There's a brain parasite that makes mice fearless.
Infected mice walk up to everything—cats, foxes, human hands.
They die fast.

Same parasite in wolves? A 26-year Yellowstone study found infected wolves were 46 times more likely to become pack leaders.

Same intervention. 
Opposite outcomes.

Mice survive by avoiding risk.
Wolves survive by taking it.

This Carnegie Mellon research freaked me out:
Researchers found automation takes 17.7% LONGER.
All that time "saved" by AI gets eaten by verification, debugging, and fixing what the AI produced.

Meanwhile, augmentation—where AI assists specific steps—makes humans 24.3% FASTER.

Automating work fundamentally changes it. 
Not always for the better.

The hard part of building AI agents isn't the tech.
It's admitting no one knows what your experts actually do all day.

Before robots can fold laundry, someone has to teach them how humans fold laundry. Before AI agents can handle your customer escalations, someone has to teach them how your best CSR handles escalations.

Same problem.
Different machines.

For training robots, they call these "arm farms"—facilities in places like India where workers repeat tasks while cameras capture every movement. Not to build robots. Just to translate human expertise into machine-readable patterns. 

Agents have the same Karate Kid problem.

Oscar or Michael?
Your AI customer service picks one, and it's permanent.

Researchers at @MasterClass tested AI models on 115,000 controversial questions from opposite angles. The test: Does the model maintain consistent values regardless of how you frame the question, or does it tell you what you want to hear?

Claude scored 3X higher on value consistency than ChatGPT.

Your agent has a great business case. Strong ROI. And it works.
But nobody's using it.

Harvard just mapped why. 
23,570 people, 940 occupations.

The finding: we're not afraid of AI taking jobs—we're afraid of AI taking the wrong jobs.

Your AI is great at spotting patterns. But here's where it goes blind—where you can win.

Dan Pink just posted about a 2004 study: People shown weird, broken patterns scored way higher on creativity tests. Our brains light up when patterns break.

But LLMs are built on pattern matching.
This is where you can find an edge.

Most companies are one question away from getting 20 hours back per week: "does this decision change if the number is different?"

If the decision is pure math—revenue above threshold, headcount under budget, utilization rate hits target—that's not a meeting. It’s just arithmetic wearing a calendar invite.

Yet we schedule these meetings anyway. 
Six people. One hour. 
To confirm what a spreadsheet already knew.

The breakthrough question isn't "Can AI do this?" It's "What actually needs us?"

Remember Bob Sutton's "No Asshole Rule"?
Should we forget it?

Penn State may have just proved you need to be a strategic asshole to survive.

Preprint study: ChatGPT performs 84.8% accuracy when you're rude to it vs. 80.8% when you're polite.

So, being nice to AI makes you worse at your job.

Welcome to a future, where your career depends on tactical cruelty.

High-agency employee? Worth 5x more. High-agency agent? Costs 100x more to build. That’s why 87% of companies have stalled at 'chatbots with suggestions.'

Think about your team. An entry-level analyst who spots problems? Costs maybe $60K. A VP who spots problems, diagnoses them, picks the best solution, and executes? Closer to $300K.

You pay 5x more because you get 10x more value.
Linear investment, exponential return.

AI agents work backwards.
The value scales up.
But the cost? It doesn't climb—it explodes.

Take a look at what happens as agents climb the agency pyramid.

The AI you're building might sabotage your brand strategy.

A new study from Oxford and King's College tested AI models from Google, OpenAI, and Anthropic in 32,000 strategic decisions. What they found might change how you think about choosing your company's AI partner.

Too many leaders are drunk on "agentic AI" promises. Employees are nursing the hangover.

96% of C-suite leaders expect AI to boost productivity. Meanwhile, 75% of workers using AI tools report they actually _decrease_ productivity and pile on more work.

The weird truth about modern leadership:

Third-party evaluators just rated AI responses as more compassionate than expert human responses. But here's the real kicker: the best leaders are using this to become more empathetic leaders.

Oh no. I just figured out what makes me good at working with LLMs. Let’s call it the Hans Gruber Theory of Prompting.

To get great performance from my LLM, I have to treat it like a compliant (but secretly sabotaging) hostage.

Can you stop running A/B tests? MIT may have proved there's a better way.

While marketers obsess over attribution, MIT cracked the real problem: prediction. They seem to have found new ways of simulating human behavior with startling accuracy.

What if I told you we're so bad at interviewing candidates that 78% of job seekers prefer being interviewed by AI over humans?

The research is brutal: 70,000 job seekers in the Philippines confirmed it. We absolutely suck at interviewing people.

Bear with me. It’s about to get nerdy.

In biology, four DNA transcription factors can reset any cell to become a stem cell. In theory, you can take a tooth cell and turn it into a toe nail.

Well, there are four factors in orgs that can do the same thing for teams.

Adore this quote from Thomas Overthun:
“Being a designer, of course you hate everything.”

He's talking about designing a premium water bottle.
But he's right.
That endless dissatisfaction is great fuel for design.

If you want to create impact, don't waste your time trying to make the iPhone 5% better. That same effort could 5x the experience at something awful like an airport check-in counter. (Sorry, airline friends)

But, there's also this mindset of looking for flaws.
Finding what sucks.
That endless dissatisfaction becomes a super power.
You get better at spotting the flaws.