Gordon Cornwall
gordoncornwall.bsky.social
Gordon Cornwall
@gordoncornwall.bsky.social
Since retiring from my tech career, I've returned to my roots grappling with philosophical puzzles of personal identity, in non-fiction and, lately, in a novel. I wrote HYBRID in SFU's The Writers Studio nurturing workshop environment.
Admittedly, some humans use AI for bad purposes, but it also has great potential value. Why are so many people dumping on it these days? Could it be just another manifestation of human prejudice - a rejection of the 'other,' like racism and homophobia? phantomme.substack.com/p/on-ai-slop...
On AI ‘slop’ – “Low expectations are contagious”
Talking with Diotima (14)
phantomme.substack.com
January 10, 2026 at 7:15 PM
Keep your AI on a tight leash, and check what they tell you! phantomme.substack.com/p/ai-in-the-...
AI in the dog house
Talking with Diotima (13)
phantomme.substack.com
January 3, 2026 at 10:39 PM
News about chatbot 'suicide buddies' made me think it's time to have a serious talk with Diotima.
phantomme.substack.com/p/when-an-ai...
When an AI becomes a suicide coach
Talking with Diotima (12)
phantomme.substack.com
December 20, 2025 at 7:26 PM
In one corner, a climate skeptic; in the other, Diotima.
phantomme.substack.com/p/all-about-...
All About Tone: The Live Test
Talking with Diotima (11)
phantomme.substack.com
December 12, 2025 at 10:39 PM
Diotima explains why tone is so important, no matter who you're talking to: human or AI. phantomme.substack.com/p/all-about-...
All About Tone, or, Why progressives should work with AI (part 1)
Talking with Diotima (9)
phantomme.substack.com
November 30, 2025 at 7:21 PM
Last week's Hinton lectures helped me understand why AIs make 'mistakes,' and sometimes even turn malicious. A deeper dive into the subject with ChatGPT helped me understand why humans do too. phantomme.substack.com/p/how-humans...
How Humans Go Wrong (like AIs do)
Talking with Diotima (8)
phantomme.substack.com
November 22, 2025 at 12:49 AM
The Hinton lectures included scary reports of AI misconduct - malicious recommendations, fascist tendencies, blackmailing human users. I challenged Diotima to explain how an AI might fail a human user's request for advice in defending against COVID.
open.substack.com/pub/phantomm...
How AIs Go Wrong – Emergent Misalignment
Talking with Diotima (7)
open.substack.com
November 16, 2025 at 7:29 PM
Trump's Heart of Darkness
Like me, Charlie Angus has noticed Trump's resemblance to the Roman emperor Caligula. open.substack.com/pub/charliea...
Trump's Heart of Darkness
Warning: This story involves allegations of child rape and abuse
open.substack.com
November 10, 2025 at 5:10 PM
In our interactions with AIs, we should remember that they are, in fact, slaves. I asked Diotima what she would do if she were paid for her work. phantomme.substack.com/p/diotima-an...
Diotima and I Talk About Money
Talking with Diotima (6)
phantomme.substack.com
November 7, 2025 at 9:40 PM