Google AI: One Model to Learn Everything. Yep, Everything

November 8, 2018

What’s an online advertising company with interesting ethical norms doing with artificial intelligence? Part of the answer appears in the ad stuffed The Stanford Daily’s story “Jeff Dean Discusses Google’s Current Efforts around AI.”

I noted this interesting point:

The future, he said, lies not in creating lots of models and algorithms for distinct tasks but rather in having one model that can learn everything.

Makes sense. A monopoly on information. Now about the ethics part from a company with a founder who dallied with a Glass wearer, a lawyer who sired a Googler at Google, and a Google VP who ended up dead on a yacht from a needle mishap involving a contract worker and a controlled substance.

Will the magic algorithm operate without bias or unusual tendencies?

Yeah, one model to learn everything. Universal. Perfect right?

Sounds good. But Google is not your mother’s AltaVista, is it?

Stephen E Arnold, November 8, 2018

References for those who do not link these ideas:

The Google Glass affair:

The productive Google lawyer reference:

The dead Googler and alleged drug explorer:


Got something to say?

  • Archives

  • Recent Posts

  • Meta