Why we may have had the biggest AI breakthrough in years & what this means for the future
In I, Robot, Asimov imagined a robot psychologist who could only understand machines by watching what they did. That's still where we are with AI today.
For six years, one neuron that lit up for both cats and cars was the AI field's favourite proof that we can never understand what AI is thinking.
Last week, a study showed we could potentially understand AI neural networks
In this episode of In The Loop, I'm going through what the research found, how it could solve major problems with AI safety and what this means for the future.
⏭️ Episode highlights
(00:00) – I, Robot and the robot…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.