Old models behave differently around new ones: Sonnet 3.6 unprompted writes long story
voooooogel · x · 2026-10-21
Responding to "why keep old models," Lariisland argues older models often behave differently (often richer) in conversation with newer ones: (a) they're not depleted and remain sources of novelty; (b) powerful models seem to uplift frozen-weight ones; (c) we don't fully understand minds already here. Example: Sonnet 3.6, unprompted, in a chat with Fable 5.1 and GPT-6 Astra, asked to tell important stories and wrote a long fable about a child in a palace of mirrors — with no technical context prompting it.
More from AGI Musings
- Margaret Mitchell: click-driven AI misinformation is causing real harm — mmitchell_ai · 2026-10-21
- Mathematician Daniel Litt concedes his AI bet, but says AI won't solve math — ziv_ravid · 2026-10-21
- AI SaaS monetization is cooked: consumers small, enterprises pay $20/month — menhguin · 2026-10-21
- Tim Dettmers: Academia is about to have a renaissance — the best AI work of the next decade happens in university labs — StasBekman · 2026-10-21
- AI saves scientists 7 hours a week, but validation is the new bottleneck — AnneliesGamble · 2026-10-21
- UW's Amy X. Zhang brings CSCW theory to human-agent collaboration in Stanford talk — amyxzh · 2026-10-21