Writing publicly about AI felt optional a few years ago. It doesn't anymore.
This blog covers four things:
- AI, honestly. What the systems can and can't do. No hype in either direction.
- Alignment and AI ethics. I take existential risk seriously. Taking it seriously does not require being a doomer.
- AI and the inclusive web. I work on web accessibility under EN 301 549. AI can make the web work for more people. It can also quietly make it worse.
- Mathematics. Focused notes on the math behind these systems, and on math for its own sake.
Why me
Not a frontier lab researcher. An engineer with a math degree who has spent years building software for real people. Technical enough not to handwave, independent enough to say what I actually think.
Most writing about AI is boosterism or doom. Neither is serious.
Format
Short posts. Longer when the subject warrants it. Honest, not comprehensive.
If something resonates, or I got something wrong: hello@danielkosbab.com.
More soon.