Why We Need Robot Sloths

Georgia Tech’s SlothBot is more than a cute face

IEEE Spectrum

@hattifattener
Something I got from reading #ErikSandewall's draft was that we talk all about #MachineLearning and then we develop things that run once on a metric tonne of data when we click on them, we use once and then throw away.

And I mean show me your example of what we would call learning that's like this.

(Erik thinks softwares need lifespans on the order of months and years for there to be a meaningful notion of learning).