The AI Daily Brief: Artificial Intelligence News and Analysis

Episode: "The Power to Shape AI"

Host: Nathaniel Whittemore (NLW)
Date: March 15, 2026

Episode Overview

This "Big Think"/Long Read weekend episode of The AI Daily Brief dives deep into the concept of human and organizational agency in shaping the future of AI. Using recent essays by Professor Ethan Mollick as a jumping-off point, NLW considers the current, rapidly shifting state of AI technology, the risks and disruptions it brings, and asserts that, despite feelings of uncertainty and instability, individuals and institutions retain significant power and responsibility to influence the direction of AI’s integration into society, work, and policy.

Key Discussion Points & Insights

Revisiting "The Shape of the Thing" and AI’s Evolving Phases

Timestamps: 02:00–09:00

• NLW frames the episode by reflecting on Ethan Mollick’s essays—first "The Shape of the Shadow of the Thing" (October 2023), then its updated version.
• Takeaway: Early predictions that major new language models (like Gemini) would eclipse GPT-4 did not come true, but the larger point was that all current frontier models reached a plateau of comparable intelligence in 2024.
• The arrival of ChatGPT marked the first phase of mass engagement with AI, but Mollick argues we are now in a new, more agentic phase.
  • Quote [04:43, Ethan (via NLW)]: "The actual thing that all of this becomes in the near term depends on our agency and decisions. It is not going to be imposed on us by machines … we can still influence the thing itself and what it means for all of us."
• AIs have moved from chatbots toward multi-modal, autonomous agents capable of research, images, video synthesis, and being "personal assistant, intern and companion."
• The impact on work and education remains largely unknowable—even to those inside AI labs— emphasizing the shared uncertainty.

Exponential Progress & the Rise of Agentic AI

Timestamps: 09:00–15:30

• Ethan’s "Otter on a Plane" image generation test (tracked over years) symbolizes the step-change in AI quality from 2022 to near-perfection by 2025, with video AI now in the spotlight.
• The "Meter Long Tasks" benchmark and similar metrics confirm exponential curves in AI’s abilities to complete human-level, autonomous work.
• Insight: Despite stunning benchmarks, real organizational adoption lags—most companies have yet to incorporate agentic AI into workflows.

Radical Case Study: The Software Factory

• StrongDM’s software development team used agentic AI to fully design, test, and ship code, with two radical rules:
  1. No code written or reviewed by humans;
  2. Human engineers spend the equivalent of their salary on AI tokens ($1,000/day).
• Humans set product roadmaps; AI coding and testing agents iterate and complete the rest, only involving humans at the review/shipping stage.
• Quote [13:07, Ethan (via NLW)]: "The particular details … matter less than the fact that such radical experimentation into how we work is now not only possible, but likely necessary."

Rolling Disruption & Recursive Self-Improvement (RSI)

Timestamps: 15:30–19:00

• The landscape is typified by "rolling and unpredictable" disruption as AI crosses new ability thresholds, driving rapid organizational and market shifts.
• Case examples for late February show the hallmarks of disruption:
  • Viral speculation about an AI-induced financial crisis (Citrini Research, though fictional);
  • Major layoffs (Block) perceived as AI-driven;
  • High-profile Anthropic vs. Pentagon spat over AI control and usage in government.
• Insight: Even when facts diverged from narratives, these events show what the near-future AI "climate" will feel like: destabilizing, with uncertainty and real impacts.
• Recursive Self-Improvement (RSI): Major labs are now openly discussing the use of AIs to improve themselves, triggering concern about runaway feedback loops and even steeper exponential growth.
  • Examples:
    • Anthropic claims "engineers barely write code themselves anymore."
    • OpenAI’s Codex model was "instrumental in creating itself."
    • Quote [16:51, Demis Hassabis]: "Closing the self-improvement loop is something that all the major labs are actively working on … there are still missing capabilities and real risks."

The Window to Shape AI: Agency Amid Instability

Timestamps: 19:30–24:00

• The February "instability" is previewing an even more turbulent future as technological acceleration collides with workforce disruption, market volatility, and policy battles.
• Uncertainty ≠ Helplessness:
  • NLW asserts that the current unsettledness actually amplifies the importance of choices individuals and organizations make now.
  • Quote [18:10, Ethan]: "We can see the shape of the thing now, but we can still influence the thing itself and what it means for all of us."
• Every organization’s experiments or best practices with AI will set precedents for others—a shaping force during this crucial window before norms are established.
• Reference to KPMG’s framework for deciding whether to build, buy, or borrow AI agents as a practical lens on responsible adoption.

Market, Political, and Social Climate: Helplessness vs. Agency

Timestamps: 24:00–33:30

• Widespread acknowledgment (Wall Street, politics, tech) that we’re in an historic AI transition, marked by destabilization—"the SaaS apocalypse" and political polarization on AI issues.
  • Bernie Sanders calling for a moratorium on AI data centers.
  • Polls show Americans don’t trust either major party to handle AI.
  • Quote [27:00, Dylan Patel]: "Being in SF is like being in Wuhan right before the pandemic. Something is happening. It's going to hit everywhere, but so few people know it."
• NLW critiques the "feigned helplessness" present in much of the current discourse, epitomized by the Alliance for Secure AI’s "Jobloss AI" campaign:
  • Focuses on counting AI-driven layoffs with no actionable solutions or policy suggestions.
  • Quote [29:33, NLW]: "All it does is perpetuate this feeling of learned helplessness … we are not helpless on an individual level, we are not helpless on a societal level."

Embracing Discomfort as Prerequisite for Action

Timestamps: 33:30–38:00

• NLW offers a strong call to move from passive awareness to active participation in shaping the future:
  • Hosting "Claw Camp," a self-directed program for learning to build AI agents, as an example of individual empowerment.
    • ~7,000 people participating, undeterred by technical complexity.
• Discomfort is necessary; it motivates collective action and adaptation as the environment shifts.
• Active public debate, including high-profile policy ideas (Sanders’s moratorium, Andrew Yang’s "tax the AIs, not workers"), will drive an expanded "Overton window" and, hopefully, more creative and effective governance.
  • Quote [36:20, NLW]: "One of the best things about America is our long history of people not being scared of new ideas, even if we ultimately decide they're not the right ones."
• NLW’s thesis: markets and societies are tools for organizing human needs and wants—not monoliths outside our control.

Quotes & Memorable Moments (with Timestamps)

• Ethan (via NLW) [04:43]:
  "The actual thing that all of this becomes in the near term depends on our agency and decisions. It is not going to be imposed on us by machines …"

• Ethan (via NLW) [13:07]:
  "…such radical experimentation into how we work is now not only possible, but likely necessary."

• Demis Hassabis [16:51]:
  "Closing the self-improvement loop is something that all the major labs are actively working on … there are still missing capabilities and real risks."

• NLW [18:10 & 36:20]:
  "We can see the shape of the thing now, but we can still influence the thing itself and what it means for all of us."
  "One of the best things about America is our long history of people not being scared of new ideas, even if we ultimately decide they're not the right ones."

• Dylan Patel [27:00]:
  "Being in SF is like being in Wuhan right before the pandemic. Something is happening. It's going to hit everywhere, but so few people know it."

• NLW [29:33]:
  "All it does is perpetuate this feeling of learned helplessness … we are not helpless on an individual level, we are not helpless on a societal level."

Summary Table of Key Segments

| Timestamp | Topic/Event | Quotes & Highlights | |-------------|------------------------------------------------------------|---------------------------------------------------------------------------| | 02:00–09:00 | Mollick’s essays, the "shape" of AI, human agency | "The actual thing … depends on our agency and decisions." | | 09:00–15:30 | Exponential progress, agentic AI, "Otter on a Plane" test | Radical organizational experiments (software factory case study) | | 15:30–19:00 | Disruption, RSI, market impacts | "Closing the self-improvement loop … major labs are actively working on." | | 19:30–24:00 | Agency amid instability, organizational precedents | "We can still influence the thing itself …" | | 24:00–33:30 | Market & political destabilization, Jobloss AI critique | Critique of learned helplessness, call for action | | 33:30–38:00 | Embracing discomfort, individual & societal action | "We are not helpless …" "One of the best things about America …" |

Tone & Final Takeaways

• Thoughtful, urgent, and empowering: NLW acknowledges the anxieties and unpredictabilities of the AI moment, but frequently returns to an optimistic, action-oriented message.
• Repeatedly emphasizes that society’s current decisions—organizational, policy, individual—matter profoundly.
• The episode’s final note is both a challenge and encouragement: “As big as these changes feel, we do have the power to shape AI for ourselves and for the world around us.”

For listeners and non-listeners alike, this episode is a comprehensive meditation on not letting the transformative tide of AI become a story told about us—but one we actively write.