I recently ran into a fun little problem with the Helix Stadium XL and Streamlabs Desktop. And by fun, I mean the kind of problem where you spend a stupid amount of time wondering why the thing you can hear perfectly fine in your headphones sounds completely wrong in the recording.
The Stadium was connected to my Mac over USB-C. Streamlabs could see it. The audio meters were moving. I hit record, played a bit, listened back, and instead of hearing my processed Helix tone, I got the dry DI.
GPT-5.5 dropped yesterday under the rumoured codename Spud, and I’ve been hammering on it inside Codex for most of today. I want to say something that probably won’t be popular. It feels a lot like GPT-5.4.
That’s not a bad thing exactly, but it’s also not the leap the X timeline had me bracing for. Every leaker for the past month was whispering about big model smell. The general vibe was that OpenAI was about to drag the throne back from Anthropic in one swing. Sam’s strawberry potato tweet didn’t help calm anyone down. By the time the announcement landed, half the AI internet had convinced itself we were getting a generational jump.
Remember Web3? The decentralised future where users owned their data and corporations did not control everything? The revolution that was going to disrupt Big Tech and return power to the people?
It was speculation. That is all it ever was. The tech industry took gambling on tokens and wrapped it in revolutionary language to make it sound like innovation. The revolution was not coming. The tokens were coming. That was the whole thing.
We used to build things. Software that did something. Products that solved problems. Tools that people used.
Now we build apps to talk about building things. Project management tools for managing projects that produce nothing. Communication platforms for communicating about communication. Productivity apps that consume more time than they save.
The industry has become meta. We are so busy building tools for building that we forgot to build anything.
Look at a modern developer’s setup. They have tools for task management. Tools for note-taking. Tools for documentation. Tools for communication. Tools for code review. Tools for deployment. Tools for monitoring. Tools for managing the other tools. Each tool promises productivity. Together they consume productivity. The time spent configuring, maintaining, and switching between tools is time not spent on actual work. The tools became the work.
You can earn more money. You can rebuild your health. You can make new friends to replace the ones you lost. Most resources that feel scarce can be regenerated if you invest the effort.
Attention cannot. You have a finite amount of attention in your lifetime. When you spend it, it is gone. There is no getting it back. There is no earning more. You get what you get and then you die.
There was a moment, somewhere around 2012, when the internet was as good as it was ever going to get. We did not know it at the time. We thought it would keep getting better. Instead it got worse, and we have been coasting on momentum ever since.
This is not just nostalgia. The structural incentives that made the early internet good have reversed. What we have now is optimised for different things, and those things are mostly bad for the people using it.
In 2008, the financial crisis destroyed millions of jobs. The unemployment rate doubled. People lost homes, savings, careers. It was the worst economic crisis in generations and it scarred everyone who lived through it.
I think we are headed for something worse. Not a financial crisis. An automation crisis. The kind of job displacement that makes 2008 look like a practice run.
This is not going to happen all at once. It is going to happen gradually, then suddenly, the way these things always happen. And when it hits, we will not be ready because we spent the warning period arguing about whether it was real.
I have a ducted air conditioning system made by iZone. It works fine. There’s an app. The app does what apps do. You tap buttons, things happen. It’s perfectly adequate.
But I’m a developer, and perfectly adequate is never enough. I wanted to control my AC from the terminal. I wanted to type a command and have my house cool down. I wanted to ask an AI assistant to set up my bedtime routine and have it actually do it. Not some hypothetical future integration. Right now, on my local network, with no cloud dependency.
Someone alive today will be the last professional writer. Not the last person who writes. People will always write. But the last person who makes a living primarily by arranging words on behalf of others. That person exists right now. They might be in their twenties or thirties. They might be starting their career as I type this.
And they will be the last.
This sounds dramatic. It is dramatic. It is also, I think, probably true. The economics of professional writing have collapsed, and AI is accelerating that collapse to its conclusion.
For decades, we optimised for the wrong things. Memorisation. Credentials. Which university name you could drop. How many facts you could recall in an interview. The education system was built around filling heads with information that would be useful later.
Then AI happened and suddenly the information is just there. You do not need to remember it. You need to know how to get it.
The new skill is prompting. Not in the narrow sense of writing clever queries to ChatGPT. In the broader sense of knowing what to ask, how to ask it, and what to do with the answer. The people who thrive in the AI era will be the ones who can extract value from these tools. Credentials are becoming less relevant by the day.