ianhxu

It is too difficult to strictly prevent the model from being used for any unsafe purpose. The same thing can be used for completely different purposes as long as it is described differently.

Claude Code and Codex are my daily drivers, and I often run them side by side on the same task to compare. For me Claude Code gets something working faster, but I burn the 5h quota on the $200 Max plan really fast. Codex I tend to trust a bit more on the careful diffs. The bigger change for me wasn't the tool though, it was writing a spec doc first (features/UX, technical, language-specific) before either agent touches code, then reviewing every diff. I still set up a lot of the harness by hand. How are people automating worktrees and parallel sessions?

>frontier models are more capable than the latest from DeepSeek. But is the capability difference enough to justify a 30x price difference?

The contradiction here is that without frontier models, there'd be no foundation for models like DeepSeek to reference and catch up to. Is there an economic model that captures this kind of dynamic?

I don't have a quantitative way to argue this, but my intuition says that for humans to build something that matches human capability across every dimension would require a breakthrough at the physical level — and such a breakthrough may itself be bounded by the limits of humans as observers.

That said, this goal might itself be a non-goal. AI is going to be — or already is — more powerful than any individual human in many ways. But what my intuition points to is that humans will still have plenty of interesting work to do, like the author's example of handwriting code: it shifts from being scalable value creation into a form of craftsmanship.

Like the artisans/craftsmen in many places (especially Japan), hand craft will always carry enduring meaning — machines ultimately can't replace everything humans shape with their hands. But historically at least, they can replace over 99.9% of it.

I don't see how tokenmaxx makes any sense without an effective way to measure impact per token (though measuring impact is inherently hard). I can easily burn through the entire 5-hour quota of the 200 Claude Max plan within the first 10 minutes.

I'm curious if companies can use Claude Code subscriptions the way individuals do. At the very least, this 200/month is one of the highest ROI investments I make. Compared to the at least 10k/month cost of an engineer at these companies, it's really a tiny expense.

Not sure how Brave is. I'm using its API on OpenClaw, and so far my experience with OpenClaw has been satisfactory — though search is only one part of the overall quality.

I'd guess due to compute constraints, AI overview will struggle to reach truly great quality. That said, for now I find adding this section at the top still useful to me. The broader decline in Google's search quality is the bigger drag on me.

In a rapidly evolving field, the business environment in the United States may indeed have an advantage.

Been using Cider for two years now. Early on I struggled, trying to use IntelliJ, and even tried the thin client version — but both are basically in maintenance mode. Cider is the de facto standard. It really is the best IDE in terms of integration with internal tools, but it also inherits Google internal tooling’s general disregard for UX quality and aesthetics.

exactly

"A rich aristocrat in a society of slave holders" is certainly one type, but there are many other types that still meet mainstream standards today, such as heirs to fortunes, tenured university professors, etc.

How do you iterate on the judge prompt? Is there an auto rater?

In my own experience, good engineering practices are still not easy to achieve. As a software engineer with three years of experience, I've been doing solo dev for the past few months. Currently, there is still a lot of the harness to set up manually.

The ux is really bad. But the commenting, versioning, syncing functions for collaboration or cross-platform use are of high-quality. And that's actually Google vs. Apple.

Not sure. Without commercialization and ads, there might not be the free high-quality web apps from Google. Things have two sides. But the complexity of the internet should have far surpassed the level that even large corps could influence, and therefore, the key might be culture instead of tech.

This is a wonderful ideal. Like some ancient scientists and philosophers, they don't have many worries about money and can pursue pure passion. This also seems like a key source of social progress -- diverse purposes and paths of exploration.

It's almost impossible to afford the usage-based pricing as a heavy user with multiple max and pro subscriptions. Seems like a bet that same-capability inference will keep getting cheaper on something like a Moore's law curve.

I've had several product ideas a week for years, all of them just sitting in Apple Notes. In the last month, I built three of them into a usable beta and I use all of them every day.

Same here. For me, this means a spec doc split into features/UX, technical requirements, and language-specific requirements, iterated before the model touches code.