Azeem Azhar's Exponential View Reflecting on 2025 (the K-shaped economy, AI's impact on work and human judgement, energy bottlenecks)
63 snips
Dec 20, 2025 Reflecting on the transformative shifts of 2025, the discussion highlights a K-shaped economy where AI excels while many organizations struggle to catch up. Cutting-edge models now complete lengthy tasks, prompting a change from doing to judging work. Rapid revenue growth in AI accompanies a hesitancy to experiment, revealing common mistakes that early adopters make. Additionally, constraints in physical infrastructure pose challenges, as top experts benefit disproportionately from AI advancements. Lastly, the episode offers valuable advice for newcomers in the evolving job market.
AI Snips
Chapters
Transcript
Episode notes
Models Deliver Hours Of High-Quality Work
- AI models in 2025 can deliver several hours of high-quality work and enable non-developers to build useful apps.
- This shifts work from doing tasks to judging outputs and specifying problems for models.
Model Specialization Raises Cognitive Load
- Different models excel on different dimensions, creating choice complexity for users.
- That cognitive load shifts to users who must pick the right model for each task.
Small Team's Yearlong AI Rewiring
- Exponential View reorganized in January 2024 to embed AI and gave team members budgets and time to experiment.
- The small team then built custom apps, ran AI sprints, and set a 'rule of five' to automate repetitive tasks.