Editorial illustration of the future of work: a diverse team collaborating around a translucent AI co-pilot interface overlaying documents, code, charts, and emails; subtle portrait silhouette of Sam Altman in the background as a metaphor for ideas (no logos); modern office, clean light; hints of power lines and data-center racks in the distance to symbolize energy + compute; cinematic, high-detail, balanced composition; accessible color palette; alt text: ‘AI co-pilot reshaping work, with energy and compute as the new bottlenecks.’

Sam Altman believes the cost of “usable intelligence” is collapsing, AI will reshape rather than erase most jobs, and the winners will be people and teams who learn to work with AI systems, not around them. He also argues the bottlenecks are now energy and infrastructure, not algorithms—and calls for sensible governance that protects people without choking progress.

Altman’s core thesis: Intelligence is getting 10× cheaper—fast

In early 2025, Altman wrote that the cost to use a given level of AI has been dropping roughly 10× every 12 months, which unleashes new use cases and adoption waves. That’s far steeper than classic Moore’s Law improvements. Translation: expect AI co-pilots in every workflow, from code and design to finance and customer ops.

What this means for you: Treat AI as a universal interface. Anything you do with text, numbers, images, audio, or video will be faster with a model in the loop.

Jobs: Significant impact, different mix

Altman has been frank with lawmakers: there will be “significant impact on jobs,” even if the exact shape is hard to predict. He also expects better jobs on the other side, as with prior tech revolutions. In 2025 testimony on AI competitiveness, he emphasized skilling and broad adoption as levers to capture the upside.

Recent interviews add a nuance: older cohorts may feel more disruption than Gen Z, who enter the workforce AI-native. Regardless of age, adaptability and learning velocity matter most.

What this means for you: Shift time from low-leverage tasks (summaries, boilerplate, simple analysis) to judgment, relationships, and systems thinking—the parts AI augments, not replaces.

The real bottleneck: Power and build-out

Altman has said the future of AI now hinges on a breakthrough in energy—or at least a massive scale-up of clean, reliable power. He’s repeatedly pointed to the coming compute + energy supercycle. New reporting underscores just how large that build-out may be: trillions in AI infrastructure over time.

Why you should care: The next wave of opportunity spreads to energy, grid, power electronics, construction, semiconductors, and advanced manufacturing—not just software.

Policy: Safety yes, but don’t stall progress

In his 2023 Senate testimony, Altman supported licensing or registration for the most capable models, plus rigorous pre-release testing and global coordination. By 2025, in a separate hearing on competitiveness, he stressed “sensible regulation that does not slow us down” alongside big investments in compute, talent, and supply chains. The through-line: govern responsibly, move fast on capability and access.

A values lens: Sharing the gains

In “Moore’s Law for Everything” (2021), Altman argued AI could drive abundance but policy must ensure broad distribution—even proposing mechanisms to tax capital and land and return value to citizens as an “American Equity Fund.” Whatever you think of the specifics, the signal is clear: prepare for rapid change and design inclusive systems.

The Thrive Goals 30-60-90 Day Action Plan

Days 1–30: Learn to work with AI

  • Co-Pilot Audit: List your weekly tasks. Star anything repetitive, document-heavy, or data-heavy.
  • Pick 2 workflows (e.g., drafting emails, analyzing spreadsheets). Add an AI step: prompt → draft → human edit → final.
  • Prompt patterns to master: role + goal + constraints + examples + tone.
  • Safety baseline: Establish red-line rules (no sensitive data, verify facts, keep human sign-off).

Days 31–60: Build leverage

  • Personal “AI Stack”:       
    • Text model for drafting/analysis          
    • Code assistant for scripts/automation
    • Vector notes/search for your knowledge base
    • Simple data tool (e.g., Python/Sheets) for charts and QA
  • Systems thinking: Map your team’s input → process → output. Insert AI where it trims cycle time or increases quality.
  • Create SOPs: Save prompts, checklists, and acceptance criteria so others can replicate wins.

    Days 61–90: Scale and de-risk

    • Pilot → Policy: Turn your best workflow into a team standard with guardrails and training.
    • Metrics: Track hours saved, quality scores, error rates, customer NPS. Keep what works.
    • Skill sprint: 30 minutes/day on communication, leadership, and data literacy—the human skills AI amplifies, not replaces.

    Your Future-of-Work Skills Map (mini-worksheet)

    Task Frequency Quality Standard AI Assist Idea Human Judgment Needed? Win Metric
    Example: client proposal 3×/wk 1-page brief, 3 options Draft outline + value props Pricing, tone, feasibility 50% faster, same win-rate

    Frequently asked (and misunderstood)

    “Will AI take my job?”
    It will reshape most jobs and replace parts of many roles. People who pair domain expertise with AI tools will gain an edge.

    “Is the hype just about models?”
    No—the constraint is increasingly power and infrastructure, which is why so much capital is shifting to compute, energy, and data centers.

    “Should policy slow this down?”
    Expect guardrails for top-tier models plus workforce upskilling—the debate is about how to regulate without freezing innovation.