Tokenization of your work

What running into paywalls taught me about the value of work

Every Tuesday at LinkedIn, our experience of our own product slowed to a connection speed that was at best clunky, and at its worst, made us want to throw our phones against the wall. But it was intentional. We wanted to be able to use our own product the way everyday users might in their day-to-day, including slower mobile connections. Many of us gave up our Premium subscriptions, too – another key perk – because we needed to understand better what works and doesn’t when you’re using the product as a free tier user, and what would compel us to pay for it ourselves.

At Visible, the number of phones I carried with me was always a family joke, because I constantly signed up for our own service to see if the onboarding flow held consistently, and what changed when using the service on the most affordable to the highest end phones. The point was that we needed to be intentional about going out of our way to step outside of the bubble that work built up around us with the best of our working environments, tools, and conditions. That’s not how most people live and work, and we wanted to build products for most people, not just ourselves.

I’m wondering how those with “unlimited” access to tokens and Tokenmaxxing at work are getting this part of the user experience. This was always the double edged sword of working inside of a company that comes with the benefit of access to the best and the highest tier of your own product at a subsidized cost or entirely free. You start to experience your own product in a way that most everyday people do not, and as much as you try, your own version of reality starts to take departure from the people you’re trying to serve and sell to.

My own experience integrating AI more and more into my work is that I run into the paywall quite often, at times, several times a day every few hours. I know this is by design, because there’s such high demand, and companies, of course, want us to pay more if we want extra usage and access. I’ve gone several times over my budget already, and in the end, my workaround has to be mine, not the one that’s most easily available. That means I’m being more disciplined about what tasks actually deserve my attention now vs. later, and with how much effort and time. Some tasks don’t even make the cut, as I’m doing the math and realizing it’s not worth the effort, time, and most importantly, the tokens likely required to do it. The tokenization of my work is officially here.

To be fair, I’ve always been on top of where and how I spend my time. But nothing says “unit economics of your time” like literal tokens dictating which work can continue when, and the times you’ve got to walk away because you’re cut off and in the “penalty box.” What felt like the breeziness of my initial few months with the LLMs is now entering a “just because you can’t doesn’t mean you should” phase, where discernment isn’t just about the decision making that happens inside of your integrated workflow with AI, but it’s equally about when and how you choose to do the work, if you do it at all. My work has to get better because of better prioritization, less noise, and more intentional selection and sequencing of the tasks before I even start.

The breakneck pace of innovation and progress we’re seeing today, it’s easy to forget that there’s a strategic advantage to constraints. But it shouldn’t just be on end users to exercise that discipline. The best AI habits aren’t formed by just those with unlimited access. They’re being built by the people who’ve had to think twice, and work around the constraints. I’m watching to see which companies decide that’s worth designing for.