"The technology keeps moving forward, which makes it easier for the artists to tell their stories and paint the pictures they want."

—George Lucas, Filmmaker

Where the Advantage is Moving in the New Year

Over the past year, AI capabilities jumped fast and visibly.

Reasoning models matured. Agents became much more user-friendly. Open source coding closed the gap with frontier labs faster than anyone expected. What’s becoming clear now is that raw capability is no longer the constraint.

Integration is.

Productivity Comes With a Delay

AI productivity gains are real, but they do not show up instantly. MIT research shows many organizations experience a short-term dip before seeing stronger long-term results.

Nearly 40% of companies are experimenting with AI agents, yet fewer than a quarter have scaled them inside a single function.

Still, the U.S. remains well positioned, with labor productivity up about 10% since 2019 and most economists expecting that lead to widen.

Human Judgment Is Becoming More Valuable

As execution shifts to AI, human value concentrates.

Between 2025 and 2030, the WEF projects a net gain of 78M jobs globally, even as millions of roles are displaced. In the US, many companies plan to restructure around AI, hire for AI oversight skills, and reduce headcount where tasks can be automated.

This creates a narrower margin for error. When AI handles more execution, human oversight matters more, not less.

Trust Is Becoming the Competitive Advantage

AI-generated content now saturates the internet. Deepfakes are rising fast, and studies show people can only correctly identify them about a quarter of the time.

For U.S. businesses that operate in local markets and rely on reputation, this matters. A lot.

Generic, repetitive AI output erodes confidence quickly, so trust must come from consistency, accountability, and visible human oversight.

How to Stay Ahead This Year

The organizations seeing the cleanest results are doing four things well:

1. They apply AI to decision points that compound results, not to every task that can be automated.

2. They design workflows around continuity, with clear ownership of customer intent and pause logic when humans engage.

3. They measure momentum rather than activity, focusing on handoffs, overrides, and time to meaningful next steps.

4. They constrain agents as responsibility grows, with narrow permissions, logging, and regular testing.

ChatGPT Image Update

OpenAI just shipped an update to ChatGPT Images using its new GPT-Image-1.5 model. It’s live now for all users and available through the API.

The biggest change shows up when you edit images. If you ask to tweak one thing, the rest of the image is more likely to stay the same. Lighting, faces, logos, and composition tend to hold across multiple edits instead of drifting every time you revise.

It’s also better at following instructions. Layouts make more sense, relationships between elements hold up better, and text renders more cleanly, even when it’s smaller or more detailed.

Image generation is faster too, up to four times quicker, and you can run multiple versions at once. Plus, the API pricing is lower than before, which makes batch image work and variations easier to justify.

Results still need review, but users report the day-to-day friction is noticeably lower and realism significantly improved.

Most people think “be more specific” means “add more detail.” But specific and detailed are not the same thing.

Specific means sharper. Detailed usually just means longer. If your point is fuzzy, adding detail tends to make it worse, not better. You don’t get clarity, you get a bigger pile.

Here’s a simple loop you can use to turn your next message into a better version of itself. Same person, just dressed up for church. The loop is designed to sharpen the target without growing the junk.

If the main idea is getting closer to what you meant, you’re moving in the right direction. If it still comes back muddy, you do not need more detail. You need a clearer target. Repeat as needed.

AI is already shaping how customers find, evaluate, and trust dealerships. Most stores just do not realize how much.

In this Strategy Session, Paul sits down with Brian Kramer of Cars Commerce to walk through what the end of 2025 is revealing about search, inventory, and decision making heading into Q1 2026. They dig into how AI search engines validate dealer claims across reviews and third-party sources, why automation is becoming central to appraisals, and how EV-heavy inventory will quietly reset used car strategies.

Catch the full replay to see where AI is already influencing outcomes, whether dealers are ready or not.

• For the first time, U.S. defense and commerce officials have approved a Chinese auto AI chip for sale in the country. 🍪

• CES (Consumer Electronics Show) 2026 is in full swing this week in Las Vegas. 🤖

• Police officials in Utah had to correct an AI report that said a traffic stop turned an officer into a frog. 🐸

Thanks for reading along in 2025, and cheers to a successful new year!