scripod.com

Fable 5, Edge AI, and Personalized Models

Shownote

AI news keeps moving from bigger frontier models to smarter ways of using models: when to spend tokens on Fable 5, when Sonnet-style reliability matters more than eloquence, and how smaller edge models may become faster and more personal. Beth Lyons and A...

Highlights

In this episode, the hosts explore the practical side of AI, moving beyond the hype of new frontier models to focus on strategic model selection and efficient deployment. They discuss the trade-offs between using powerful but expensive models like Fable 5 versus more reliable and cost-effective options like Sonnet 5, emphasizing the importance of choosing the right tool for the task.
03:20
Marginal gains from newer models are minimal
05:19
Sonnet 5 caught false premises better than Opus 4.8
13:42
Give Fable a day's tasks for automation ideas
17:45
Edge computing bypasses memory constraints
25:45
Crowd-sourced information is more reliable than probabilistic AI.
30:24
Chillers consume 40% of facility power
36:27
Chip shortages drive up AI memory costs and hardware prices.
40:43
LoRA fine-tunes a frozen model without retraining all weights.
53:53
AI can invent by combining concepts without bias.
55:26
LoRA adapters are more cost-effective than context tokens

Chapters

Fable 5 and Choosing Models
00:00
Sonnet 5 Versus Opus 4.8
05:18
Claude Model Naming and Access
10:17
Android Intelligence and Edge Models
17:41
AI Search Accuracy Questions
25:43
Data Center Cooling Costs
30:18
Custom AI Chips and Memory
36:26
LoRA and Personalized Small Models
40:42
Fusion Animals and Video Prompts
49:36
Combination as Invention
55:22

Transcript

Beth Lyons: Hey, everybody. Welcome to Friday. It has no bias in terms of combination, Which is a couple days after holiday Wednesday. Andy Halliday: There's a person with arthritis typing. Chillers take 40% of the power required by the facility. Beth Ly...