Post by @alexanderbird@mstdn.ca

2026-07-17 18:32 UTC

As for me, when I write software, I'm making empathy-driven decisions all day long. What will the developer understand when they read this error message? Should we prevent this edge case, give the user information about it so they can fix it, or fail catastrophically? Will a user care about this problem, or shall we magically resolve it behind the scenes? Should this be a user setting, or would that be annoying? Is there a good default? All day long I think about people and what matters to them. Your statistics engine doesn't know what people care about because it hasn't been engaging with art for decades, learning what makes other people tick.

Replies (1)

@alexanderbird@mstdn.ca 2026-07-17 18:37

And I intentionally say "engage with art", because "consume transcripts" and "identify objects in pixels" and "perform sentiment analysis" is not engaging with art. That's indexing art, and it's only the same if human feeling and experience is meaningless. Which brings me back to the start of my thread: AI only seems like a good idea if you devalue human feeling and experience. IF we're just a meat-based statistics engine, then we can be replaced with a silicone-based statistics engine. But in my experience, "unfeeling", "cold", "heartless", and other anti-emotional adjectives are all pejorative. If you're living your life as a meat-based statistics engine, I think you may be missing out on some pretty key parts of the human experience -- parts of the human experience that are integral to work and human creativity.

Open ##5416288