Our progress toward recursive self-improvement, and its implications.

An interesting look at what Anthropic sees as the future of AI research, and how their internal use of Claude has changed over the last year

Claude-written code was somewhat worse than human-written code at Anthropic in late 2025, is roughly at parity today, and we expect it to be strictly better within the year.

but also

Work (and life) ran on a gift economy of small favors between humans. ‘Can you help me get this script running?’ […] each one created a little debt, a little mutual awareness. [Claude is] faster, it creates zero debt, but each of these is a lost bid for human collaboration.

  • NewOldGuard@lemmy.mlEnglish
    51·
    19 hours ago

    Even the premise that LLMs write as good of code as a human today is marketing BS. I work in software, I’ve seen what Claude code does when set loose on our repos already. Its output is unreliable, inefficient, doesn’t meet security standards, and always requires manual intervention to bug-fix. It produces worse code much faster

  • A_A@lemmy.worldEnglish
    35·
    20 hours ago

    That Lemmy users do not appreciate this writing tells more about them that it tells about the state of AI research and perspective impacts on society as discussed within.

    • Telorand@reddthat.comEnglish
      91·
      16 hours ago

      You’re reading a marketing piece from Anthropic. This is not an unbiased look at the AI landscape, it’s targeted propaganda.

      If you think this is something to “appreciate,” congratulations; you’re the mark.