The hosts challenge the assumption that greater token consumption necessarily creates revenue or productivity. AI activity may be visible and measurable, but it remains valuable only when it correlates with an actual business result.

Token consumption and business outcomes may be disconnected
Code changes are only a proxy for software-development value
Activity metrics can conceal weak commercial results
Teams need direct evidence that AI changed performance

“does AI usage and token usage correlate to outcomes, or are those two things disconnected from each other?”
Kipp Bodnar · 03:00

“And they measure these software developers basically on pull requests from GitHub and like how much code are they changing and improving? And those code…”
Kipp Bodnar · 05:00

#ai-roi#business-outcomes#productivity

Hot Take· 2

Hot Take05:30

Today’s Cheap AI Tokens May Be a Temporary Learning Subsidy

Aggressive experimentation can be rational when the intended outcome is faster learning. The hosts argue that venture funding and negative model-provider margins may be subsidizing current prices, so teams could face higher costs and need greater discipline later.

Learning speed can be a legitimate AI outcome
Investors may be indirectly subsidizing token usage
Current token prices may not remain this low
Future AI usage may require more deliberate allocation

“one outcome you do have to maximize is learning and how quickly you can learn.”
Kieran Flanagan · 05:30

“Tokens will never be as cheap as they are now.”
Kipp Bodnar · 07:00

#ai-economics#learning#token-costs

Hot Take07:00

AI Makes Weak Practitioners More Dangerous, Not More Skilled

Heavy AI use can amplify the judgment of the person directing it, whether that judgment is good or bad. A practitioner without strong craft may generate more work and consume more resources without improving quality or outcomes.

AI scales the user's existing judgment
High token consumption is not evidence of competence
Weak craft combined with rapid generation creates organizational risk
Teams must evaluate quality as well as activity

“But I say the most dangerous person in a company today is the person who is like token maxer and bad at their craft.”
Kipp Bodnar · 07:00

“one of the biggest epidemics in business has been measuring and reporting on activity instead of outcomes.”
Kipp Bodnar · 07:30

#craft#ai-risk#judgment

Explainer· 2

Explainer01:00

Why Companies Are Suddenly Maxing Out Their AI Tokens

Token maxing describes the push to spend aggressively on AI models, often without knowing whether that usage produces meaningful results. Enterprises are consuming far more tokens, while sophisticated models and unrestricted access make costs difficult to predict.

AI token consumption is surging across enterprises
Leaders are encouraging teams to use models more aggressively
More sophisticated models can rapidly exhaust budgets
High usage does not automatically demonstrate business value

“The average enterprise, they're burning through 13 times more tokens this year than they were last year.”
Kieran Flanagan · 02:00

“Most people I talk to are burning through the budget the first half of the year, Kieran.”
Kipp Bodnar · 03:30

#ai-spend#token-maxing#enterprise-ai

Explainer06:00

The Right AI Outcome Depends on the Business Function

AI value should be measured against the natural result of each department. Sales can track productivity and closed deals, while support can combine ticket deflection with customer-experience quality; marketing requires more nuanced measures.

Sales can measure productivity per representative and closed deals
Support can track ticket deflection and experience quality
Marketing outcomes are less binary and require careful attribution
Department-level metrics make AI usage easier to evaluate

“Support would be ticket deflection. It gets harder in marketing, actually.”
Kieran Flanagan · 06:30

“In sales, there's like a real binary outcome. You closed a deal, you did not close a teal.”
Kieran Flanagan · 06:30

#ai-metrics#sales#customer-support#marketing

Story· 1

Story10:00

AI Lets Every Idea Ship—even the Bad Ones

The episode compares AI management to editing creative talent: easier production can release valuable experimentation, but it also removes the friction that once killed weak ideas. Teams need constraints to prevent novelty, distraction, and disposable projects from consuming time and tokens.

AI dramatically lowers the cost of acting on an idea
Lower friction encourages both useful experiments and bad distractions
Management must constrain low-value creative activity
The ability to build something does not mean it should be built

“we're basically just letting everybody take every idea that's whether it's good or crazy and just do it.”
Kipp Bodnar · 10:30

“what you actually need is like a management structure and a team structure to put a little bit of constraints”
Kipp Bodnar · 10:30

#management#experimentation#focus#ai-waste

Takeaway· 4

Takeaway08:30

Give AI Projects Quarterly Targets Instead of Open-Ended Budgets

Executives can connect AI usage to value by selecting a bounded set of projects and defining what should improve. Examples include increasing content quality and speed or reducing social-media agency costs while improving engagement.

Choose a quarterly set of AI projects
Define the desired result before measuring usage
Measure content teams on both speed and quality
Tie social-media AI usage to spend reduction and engagement

“the best way to do this is to have a quarterly set of projects and outcomes you want to drive from AI usage.”
Kieran Flanagan · 08:30

“there's some sort of outcome you have to derive and say, well now I can say based upon this AI usage, we're actually seeing like…”
Kieran Flanagan · 09:00

#ai-strategy#quarterly-planning#marketing-operations

Takeaway09:00

Stop Using the Most Expensive AI Model for Every Task

Employees commonly select the strongest available model without considering task complexity or cost. The hosts expect companies to map simple tasks to cheaper models and eventually deploy fine-tuned open-source models for specialized internal work.

Simple tasks should use cheaper models
Employees rarely consider model economics
Companies can map task categories to suitable models
Fine-tuned open-source models may lower costs for specialized work

“this is a simplistic task. I should use a really cheap model.”
Kieran Flanagan · 09:00

“I think companies will start to integrate open source models and fine-tune them for their own companies and then you'll actually have tasks mapped to…”
Kieran Flanagan · 09:30

#model-routing#ai-costs#open-source-ai

Takeaway12:00

Trace Every AI Build to Its Downstream Business Value

Building an AI tool is not itself the outcome. The hosts illustrate how a second brain becomes strategically useful only when its chain of effects—faster responses, faster decisions, and faster results—is made explicit.

Treat the artifact as a means rather than the final result
Explain why the tool saves time or improves quality
Trace immediate benefits to downstream business performance
Reject projects whose value chain cannot be articulated

“you have to answer a couple of a series of questions to really connect the thing you're building to the outcome.”
Kipp Bodnar · 12:30

“I think sometimes people get stuck on like, hey, I just want to build this thing without fully connecting it to the outcome that they're…”
Kipp Bodnar · 13:00

#ai-value#second-brain#decision-making

Takeaway13:00

Use Reusability to Filter Out Disposable AI Projects

A simple filter for proposed AI work is whether the result will be used repeatedly. If an artifact is disposable and requires more than a few minutes, its expected value may not justify the time or token expenditure.

Ask whether the output will be used more than once
Prioritize repeatable actions over one-off artifacts
Keep disposable work extremely cheap
Use expected reuse to protect focus and budget

“Am I gonna use it more than once?”
Kipp Bodnar · 13:30

“If this is really like a completely disposable thing and it's gonna take me more than a couple minutes, like I probably shouldn't do it.”
Kipp Bodnar · 13:30

#prioritization#reuse#ai-projects

Frameworks in this episode

AI by Outcome Equals Strategy

Constraint-Based AI Creativity

Outcome Maxing Operating Model

Repeat-Use Build Filter

Task-to-Model Routing

Insights & moments

⚡Myth Buster· 1

More AI Usage Does Not Automatically Produce Better Results

◆Hot Take· 2