This seems odd. I think broadly there are two ways of structuring how you interact with AI agents:

* The first is where there's you and your computer, and you're doing pre-AI work. You hit some hotkey and pass off some task to AI.

* The second - and where I think we should probably be going, is there's you and you interact with the an agent. You aren't handing the Q4 report off to the agent, the agent is bringing the Q4 report to you.

I think the first scenario is trying to pry agentic work into legacy workflows. It will be more powerful when we simply go straight to the second, where orchestration and interaction with your agents is the interface.

This article is answering a different question to what it is asking. It's asking "What is the most effective strategy to freeze your eggs if you're absolutely certain you will need to".

The reason women freeze their eggs in their early 30s is because they still have a good chance for it to be effective and they now have a strong idea they'll need to. You don't have that second piece of information at age 19.

Or to be specific: What is the size of the cohort of women you are expecting to freeze their eggs at the age of 19, who will use those frozen eggs. How many of them will give birth to children without the help of IVF, and how many will choose never to have children.

I think this article is a good example of rationalism. Which is basically getting very mathsy about 1 specific very of the data, without viewing the data in the context of the decision that is being made.

For example, what is the percentage of women you expect to freeze their eggs at age 19, who you then expect to be unable to afford the $500 every year to keep those eggs frozen over the next decade?

Given that the UK is structurally uncompetitive in industry because of extremely high energy costs, it would seem crazy to me to build a data centre here. Even if we did build plenty of data centres here, it's simply not a driver of employment. We used to have shops. Shops employed lots of people. Then we moved to out of town shopping centres, which were cheaper and employed fewer people. Then we moved to online food delivery which was more efficient and employed even fewer people. Now we're moving to data centres which are enormous and employ practically no one. You can inflate the numbers by 10x or 100x or 1000x but you're simply not going to get numbers comparable to a car production line or anything that's actually a big driver of employment.

>we're no longer hiring: specialists. In this new era, generalists win.

I am truly astonished someone said this. My experience with AI has been the opposite. It can knock up something reasonable given the right prompting. It can generate a lot of code. It will hit deliver whatever goal you have, but it won't do it in the simplest neatest way, it won't have context about how you way to grow the codebase in the future.

The inevitable result of this is that it produces much more complexity to do a given thing than a senior engineer will. This is a big problem, because their performance degrades as their context window fills. So you will reach a point where the AI has coded you into a corner, it has created such complexity that it can't understand what is going on without filling its context window, at which point it can't do anything and you're left with a massive stinking pile of slop that's too big for you to work on and too complex for it to unpick.

The only way of sustainably using AI to accelerate your productivity is to be the specialist who can give guidance on structure and approach so that the codebase doesn't spiral.

I'm kind of surprised there has been so much talk about benchmarks. The CPU is probably the single strongest part of the product. The RAM and SSD are clearly going to be limiting factors for a lot of people, but they're not crazy, they're reasonable minimums for a budget device. If you're thinking about performance this device just really isn't for you.