2024-2025 was filled with huge improvements. 2025-2026 has not been, outside of open source.
The idea that we’re at the point where it’s superseded our ability to tell just makes no sense. I’ll be happy if we can get to a point where I don’t have to tell Claude not to tail every bash command or make a job that writes throughout instead of once at the end. I’ll be happy if “continue this interaction naturally, you are taking over from an independent subagent” works.
But I’m not holding my breath. It’s still really cool that any of this stuff is possible.
Claude in feb of 2025 was barely able to code. Sure, it could write you a nice function, it could even write you a complex 200-line algorithm, but give it a codebase, and it would quickly get overwhelmed.
Claude in feb of 2026? Still far from perfect, but there's definitely a huge improvement here.
You're a good contributor - it's just all too easy for unintentional sharpness to downgrade the conversation, and when it's a good conversation like this one, that's especially regrettable.
The idea that we’re at the point where it’s superseded our ability to tell just makes no sense. I’ll be happy if we can get to a point where I don’t have to tell Claude not to tail every bash command or make a job that writes throughout instead of once at the end. I’ll be happy if “continue this interaction naturally, you are taking over from an independent subagent” works.
But I’m not holding my breath. It’s still really cool that any of this stuff is possible.