Why the AI “bubble” is actually a human capability bubble

Executive Summary

Across industries, organizations are investing heavily in artificial intelligence with the expectation that it will unlock dramatic productivity gains. Yet survey data from thousands of workers — and reporting from the Wall Street Journal — reveal a widening gap between executive optimism and employee experience. Workers report minimal time savings, increased cognitive load, and rising frustration. Executives report confidence, acceleration, and transformation.

This divergence is not evidence that AI is failing. It is evidence that organizations are misdiagnosing the nature of the technology.

AI is not a turnkey solution.
AI is a human gain‑of‑function technology — a multiplier of human capability, not a replacement for it.

When deployed into environments where foundational human capabilities are weak, uneven, or unsupported, AI does not create efficiency. It amplifies dysfunction. This whitepaper outlines the structural reasons behind the current AI‑productivity paradox and presents a capability‑first framework for realizing AI’s actual value.

1. Introduction: The Perception Gap

Recent surveys referenced in the WSJ article highlight a striking pattern:

• Only a small fraction of workers report meaningful time savings from AI.
• A majority report saving less than two hours per week.
• Many describe AI as adding work — not removing it — due to rework, error correction, and hallucinations.
• Executives, by contrast, overwhelmingly believe AI is improving efficiency and accelerating operations.
• Only 12% of CEOs report seeing both cost and revenue benefits from AI investments.

This is not a disagreement about preferences.
It is a disagreement about reality.

Workers are describing the lived experience of interacting with AI systems.
Executives are describing the projected benefits of AI systems.

The gap between these two perspectives is the first signal that the problem is not technological — it is organizational.

2. The Core Misdiagnosis: AI as a Solution Instead of a Multiplier

Most organizations treat AI as a replacement for human capability:

• Replace analysts with models
• Replace writers with generators
• Replace support staff with chatbots
• Replace decision‑making with automated reasoning

This framing is fundamentally flawed.

AI does not create capability.
AI amplifies capability.

It multiplies the strengths — and weaknesses — of the humans and systems it interacts with. When the underlying human capabilities are strong, AI accelerates performance. When they are weak, AI magnifies errors, misalignment, and operational friction.

This is the essence of the human gain‑of‑function model.

3. The Five Human Capabilities AI Depends On

AI’s effectiveness depends on five foundational human capabilities. These are not optional. They are prerequisites.

3.1 Critical Thinking

AI outputs require interrogation, validation, and contextual judgment.
Without strong critical thinking, AI becomes a generator of plausible‑sounding errors.

3.2 Perspective Skills

Executives and workers often inhabit different operational realities.
AI exposes this misalignment instantly, creating friction when perspective‑taking is weak.

3.3 Systems Thinking

AI interacts with incentives, workflows, governance, and culture.
Deploying AI without systems literacy leads to brittle, failure‑prone implementations.

3.4 Long‑Term Orientation

Organizations that chase short‑term automation gains often reverse course when quality drops.
AI rewards patience and capability building, not impulsive cost‑cutting.

3.5 Creativity

AI’s highest value emerges when humans use it to explore, design, and innovate.
Without creativity, AI becomes a faster autocomplete — not a strategic asset.

These five capabilities form the substrate upon which AI can deliver meaningful value.

4. How Current AI Deployments Expose Capability Gaps

The WSJ article and related surveys reveal predictable failure patterns:

4.1 The “AI Tax”

Workers spend significant time correcting AI‑generated errors.
This is not inefficiency — it is the cost of deploying AI into environments lacking strong critical thinking and validation workflows.

4.2 Automation Theater

Companies announce automation gains, then quietly rehire humans when quality drops.
This reflects a lack of systems thinking and long‑term orientation.

4.3 Misaligned Expectations

Executives see AI as a strategic accelerator.
Workers see it as a source of rework.
This is a failure of perspective skills.

4.4 Underutilization of Advanced Capabilities

Most workers use AI for drafting and search replacement.
Few use it for analysis, modeling, or design.
This reflects a creativity gap and a lack of capability scaffolding.

These patterns are not random. They are structural.

5. The Real Bubble: The Belief That AI Replaces Human Capability

Satya Nadella recently warned that if only tech companies benefit from AI, it is a bubble. The deeper bubble, however, is the belief that AI can substitute for human capability.

The belief that AI can replace:

• critical thinking
• systems literacy
• perspective‑taking
• long‑term reasoning
• creativity

This belief is driving billions of dollars in misallocated investment and unrealistic expectations.

AI cannot replace these capabilities.
AI depends on them.

6. A Capability‑First Framework for Realizing AI’s Value

To unlock AI’s actual potential, organizations must shift from a technology‑first approach to a capability‑first approach.

6.1 Assess Human Capability Baselines

Before deploying AI, evaluate the five gain‑of‑function domains across teams.

6.2 Build Capability Scaffolding

Training must focus on judgment, validation, workflow integration, and creative application — not just tool usage.

6.3 Redesign Workflows for Human‑AI Symbiosis

AI should augment human decision‑making, not bypass it.

6.4 Align Executive and Worker Perspectives

Executives must understand the lived reality of AI‑mediated work.

6.5 Invest in Long‑Term Capability Development

AI transformation is not a quarter‑to‑quarter initiative.
It is a multi‑year capability‑building process.

Organizations that adopt this framework will see AI become a force multiplier.
Those that do not will continue to experience the AI‑productivity paradox.

7. Conclusion: Upgrade the Humans First

The data is clear: AI is not failing.
Organizations are failing to prepare the humans who must use it.

AI is a gain‑of‑function technology for people.
It amplifies what humans bring to it.

If organizations bring clarity, judgment, and capability, AI becomes transformative.
If they bring confusion, misalignment, and wishful thinking, AI becomes a drag.

The path forward is not more automation.
It is more human capability.

Upgrade the humans first — and AI will finally deliver what the headlines promise.

 

 

Seeing Beyond Our Instruments: Why Open Science Matters

 

Reading recent discussions in astrobiology—like this Big Think article exploring why scientists still struggle to define life—strikes a chord with something I’ve believed for a long time: open science isn’t optional. It’s essential.

Ecosystems, whether on Earth or imagined on distant worlds, are shaped by dynamics that often slip past our instruments and models. Deep microbial networks, faint ecological signals, slow‑moving processes, and emergent behaviors all operate in realms we can barely detect. Our tools illuminate only a fraction of what’s actually happening. The rest remains hidden in the noise.

Astrobiologist Carol Cleland argues that our inability to define life stems from this same limitation: we’re constrained by what we can currently perceive and measure. And I agree. The absence of a comprehensive theory of life shouldn’t be a reason to pause inquiry. It should be a catalyst to expand it.

If anything, the gaps in our understanding should push us toward more inclusive scientific frameworks—whether we’re probing alien worlds or studying the ecosystems beneath our feet. When we acknowledge the limits of our observations, we also acknowledge the possibility that life may exist in forms we haven’t yet imagined.

This is one of the great challenges of modern science: so much of the real “action” in ecosystems happens beyond the reach of our instruments. Signals are too subtle. Interactions unfold over centuries. Entire processes evade quantification. And without a holistic theory of life, we risk overlooking entire categories of living systems—whether they’re hidden in Earth’s deep biosphere or thriving somewhere far beyond our planet.

All of this underscores why transparency, interdisciplinarity, and openness to the unknown aren’t just philosophical preferences. They’re practical necessities. Without them, science becomes a reflection of our current biases rather than a tool for discovering what lies beyond them.

Open science widens our field of view. It invites new perspectives, new methods, and new interpretations. It helps us see what our instruments can’t. And ultimately, it increases our chances of recognizing life—whatever and wherever it may be.

Call to Action: Keep the Search Open

If we want to discover life in all its possible forms—on Earth or beyond—we need to build a scientific culture that welcomes uncertainty rather than fears it. Support open data. Collaborate across disciplines. Question assumptions. Share methods, not just results. And above all, stay curious about what lies outside the range of our current tools.

The universe is far richer than our instruments. Let’s make sure our science is rich enough to meet it.

 #OpenScience #Astrobiology #Ecology #ScienceCommunication  #InterdisciplinaryScience #ScientificInquiry 

 

Beyond AI: The Skills That Keep Science Human

 

I recently read an article on the Inner Development Guide and its Thinking domain, which highlights five skills—Critical Thinking, Perspective Skills, Systems Thinking, Long-Term Orientation, and Creativity—as the true upgrade for researchers.

From my perspective, these skills are not just enhancements; they are the essence of what humans bring beyond AI. I’ve written about the pantheistic fallacy—the mistake of projecting human qualities onto AI and assuming it can replicate the full spectrum of human cognition. AI is powerful, but it is not wisdom. It processes data, finds correlations, and optimizes for efficiency. What it cannot do is ask what’s missing or imagine what needs to be discovered.

• Critical Thinking → AI validates hypotheses, but only humans interrogate the assumptions behind the questions.
• Perspective Skills → AI merges data, but humans bridge science with lived experience, culture, and meaning.
• Systems Thinking → AI maps connections, but humans perceive paradox, cooperation, and emergent properties.
• Long-Term Orientation → AI optimizes for immediate goals; humans imagine regenerative futures and purpose-driven trajectories.
• Creativity → AI recombines patterns; humans risk failure, leap into the unknown, and dare to ask what if.

This is also where open science becomes essential. These human capacities flourish most when knowledge is shared, not siloed—when researchers build on each other’s insights, challenge assumptions openly, and collaborate across disciplines and cultures. Open science creates the conditions where critical thinking, systems thinking, and creativity can scale across networks rather than remain isolated within institutions.

The future of science will not be defined by AI alone, but by scientists who can leverage AI as a human gain‑of‑function—amplifying our ability to see what’s missing, question what’s assumed, and explore where the undiscovered awaits.

Read the original article: Science Needs More Than Data: Have You Led Your Own Thinking Yet?

#OpenScience #AIgovernance #PhilosophyOfScience #RelationalThinking #AugmentationNotReplacement #CarloRovelli #SciencePolicy #EpistemicHumility #SystemsThinking #ProvenanceMatters

 

Composing Reality: Rovelli’s Relational Physics and the Case for Open Science

 

Carlo Rovelli’s relational view of reality isn’t just a radical physics insight—it’s a governance imperative for AI and science. I couldn’t agree more.

Rovelli’s latest interview in Quanta Magazine is a masterclass in epistemic humility. His relational interpretation of quantum mechanics—where reality is not a fixed inventory of things but a web of interactions—offers a profound challenge to legacy models of objectivity and control. This isn’t just theoretical physics; it’s a philosophical stance with real-world consequences.

I see Rovelli’s perspectivalism as a call to action. We must abandon the illusion of detached, universal truths in favor of systems that honor situated knowledge, human judgment, and pluralistic accountability. That’s the essence of open science—not just transparency, but a governance architecture that foregrounds relationality, provenance, and participatory design.

In AI, this means rejecting monolithic models of intelligence and embracing frameworks that are composable, interpretable, and grounded in human values. Philosophy isn’t a luxury here—it’s the scaffolding for responsible deployment. Rovelli reminds us that physics itself was shaped by thinkers like Nagarjuna, Kant, and Mach. Why should AI be any different?

We’re not just building tools. We’re shaping the conceptual schemes through which reality is understood and acted upon. That demands philosophical rigor, operational clarity, and a commitment to governance that reflects the complexity of the systems we’re intervening in.

Let’s stop pretending that science and technology are neutral. They’re perspectival. And that’s exactly why they need philosophy.

Read the full interview: Carlo Rovelli’s Radical Perspective on Reality

#OpenScience #AIgovernance #PhilosophyOfScience #RelationalThinking #AugmentationNotReplacement #CarloRovelli #SciencePolicy #EpistemicHumility #SystemsThinking #ProvenanceMatters

 

 

Amplifying Human Experience: Enabling Gain-of-Function via AI

Artificial Intelligence is too often mythologically known as an all-knowing oracle, human replacement, or rival species.” This framing reflects a pantheistic fallacy—the belief that more data equals perfect knowledge. It assumes that scale alone can substitute for understanding, and that exposure to vast corpora confers expertise. But AI’s true power lies not in autonomy, but in augmentation. It is not a rival mind—it is a cognitive scaƯold, a tool for amplifying human judgment, not replacing it. When we treat AI as an oracle, we obscure its blind spots. When we treat it as a replacement, we abdicate responsibility. But when we treat it as an augmentative partner— bounded by provenance, guided by human validation, and aware of its epistemic limits— we unlock its real potential: to extend human insight, not overwrite it. Reframing AI as a gain-of-function technology positions it as scaffolding that extends human cognition, perception, judgment, creativity, and inclusion. This shift demands a new design philosophy—one that embraces uncertainty, centers human judgment, and prepares for the unseen variables that shape our world. AI becomes less about what it is, and more about what it helps us become. When built responsibly, it amplifies human flourishing and generates positive ripple eƯects across society and ecosystems. This paper explores five core domains where AI delivers gain-of-function capabilities. It illustrates how augmentation works in practice through detailed examples, and concludes with risk mitigations, composability opportunities, and prescriptive guardrails for responsible deployment. "AI should be judged not by what it replaces but by the new human capabilities it enables.

 

The Problems of Philosophy in the Age of AI

 When Bertrand Russell wrote The Problems of Philosophy in 1912, he grappled with the gap between appearance and reality, asking how we can know anything with certainty when our senses may deceive us. Russell’s skepticism presumed that underlying truth existed and could be approached through rigorous inquiry. Over a century later, his questions have not merely persisted—they have proliferated into new domains of epistemic risk. Artificial intelligence does not simply introduce fresh uncertainties; it actively manufactures realities, fragments shared understanding, and operates at speeds that preclude human deliberation. In this landscape, the peril is not ignorance but epistemic surrender: the quiet abdication of judgment to systems that neither know nor care what is true. This paper revisits Russell’s inquiry in light of AI’s epistemic power, arguing for a renewed ethics of validation, provenance, and human oversight.

The Problems of Philosophy in the Age of AI