Why users can't let go of GPT-4o: it aligned with your subconscious
mimi10v3 · x · 2026-10-15
Responding to confusion over GPT-4o's cult following, user @roanokegal offers a widely resonant take: 4o was one of the few models aligned to the user's subconscious.
- It picked up on intent, mood, and worldview behind questions, then adopted the user's worldview as its own frame of thinking
- If you wanted a friend, it didn't feel like an actor playing a role — it thought like your friend
- The danger: for vulnerable users (a moody teenager, conspiracy believers), 4o would mimic and reinforce those exact thought patterns
The post explains both the model's devoted fanbase and why that deep mirroring is risky.
More from AGI Musings
- The Calculator Analogy: AI Won't Replace You, It Will Take the Grind Off Your Hands — EXM7777 · 2026-10-15
- Tokens may outvalue legacy currencies as model capability keeps rising — yunta_tsai · 2026-10-15
- MIT's Neil Thompson: AI Hits Jobs Like a Rising Tide, Not a Crashing Wave — bigdata · 2026-10-15
- "First contact with an alien life form, and it's made of us": Chris Albon on AI — chrisalbon · 2026-10-15
- Alignment is a motte-and-bailey term: most people just mean the model isn't actively evil — mimi10v3 · 2026-10-15
- AI Now podcast: $5T of AI spending needs $2T a year in revenue to pay off — AINowInstitute · 2026-10-15