[Banditoz]

This is what I don't really understand. It's a bit difficult to take "wait x months" at face value because I've been hearing it for so long. Wait x months for what? Why hasn't it happened yet?

Things seem to be getting better from December 2022 (chatgpt launch), sure, but is there a ceiling we don't see?

[wlindley]

"Self-driving cars" and Fusion power also come to mind. With the advent of photography, it was widely believed that drawing and painting would vanish as art forms. Radio would obsolete newspapers, becoming obsolete themselves with television, and so on. Don't believe the hype.

[elzbardico]

Kind of funny that some of those wait-6-month people are basically the same ones behind "no human driven cars being sold after 2025" and "computer vision is all you need"

[dham]

My car has driven me back and forth with no issues for 6 months now. But yes it's been a long time coming.

[XenophileJKO]

And yet.. my car was surrounded by 5 self-driving cars with no people in them on the way to work on Thursday.

[XenophileJKO]

Um.. Claude Code has been out less than a YEAR.. and the lift in capability in the last year has been dramatic.

It does seem probable based on progress that in 1-2 more model generations there will be little need to hand code in almost any domain. Personally I already don't hand code AT ALL, but there are certainly domains/languages that are under performing right now.

Right now with the changes this week (Opus 4.6 and "teams mode") it already is another step function up in capability.

Teams mode is probably only good for greenfield or "green module" development but I'm watching a team of 5 AI's collaborating and building out an application module by module. This is net new capability for the tool THIS WEEK (Yes I am aware of earlier examples).

I don't understand how people can look at this and then be dismissive of future progress, but human psychology is a rich and non-logical landscape.

[stoneforger]

Because then you won't be important, the model will be important. And then everyone will have to use their model, that's their dream. Why isn't that your nightmare too? Why will you be special if it can just code whatever it needs to code? Then anthropic can just employ all the programmers that will ever be needed, to just review new skills and modules of code. It was predicted early on there would be a need for about six big computers worldwide. Well now we'll just need six AI shepherds. And then literally everyone else will forget how anything works because it will be a solved problem. People already treat computers like magic, it will literally become a dark art. And I guess it's fine, what can we do, right? Go with the flow I guess. "If I don't , someone else will. Maybe I can be one of those six real people at Anthropic".