Find a Place That’s Curious

There are moments when the world changes its interface. People still want the same things: good products, less busywork, fewer meetings, more time, more leverage. But the way you get there shifts.

The internet did this. Spreadsheets did this. Search did this.

LLMs are doing it now.

Which leads to an awkward truth: if you’re at a company where not everyone is genuinely excited to learn and use LLM tools, you’re not in the right place right now. Not because they’re bad people. Not because they’re incapable. Not because your job is doomed tomorrow. You’re just early. And the early stage is where culture matters more than strategy.

This Isn’t About Code

Most people still frame LLMs as a developer tool. That is like framing electricity as a better candle.

Yes, engineers get an immediate benefit. The first time you watch an agent make a consistent sweep across a codebase, you feel it in your bones. But if the only people using LLMs are engineers, your company is basically buying computers for Accounting and letting only the programmers touch them.

The real change is that LLMs move “writing” and “analysis” from being a slow, high-friction act into being a fluid loop. Write. Check. Rewrite. Summarize. Compare. Generate options. Pick one. That loop exists in every department.

If your company doesn’t get that, you will run into invisible walls all day. You’ll be faster personally, but you will still be stuck inside a slow organization.

Excitement Is the Signal, Not Usage

You can mandate “everyone must use AI” the way some companies mandated OKRs. You can buy licenses. You can run a training. You can add an “AI first” slide to the all-hands deck. That doesn’t matter.

The signal that matters is excitement. Curiosity. A kind of restless optimism.

Because LLMs are not a feature you roll out. They are a skill your organization learns. And skills are social. They spread through imitation, not memos.

When excitement is missing, you get the same pattern in every department:

• The first experiment fails because it was under-specified.
• Someone declares the tool “not ready” based on that one experience.
• Everyone goes back to the old workflow, immunized against trying again.
• The people who could have learned fastest leave, because nothing is more depressing than being the only person trying to get better.

Healthy excitement is concrete. It looks like people sharing small workflows that remove friction:

• Support drafts replies and KB articles, then edits for accuracy and tone.
• Product turns messy notes into crisp assumptions, risks, and success metrics.
• Security uses LLMs for threat-model checklists and faster incident comms drafts.
• Finance writes clearer variance narratives and scenario assumptions.
• Engineering uses agents for boring PRs, docs, and review checklists.

Why This Has to Be Everyone

If only one department adopts LLMs, the organization still moves at the speed of the slowest interface. Engineering gets faster, but Product writes vague specs and resists iteration. Marketing publishes more content, but Sales doesn’t know how to use it. Support summarizes issues, but Engineering ignores the summaries. Security bans tools, so everyone uses them quietly and unsafely.

You end up with a company where tools amplify misalignment. The people who are excited begin to feel like they’re dragging a reluctant machine. They stop sharing. They stop proposing improvements. They learn in private and quietly look for a place where the default is curiosity. That is what I mean by “wrong place.”

Yes, LLMs make mistakes. So do humans. The difference is speed. The correct response is not faith or rejection, but verification and guardrails: safe tasks (drafting, summarizing, brainstorming), high-risk tasks (numbers, legal text, production changes), privacy rules, review loops, and test plans.

This is not about loving technology. It’s about compounding. Right now, excitement about LLM tools is a proxy for learning rate. If you’re the only person pushing, you’ll feel it as friction every day.

The Point

This is not about loving technology for its own sake. It’s about learning rate. Companies with a high learning rate compound. Companies with a low learning rate decay.

Right now, excitement about LLM tools is a proxy for learning rate. It’s a proxy for whether the organization will adapt, whether it will build taste and process, whether it will attract curious people, and whether it will keep them.

If you’re somewhere that treats LLMs as a toy, a threat, or a fad, you’re not crazy for feeling friction. You’re simply in the wrong place right now.

Find a place where curiosity is normal.