How to Help People Thrive With AI

Models Alone Don’t Drive Value
- Model improvements matter but are insufficient without training people to use them effectively.
- KPMG-style AI proficiency findings show agents are present but only 16% of workers actually use agentic tools at work. (Time 0:00:35)
AI Often Increases Work Intensity
- Early adopters using AI often work more intensely, not less, increasing messaging and software use.
- ActiveTrack and UC Berkeley research found email/chat time doubled and focused work fell by 9%, creating ‘AI brain fry.’ (Time 0:01:50)
When Intelligence Is Plentiful Volition Becomes Valuable
- AI doesn’t reduce total work for many people; it shifts what work looks like and increases intensity (more email, more multitasking, less focused time).
- The ‘infinite backlog’ feels worse with agents because they can run constantly, shifting limits from time to oversight and planning.
- Brooks’s guiding principle: when intelligence is plentiful, volition — willingness to engage mental effort — is what becomes scarce and valuable.
- People differ by need for cognition: high need enjoy hard thinking, cognitive misers avoid it, and middles engage selectively.
- These differences predict who will use AI to relax versus who will lean into AI to amplify their capabilities. (Time 0:03:27)
Volition Beats Raw Intelligence
- When intelligence is abundant, volition and willingness to undertake mental effort become the key differentiators.
- Brooks’ archetypes: productive passengers, reluctant optimizers, and mental marathoners illustrate varying relationships to cognitive effort. (Time 0:04:09)
Use AI To Challenge Not Replace Your Thinking
- Ask AI to challenge your thinking rather than produce it so you preserve cognitive agency.
- Use AI for rote tasks but reserve creative essays or memos for your own thinking when you want to build skill. (Time 0:12:13)
Stretch Capabilities By Doing New Things With AI
- The biggest opportunity is using AI to do things you couldn’t do before, not just to speed existing tasks.
- Building agents requires humility, iterative troubleshooting, and learning technical language even for non-coders. (Time 0:12:41)
Make Champions Demonstrate Not Just Promote AI
- Use internal AI champions not just for PR but to demonstrate real, practical use cases by showing what they can build.
- Champions should pair with business teams to exemplify agentic workflows, not only answer questions about tools. (Time 0:14:58)
Uber Agentic Pods Rapidly Rebuild Workflows
- Uber created Agentic Pods pairing ~30 AI-proficient engineers with domain experts to build agents in two-week sprints.
- Pods shadowed experts, prioritized, built, validated, and shipped agents, producing 2,500+ agent skills and big workflow redesigns. (Time 0:16:49)
How Uber Scales Agentic AI With Two-Week Pods
- Uber paired ~30 AI-proficient engineers with domain experts to form Agentic Pods focused on real workflows.
- Each pod had a strict two-week process: shadowing (days 1–2), prioritization (day 3), build (days 4–5), validation (days 6–9), and ship (day 10).
- Pods target complex, manual, cross-system tasks that can’t be automated from diagrams alone—engineers learn how work actually gets done.
- The approach emphasizes quick iteration and validation with actual workers to ensure generalization and real productivity gains.
- Praveen Nepalani reports this method helped move agentic AI beyond engineering into functions like finance, legal, and HR. (Time 0:17:52)
Redesign Workflows Not Just Tasks
- The biggest wins come from rethinking entire workflows rather than automating individual tasks.
- Redesigning workflows around AI eliminated handoffs, approvals, and vendor spend while accelerating decisions at Uber. (Time 0:19:02)
Reinvest AI Gains Into New Work
- Reinvest productivity gains into new, orthogonal work rather than more of the same tasks.
- Train business people through embedded collaboration so they design how to spend reclaimed time on novel initiatives. (Time 0:19:50)
AI Can Unlock Untapped Human Potential
- People often have untapped potential because work hasn’t stretched them; AI can enable higher ambitions if supported correctly.
- Whittemore argues institutions should scaffold challenges and training so more people become mental marathoners. (Time 0:21:41)