Amthropic CEO claims we are 1yr away where AI can do everything SWEs (twitter.com)

I expected Anthropic to be clever enough to position themselves differently from OpenAI. This sounds like a failing marketing message; their critical long-term customers are people, not CEOs.

That's the same timeline as commercially viable fusion power plants, or fission power plants in a shipping container, or off-the-shelf self-driving cars that work in non-ideal conditions without human intervention, or batteries that charge in minutes, or work well in cold weather, or don't weigh a literal ton to get a half-usable towing range, etc…

Ummm... no time line on when AI's excpected to outsource to less expensive SWEs.

Logically, that means that Anthropic will fire all their SWE’s one year from now.

Titlegore aside, we've been 6-12 months away from automating SWEs since GPT3.0 came out.