[jcims]

This is why I give that engineer at Google a lot of slack for his concerns that Lambda was conscious.

There’s a bit of an imperative, in my opinion, to over-index on that concern, because we could be building and torturing things for decades because people are afraid of losing their jobs.

[guerrilla]

Fair enough, it does make sense to have a higher tolerance for false positives than false negatives in this domain. I think that's a hard thing for people to accept though since so far it's been very obvious the objects and systems we interact with have not been conscious (except to pansychists of course.) It might be good to get practice in and build the habit early though, even if we don't think these current systems are conscious.

[jcims]

Exactly. In the instance it doesn't seem warranted at all, but at some point we will have to accept 'whistleblower' type accounts of the same nature with a bit less excoriation.

[antipotoad]

Also in a subsequent interview with the Google engineer [1], it comes out that the sentience claim was a bit of an attention grab, but with the reasonable purpose of airing the topic of AI ethics, and how machine learning is running way ahead of our thinking on its consequences.

[1]: https://www.youtube.com/watch?v=kgCUn4fQTsc

[jcims]

Oh wow, thank you! I hadn’t seen that one.