn't actually filled yet. Trying to remember that when everyone's convinced they're right about something this complicated, that's usually when we're all missing something.
