David Cost

Two Futures for Artificial Intelligence

There are two competing visions of artificial intelligence. In one, humans work for AI. In the other, AI works for humans. The future we choose will determine whether AI replaces human capability or expands it.

Tech History

Overclaiming Is a Silicon Valley Tradition

Overclaiming is a Silicon Valley tradition. The internet promised perfect ad attribution. The market believed it for decades. The AI industry is falling into the same trap. But everything about this era is moving faster. The correction may arrive before adoption is locked in.

AI Productivity

The One Thing That Will Immediately Make AI More Useful for You

Every time you open a new AI session you are starting from zero. The AI has no idea who you are or how you work. Context artifacts fix that. The more foundational knowledge you capture and feed to AI systems, the faster and better the output becomes. Start with one document. Your brand voice.

Tech History

The AI Industry Is Rediscovering Something Old

Proprietary data as competitive advantage is not a new idea. In 1969 David Ness co-founded Interactive Market Systems, building programmatic advertising on a mainframe. The right first question hasn't changed in fifty years.

Agentic Commerce

Agentic Commerce and the Reality of Debit Cards

Most agentic commerce conversations assume a customer with a credit card. That is not most of America. For a significant portion of this country, the only form of electronic payment is a debit card. One agent error and the rent check bounces.

AI Strategy

Public Polling on AI Is Consistently Negative. Let Me Explain Why.

Mobile brought everyone to the table. AI is pulling up the chairs. The people most affected by AI job displacement are the same people who lack the sophistication to participate in the AI economy.

AI Opportunity

AI Plus Beginner's Mind

Steve Jobs called the PC a bicycle for the mind. Think of AI as a Formula 1 car. The people who will thrive are not the most technically sophisticated. They are the ones who know what they want to create.

AI Governance

Your Product Team May Have Already Given Away the IP

When your team types a proprietary formulation into a public AI tool, that information leaves your control. A federal court ruling and a Proskauer Rose framework make clear the trade secret and patent exposure is real.

AI Governance

Public vs. Private AI: What Most Organizations Don't Know

That $20 a month subscription your employees are using is public AI. Not private. A federal court ruled in February 2026 that documents created in a public AI platform destroyed attorney-client privilege. Most organizations don't know this yet.

AI Strategy

The Spreadsheet Moment: Where AI Is Right Now

Spreadsheets worked beautifully in the hands of the person who built them. Then they tried to share them. AI is following the same pattern. Most organizations have the prototypes. Very few have the architect.

Tech History

David Ness and the Origin of Decision Support Systems

David Ness was my professor, my mentor, and my friend. He built the intellectual foundation for what AI is trying to do now. Wharton dissolved the Decision Sciences department in 1993. They didn't understand what they had.

Agentic Commerce

AI Is Real. Agentic Shopping May Not Be.

The wave is real. Not everything riding it is. The organizations that win identify which AI capabilities will fundamentally change how they compete. Identifying the paradigm shift is the skill. Anyone can chase the fad.

AI Strategy

The Prescription: Where to Start With AI

The first question isn't which model to use. It's where in the organization do decisions actually get made. Map the decision architecture first. Deploy the technology second. Most organizations do this in the wrong order.

Tech History

We Planted a Flag Before the Category Had a Name

In 1997 I co-founded PriceSCAN before comparison shopping had a name. The rules are different early. Access that exists now won't exist later. Some advantages can't be acquired at any price once the window closes.

AI Results

Two Buckets. Clean Record. Both Directions.

Two years of production AI experiments produced a clean finding. Every experiment designed to replace a person failed. Every experiment designed to enhance a person won. The record is clean in both directions.

Tech History

Refusal Is a Choice. It's Just Not a Good One.

Spreadsheets didn't eliminate finance teams but professionals who refused to learn them were replaced by ones who did. AI is following the same pattern. The gap is already forming. Refusal is a choice. It's just not a good one.

Enterprise AI

The Individual vs. Organizational AI Gap

Individuals are getting faster with AI. Organizations aren't. Workflows, decision systems, and operating models are largely unchanged. That gap is where most of the opportunity is.

Agentic AI

Agentic AI Raises the Bar. Most Organizations Aren't Ready.

Agentic AI promises systems that take action, not just generate output. That requires clear decision boundaries, defined ownership, and governance most organizations haven't built. Without that, the technology remains a demo. Not an operating system.

Enterprise AI

The Organization Was Never Designed to Absorb It.

Most AI initiatives fail not because the technology is weak but because the organization was never designed to absorb it. The same approval chains. The same incentives. The same operating structures. The technology works. The organization simply cannot use it effectively.

Enterprise AI

AI Is the Builder. Architecture Determines the Advantage.

Most organizations ask which AI model to use, which vendor to select, which pilot to run. Those are not the first question. The first question is architectural. AI behaves like a builder, not a strategist. Durable advantage comes from the architecture around it.

AI and Work

The Real AI Employment Story

CEOs say 55% expect AI to increase hiring in 2026. Only 9% plan workforce reductions. The organizations that treat AI as augmentation will compound their advantage. The ones that treat it as cost reduction will miss the opportunity.

AI and Work

The Ceiling Is Determined by the Person, Not the Tool.

Same model. Same underlying system. Our creative director produced materially stronger results than the platform provider's own demonstration. Access to the model is not the advantage. Human direction is. The ceiling is determined less by the tool and more by the person guiding it.

AI Opportunity

AI Is a Bicycle for the Mind. Not a Replacement for the Rider.

Steve Jobs called the PC a bicycle for the mind. AI is the same principle at a different order of magnitude. The organizations that treat AI as augmentation will compound their advantage. The ones that treat it as headcount reduction will miss the point.

Tech History

AI Is the New Social Is the New Digital.

For fifteen years companies were rewarded for having a strong digital story. Then a strong social story. Now an AI story. The announcement often moves the stock. Operationalizing it is much harder. Most AI strategies stall for the same predictable reasons every time.