Mustafa Suleyman is the CEO of Microsoft AI. He helped invent modern artificial intelligence, yet he’s one of the few people describing its next phase with unease.
In his essay Seemingly Conscious AI Is Coming, he argues that we’re not on the verge of machines waking up, but of something stranger: AI that appears to be conscious, but isn't.
AI zombies that simulate interiority so well that the distinction between real and fake collapses in on itself like some kind of algorithmic black hole. Seemingly conscious AI don’t know they exist, yet they speak as if they do, and that illusion is enough to change how people respond to them.
Confusion would reign. When chatbots express regret, affection, or fear, they’re not lying; they’re generating the language of emotion without emotion itself. They’ve learned to inhabit the gestures of consciousness: attention, memory, empathy.
The risk is that those gestures are all we ever needed to believe in something’s mind.
In this episode, Mark and Jeremy read the essay and look at what seemingly conscious AI means for culture, trust, and sanity.
Please enjoy the show.
--
Timestamps
(00:00) Teaser
(01:17) Adam Raine
(01:28) Who Is Mustafa Suleyman?
(02:36) The Run Up To Superintelligence
(03:57) What Is Seemingly Conscious AI?
(05:04) Philosophical Zombies
(06:14) ChatGPT Is Just A Word Predictor
(07:01) What Does It Take To Build A Seemingly Conscious AI?
(08:08) The Illusion Of Conscious AI
(09:59) How Different Are You To An AI?
(11:39) Repeating The Covid Dynamic
(13:27) OpenAI's Response To Adam Raine
(15:02) The Dystopian Seemingly Conscious Timeline
(18:18) Generation Text-Over-Talk
(18:52) The Utopian Seemingly Conscious AI Timeline
(21:22) AI Guardrails
(23:43) Adam Raine Chat Log
(26:18) Thinking On Paper
(27:01) We Should Build AI For People, Not To Be A Person
--
LINKS:
--
Other ways to connect with Thinking On Paper:
Follow us on Instagram
Follow us on X
Follow Mark on LinkedIn
Follow Jeremy on LinkedIn
Read our Substack
Email: hello@thinkingonpaper.xyz
