Questions prompted by the Clearleft gathering in Norway to discuss AI.
I like Richard’s five reminders:
- Just because the technology feels magic, it doesn’t mean making it understandable requires magic.
- Designers are going to need to get familiar with new materials to make things make sense to people.
- We need to make sure people have an option to object when something isn’t right.
- We should not fall into the trap of assuming the way to make machine learning understandable should be purely individualistic.
- We also need to think about how we design regulators too.
Here’s a fun cosmic hypothesis on the scale of an Olaf Stapeldon story. There are even implications for data storage:
By storing its essential data in photons, life could give itself a distributed backup system. And it could go further, manipulating new photons emitted by stars to dictate how they interact with matter. Fronts of electromagnetic radiation could be reaching across the cosmos to set in motion chains of interstellar or planetary chemistry with exquisite timing, exploiting wave interference and excitation energies in atoms and molecules.
I, for one, welcome our slime mould overlords.
The slime mould is being used to explore biological-inspired design, emergence theory, unconventional computing and robot controllers, much of which borders on the world of science fiction.
Vernor Vinge’s original 1993 motherlode of the singularity.
Wonderful musings from Matt on meeting the emerging machine intelligence halfway.
An excellent rebuttal by Steven Pinker to Nicholas Carr's usual trolling.
Crows is smart. And yes, I am using the "Bookmark this..." link at the end of the article.
The Dunbar number gets bandied about a lot in conversations about social networks these days. Here's the original paper that shows the research behind the oft-misused term.
A good, if somewhat dispiriting, overview of Artificial Intelligence. (There's some nice typesetting on this page)