I've been in this game for twenty years. A decade at Google, a handful of startups, hundreds of engineering hires. And in all that time, there's one signal that's never let me down when sizing up whether a company actually gives a damn about its engineers: the tools they give them.
The MacBook Pro Test
Cast your mind back to 2012. You're interviewing at a company. You show up on day one, and they hand you a mid-range ThinkPad running Windows 7. Nothing wrong with ThinkPads — fine machines. But if every other serious tech company was giving engineers a fully-loaded, top-of-the-line MacBook Pro, what did the ThinkPad tell you?
It told you the company didn't get it. Not because macOS is inherently superior (calm down, Linux people), but because the MacBook Pro had become a cultural signal. It said: we respect your craft, we want you to have the best, and we're not going to nickel-and-dime you on the tools of your trade. The laptop cost what, two, three thousand dollars? A rounding error on an engineer's salary. But the signal it sent was worth ten times that.
The 2026 Equivalent
Here's my argument: in 2026, the equivalent of that MacBook Pro is tokens. Specifically, giving your engineers unconstrained access to AI coding tools: Claude, Copilot, Cursor, whatever they prefer. No penny-pinching on subscriptions or imposing arbitrary usage caps.
I keep hearing stories that make me want to tear my hair out. Engineers at well-funded companies who have to share a single Claude subscription across a team of five. Developers who hit their token limit at 2pm and just... stop using AI for the rest of the day. Organisations that block AI tools entirely because someone in legal got nervous. In 2026!
Let's do some quick maths. A Claude Max subscription costs a few hundred dollars a month. An experienced software engineer costs — conservatively — fifteen to twenty-five thousand dollars a month in total compensation. If that subscription makes your engineer even five percent more productive (and it's dramatically more than that), you've paid for it many times over. This isn't complicated arithmetic. And yet.
It's Not Really About the Money
Just like the MacBook Pro, the token budget was never really about the dollar amount. It's about what it signals.
When a company gives engineers unlimited AI tokens, it says:
- We trust you to use good judgement about how you work.
- We care about the quality of your output, not the theatre of your process.
- We're paying attention to how software is actually built right now.
- We're not going to make you fight procurement for the tools of your trade.
When a company restricts them, it says something different entirely. It says: we don't trust you. Or worse, we haven't bothered to think about it. Neither is a good look if you're trying to hire (or retain) the best people.
Find your AI-friendly employer
Browse CompaniesThe Companies That Get It
This is actually why I built this directory. I wanted to make it dead simple to find companies that are embracing AI-assisted development, not just paying lip service to it in job ads, but actually giving their engineers the tools and the freedom to work this way.
The companies in there aren't all big names. Some are early-stage startups where the founders are building product themselves with Claude Code. Some are scale-ups that made the deliberate decision to go AI-first across their entire development lifecycle. What they have in common is that none of them are constraining their engineers with silly token limits or blanket AI bans.
And look, I get that there are legitimate questions around security, IP, and compliance. I'm not saying throw caution to the wind. But those are problems to be solved, not excuses to do nothing. The companies that have figured this out are already moving faster than the ones still drafting their "Acceptable Use of AI" policy. (If your AI policy is longer than your engineering onboarding doc, you might have your priorities backwards.)
A Simple Hiring Litmus Test
If you're an engineer evaluating your next role — and you should always be evaluating — here's a question to add to your interview list: "What AI tools do your engineers use day-to-day, and who pays for them?"
The answer will tell you more about the company's engineering culture than any number of slides about their "tech stack" or "innovation mindset." If the answer is a shrug, or "we're still figuring that out," or "engineers can expense up to $20 a month"? That's your ThinkPad moment. Keep looking.
If the answer is something like "everyone gets Claude Max, we encourage vibe coding, and we judge people on what they ship"? Now we're talking.
Companies that care about their software engineers doing good work will give them the tools to do their jobs well. That was true in 2012 with MacBook Pros, and it's true in 2026 with tokens. The signal hasn't changed. Only the currency has.