AI budgets are exploding, and not always for the right reasons. After listening to Kipp Bodnar and Kieran Flanagan from Marketing Against the Grain, I’m convinced that most teams are chasing volume instead of value. My stance is simple: stop worshiping token spend and start measuring outcomes. That’s how growth happens, and it’s how smart operators win.
The Hype: Spend More, Learn More—But To What End?
Kipp and Kieran pulled no punches about the current craze. Developers and teams are “maxing” tokens as if more usage equals progress. In Silicon Valley, that attitude is fashionable. But fashion doesn’t pay the bills. As a marketer, builder, and crypto guy who’s seen a few cycles, I’ve learned that spending is not strategy. Outcomes are strategy.
That line, shared from a post by their CEO Yamini Rangan, nails it. It’s not anti-AI. It’s pro-results. I support aggressive learning, but not blind burn. There’s a difference.
What They Got Right About AI Spend
The show surfaced a few hard truths that leaders need to hear. One was the ballooning cost of AI usage across enterprises. Kieran noted that the average enterprise is using 13 times more tokens than last year. Kipp raised a deeper concern: does heavy usage tie to growth, or is it just noise?
“The Uber CEO talked about the fact that they had burned through their entire 2026 budget already for AI.”
I’ve seen this behavior before with ad spend, cloud credits, and even crypto gas fees in peak hype. The spending feels like momentum. **But momentum without a scoreboard is just motion.**
Yes, learning matters. Kipp argued that tokens may be cheaper now due to subsidies and that teams should learn fast while the window is open. Fair point. But unbounded tinkering often rewards the tool, not the team. That’s how vendors get rich while operators fall short.
Outcome Maxing, Not Token Maxing
The practical frame from the show is gold: tie AI work to a single, clear outcome. Kieran and Kipp put it crisply—if your team can’t say in one sentence what AI will change, you don’t have a strategy. You have usage.
That’s how you link inputs to revenue. In sales, it’s clearer: more qualified meetings and higher productivity per rep. In support, you can track ticket deflection and satisfaction. Marketing is trickier, but not impossible—speed to publish, content quality, brand lift, and channel ROI can all be measured.
Here’s my take, as someone who’s shipped products, content, and crypto projects for years: repeatable wins beat one-off stunts. Tools that get used weekly are worth the spend. One-time “experiments” that eat hours are usually vanity.
- Demand a one-sentence outcome for every AI initiative.
- Map tasks to the right model cost, not the “best” model.
- Favor repeatable workflows over disposable builds.
- Review impact weekly using one or two core KPIs.
If that list feels strict, good. Guardrails protect budgets and sharpen focus.
Where Teams Go Wrong
Kipp called out a real risk: the “token maxer” who isn’t great at the craft. AI doesn’t fix weak judgment. It magnifies it. Leaders should apply constraints—clear goals, sprints, and tight feedback loops—to filter out busywork. You don’t need a hundred tiny changes to a button. You need one well-run experiment that moves a key metric.
Kieran pushed another smart point: model choice. Most users always click the priciest model. That’s lazy. Some tasks need a heavy model; many don’t. Assign models by task, and you’ll save money without losing quality.
My Playbook For AI That Pays
Here’s how I run it in my own work across content, social, and crypto:
- Set quarterly AI projects with a single, numeric target.
- Instrument every workflow so speed and quality are visible.
- Kill any tool not used more than once a week.
- Treat “learning” as a budget line with a cap, not a blank check.
Then ask the only question that matters: did this make more money, save real time, or raise quality in a way customers can feel? If not, cut it.
The Bottom Line
Outcomes create wealth; usage creates invoices. Kipp Bodnar and Kieran Flanagan are right to push leaders away from token vanity and into measurable gains. I’m bullish on AI. I’m not bullish on waste. Set the outcome first, choose the right model for the job, and build repeatable systems that compound.
Your next move: pick one team, one workflow, and write a single sentence for the outcome you want this month. Attach a KPI. Review it every week. If it works, scale it. If it doesn’t, stop the spend. That’s how you turn AI from a habit into a win.
