Thanksgiving 2025: What I'm Thankful For in the AI World

It's Thanksgiving week. I've been thinking about what I'm actually grateful for—not in a saccharine way, but in a genuine sense. And doing this work in AI, there are some things worth naming.

I'm Grateful the Hype Is Finally Settling

2023 and early 2024 were exhausting. Every conversation started with "AI will change everything" and ended with "or maybe AI changes nothing." The uncertainty was paralyzing for a lot of people. By Thanksgiving 2025, we've landed somewhere more grounded. AI is powerful. It's also not magic. It requires real thinking, good process, and sustained effort. That's more honest. It's also more actionable.

I'm Grateful for the People Doing It Right

I've worked with dozens of firms implementing AI. The ones I'm most grateful for are the ones that resist easy answers. They ask hard questions. They measure carefully. They're willing to fail and learn. They don't treat me like a consultant to confirm their intuitions; they treat me like someone to push back against their lazy thinking. Those relationships are rare and valuable.

I'm Grateful for the People Saying No

A partner at a law firm told me recently: "We looked at AI for document review. We realized the real bottleneck isn't analysis speed, it's decisions about what matters legally. AI can't help with that. So we're not doing it." That's hard-won wisdom. Most people would have implemented it anyway because the vendor promised they would. The ones honest enough to recognize fit/no-fit boundaries—those people are building sustainable organizations.

I'm Grateful for Open Questions

We still don't fully know how to govern AI systems. We're still figuring out the liability implications. We're still learning what productivity gains actually stick once the novelty wears off. These are open questions, and I'm grateful for the people treating them as open instead of settled. That intellectual humility is a feature.

I'm Grateful for the Tools Getting Better and Worse in Useful Ways

The capability floor has gotten higher. You can get solid AI assistance from almost any major model now. But the differentiation is more subtle—which tools actually work with your systems? Which ones stay reliable at scale? Which ones give you explainability? This is healthier. We're past the "my AI is better than your AI because I used a fancier model" era.

I'm Grateful for the People Adapting

Lots of people worried AI would displace them. Some of those same people have learned to use AI as a tool that makes them better at their work. A contract lawyer who now uses AI for initial contract review and spends their time on judgment-heavy negotiations. An accountant who uses AI for standard audit procedures and focuses on advisory and client relationships. These people haven't been displaced; they've been redeployed. That's a human story that matters.

I'm Grateful for the Skeptics

The people who ask "Why are we doing this?" are doing everyone a favor. They cut through the noise. They force you to articulate actual business value. Some firms have killed AI projects that would have wasted millions of dollars and years of effort because someone was willing to be skeptical. That's not obstruction; it's stewardship.

I'm Grateful for the Industry Progress

Professional services move slowly. Regulatory industries move slower. That's frustrating sometimes, but it's also protective. In three years, we've gone from AI being a wild-west free-for-all to having actual guidance on how to use it in regulated work. That's good. Boring governance and risk management saves a lot of people from bad outcomes.

I'm Grateful for What Didn't Happen

Nobody's been seriously harmed by mass AI hallucination in critical professional work. We didn't see the jobs apocalypse that doomers predicted. We didn't see the utopia that techno-optimists promised. We saw something more interesting: selective adoption, mixed results, genuine learning. The extreme scenarios didn't materialize. That's a gift.

I'm Grateful for the Work Ahead

2026 is going to be more complex than 2025. Agentic systems will require more sophisticated governance. Regulatory clarity will come, and some of it will constrain what you can do. The easy wins will be mostly taken, and the hard wins will require real effort. I'm grateful for that challenge, because it means this work still matters. We're not at "AI solved everything." We're at "AI is one tool among many, and using it well requires sustained thinking."

The Real Gratitude

If I'm honest, I'm grateful for the opportunity to work on this at a moment when it's genuinely hard and genuinely important. The easy phase of AI adoption is over. The real phase—integrating it sustainably into professional services work—has begun. That's the work worth doing, and I'm grateful to be doing it with people who take it seriously.

Happy Thanksgiving.