Why Tokenmaxxing is For Fools. A Rant on Fake Productivity.
The Weekend Windup #31 - Cool reads, events, links, and more
Are your Snowflake costs going crazy? Most data teams spend months analyzing Snowflake costs. You can reclaim up to 50% of yours in just hours.
Talk to a Revefi expert today
Thanks to Revefi for partnering on this newsletter.
There’s an intense FOMO I keep picking up, both online and offline: If you aren’t maxing out AI at every second and using the latest tools, you’re going to be left behind.
If you spend five minutes on social media or YouTube, you’re hit with a barrage of posts telling you that you need to be constantly on the bleeding edge of every new agent, model, and tool. Everything nowadays has to be maxed out. Clavicular popularized “maxxing,” which I think is a symptom of late-stage civilizational decline, but that’s another story. You’ve probably heard the term “tokenmaxxing.” It’s exactly as it sounds, and it’s all the rage. It’s cute in the way kids playing with gasoline and matches is cute, but I also think it’s for fools.
As I pointed out on a panel I was on earlier this week in San Francisco, instead of tokenmaxxing, I’m actually trying to tokenminimize and brainmaxx. This isn’t some attempt at hipster contrarianism, but rather an effort to strike a better balance between using AI effectively and doing deep cognitive work.
The AI Hamster Wheel of Fake Productivity
Right now, I have agent workflows running in the background of my business and my life. It’s awesome. When I code with AI and let agents run amok, I get incredibly fast iteration. This is great for experimentation and quickly proving out ideas. But if I’m being honest, there is a massive graveyard of projects that I’ve built this way. Either way, most of the stuff goes nowhere. Chatting with my friends, this is normal. You fire up Claude Code or Codex, give it an idea or some instructions, go off to do other things, and completely forget it’s even running until you come back. It’s fun, but I think superficially we are all moving really fast without asking: What are we actually building at the end of the day?
I was chatting with Tristan Handy, the CEO of dbt Labs, about this on a podcast yesterday (to be released in a week or so). We both agreed that a year from now, we’ll probably look back on this productivity theater and wonder what the hell we were doing.
We are burning ourselves out for the sake of fake productivity. It feels like you’re running faster, but you’re just on a treadmill. Or a hamster wheel. You can have an infinite number of AI agents acting as your employees, but if you’re a terrible manager of the Michael Scott variety, you’re just going to get out exactly what you put in.
Stepping Off the AI Hamster Wheel
I read and write a LOT. AI and screens are changing this for the worse, and my attention span is increasingly frayed due to lots of context switching. If I’m on my laptop, there’s always a tab to read, an app to use, or tokens to consume. I used to feel guilty if AI wasn’t constantly running in the background, cranking away at various tasks. It reminds me of how, in the pre-Lean manufacturing days, factories would run day and night because that was considered “productive” and “efficient.” Turns out, this convention was wrong, and Lean exposed the always-on factory as an expensive anti-pattern that was unproductive and inefficient. The same applies to humans. We’re not wired to be slaves to machines running 24/7.
To combat this, I’ve simplified. I posted recently about my travel setup for this San Francisco trip. It’s remarkably lightweight. I packed my reMarkable tablet, my phone, and a hardcover book. No laptop. Nothing else. It felt like a superpower. The nature of the work I do (the work that actually matters to me) requires deep cognitive cycles, not rapid-fire, shallow iterations.
That’s why I take my time when I write. I could easily use Claude to churn out a shitty, superficial book on data modeling, but why do that? If I did that, Mixed Model Arts would’ve been out, people would hate it, and I’d be known as a slop-author. Instead, why not dive deep into a topic, do something original, and build the thing that actually needs to exist in this world? The world doesn’t need more incremental or superficial solutions (I’m pitched these daily). We need to create what’s next. That requires deep work, and deep work takes real effort. It’s slow, gnarly, and antithetical to the hustle culture that makes you feel guilty if you’re not maxxing 24/7. At the same time, I haven’t seen very many maxxers who are seeing success. They’re just running faster on the AI Hamster Wheel. You’re fast bro, but you’re still on a hamster wheel.
Taking a step back, going for walks with just my Apple Watch (and often without it), and leaving my phone behind allow me to process problems the way we used to, before our faces were permanently glued to screens. Tristan mentioned something similar. He uses screens at work, but at home, it’s about avoiding social media, going screen-free and just processing at human speed.
The irony here is thick. John Maynard Keynes wrote Economic Possibilities for our Grandchildren in 1930, stating that by now, technology would have freed us up to work 15-hour weeks. Instead, people are working 15-hour days just to keep up with the machines that were supposed to do the toiling for us.
What Actually Gets You Left Behind
Please stop listening to the FOMO bullshit artists and grifters trying to scare you into thinking you’ll be obsolete if you miss a minor model update, or if you’re not using the latest AI tools. To be fair, they’re awesome, but that’s not what’s going to get you left behind.
You know what will actually get you left behind? Sucking at the basics.
If you can’t read, if you can’t do math, if you can’t communicate with people effectively, and if you can’t sell and negotiate - those are the things that will make you obsolete. The hard parts have always been the hard parts, and going forward, real people and thinking skills will be the ultimate differentiator. AI is an amplifier for whoever is using it, for better or worse.
The people who will truly get ahead are the ones who take the time to understand the problem they are solving at a deep, fundamental level. That deep domain expertise is your moat. Anyone can ask Claude how to become a billionaire and build an app. The success rate is unsurprisingly low. Doing the things that agents can’t do is how you win. And a deep domain expert with agents has real superpowers.
Take a step back. Apply the Pareto principle: find the 20% of effort that gets you 80% of the way there, and just do that. You are going to be fine. Give yourself permission to really believe in yourself, brainmax instead of tokenmaxxing, and output high-quality decisions.
In other news, I’m off to Sweden on Monday for the Data Innovation Summit. Then AI Council the following week, followed by Confluent Current in London. Rumor has it I’ll be in Detroit as well. Also, final edits are being done on Mixed Model Arts (Book 1). Aiming for June publication, but we’ll see. Also…stay tuned for news on courses and cohorts this summer and fall, and…a conference(?).
May is going to be a crazy month.
Have a great weekend,
Joe
Here’s this week’s Freestyle Friday podcast. Available on Spotify, Apple, and wherever else you get your podcasts. Please support the show with a review. It means a lot.
Modern data modelers need to live at the intersection of business and tech. Ellie.ai allows you to collaborate effectively with business while maintaining credibility with the Tech team. Get contextual support from AI, reverse engineer anything building a repository of sources with synthetic AI generated contextual metadata while delivering insights via an MCP Server and integrating anything with full blown API support.
The full stack data modeling future is here today!
Thanks to Ellie.ai for partnering on this newsletter.
Awesome Upcoming Events
Here are a few things I’m up to. Much more to come, so stay tuned.
Data Innovation Summit 2026 - Nordics
🇸🇪 Sweden! See you at the Data Innovation Summit in Stockholm.
I’m doing a keynote and workshop on Mixed Model Arts: Data Modeling in the Age of AI.
May 7 - keynote
May 8 - workshop
Here’s 10% off: SD10OFF (good for the event. Workshop is not included)
Register here
AI Council The technical conference for humans who ship
For 10+ years, AI Council has been gathering the world’s top AI infrastructure minds to share what’s working (and where we’re headed next). You’re invited for 3 days of high-quality discourse with 1,200+ technical experts, including office hours, small groups, and zero marketing keynotes. Speakers include: The co-inventor of ChatGPT, Creator of DuckDB and Codex, Engineers behind ClickHouse, Databricks, Datadog, and LangChain.
Join us May 12-14 in San Francisco.
I’ll be there too, but not speaking. If you want to grab a beer with me and see amazing talks from amazing people, register now.
Use code JRSUB20 for 20% off, valid thru 5/4
Disclosure: They’re giving me a ticket to attend. AI Council is one of the last indie data/AI events around, and they do great work. I support non-vendor events as much as I can, and so should you.
Cool Videos and Reads
Shinji Kim, founder of SelectStar (acquired by Snowflake in December), joins the show to discuss the deal, the integration into Snowflake's Horizon catalog, and where data cataloging is actually headed.
Here are some things I read this week that you might enjoy.
Who Owns the Code Claude Wrote?
On Chaos and Turning Inward | rng.md
The 20 Software Engineering Laws - by Dr Milan Milanović
The Difference Between “Replicable” and “Not replicable” is not Itself Scientifically Replicable
Find My Other Content Here
📺 YouTube - Interviews, tutorials, product reviews, rants, and more.
🎙️ Podcasts - Listen on Spotify or wherever you get your podcasts
📝 Practical Data Modeling - This is where I’m writing my upcoming book, Mixed Model Arts, mostly in public. Free and paid content.
If you’re interested in sponsoring my newsletter and podcast, H2 2026 is opening up. Please message me for details.
The Practical Data Community
The Practical Data Community is a place for candid, vendor-free conversations about all things tech, data, and AI. We host regular events such as book clubs, lunch-and-learns, Data Therapy, and more.





What exactly is a “Clavicular” ?
Token maxxing is a huge waste of resources and money. Our work as analysts/Data scientists is supposed to save or make money, not waste it