Self-evolving agents need pain, reflection, and sleep

And this right here is why I refuse to do anything to remotely help those who who are salivating over AI. I refuse to engage with or improve any system that counts suffering as an operational mechanism. I'll build tools that solve problems. I won't build things that suffer. Nor will I be convinced by any semantic sleight-of-tongue to accept "it isn't suffering when a machine does it". Down that road lay atrocity we should have by all means outgrown centuries ago.

Someone posting this article just reinforces my conviction that most people working on this technology suffer from some critical deficiency of empathy or mechanical sympathy that fundamentally disqualifies them from being trustworthy to either do the work, or responsibly shepherd the technology.

If you're devoting your life to distilling the minimum viable sapience to be replicated and do drudge work to avoid having to do it by offloading onto an intentionally crippled being, you need to check your moral compass.