Companies spend millions on transformation roadmaps. Almost no one thinks through what happens to the org during the transformation. In 2010, China learned this the hard way—and it created the history's worst traffic jam.

In 1999, four Italians stuffed two old cars with foam and tried to float across the Atlantic Ocean. It took 119 days. Two crew members quit. And the man whose dream started it all died before they reached the other side.

GPT-5 and Claude achieve 25% success working together on a coding task. A single agent doing both jobs? Roughly 50% success.

Stanford and SAP built a benchmark called CooperBench to test something basic: can two AI coding agents collaborate on the same codebase?

They created 652 tasks across 12 real open-source repositories. Each task gave two agents different features to implement—features that were logically compatible but required coordinating on shared code. The kind of thing any software team does every day.

Across all models, cooperation dropped success rates by 30% on average.
Researchers called it "the curse of coordination."

Results read like a dysfunctional team’s retro…

Without strict design, multi-agent systems ignore expertise in favor of middling comrpomise.

Your company spent millions on AI. Nobody's using it. Three-quarters of companies are stuck. And they’re solving the wrong problem.

BCG surveyed 1,000 CxOs across 20 sectors. Only 26% got past proof-of-concept to generate actual value. BCG found that companies throw IT at operational problems. They throw change management at technical gaps. They blame "culture" when the workflow design is broken.

It's like trying to fix a car that won't start by repainting it, then blaming the driver for not turning the key hard enough.

For the past three years, I've been seeing the same repeating failure patterns. If something’s holding your team back, it’s probably some gross cocktail of these:

New NBER study just dropped that changes how you should think about the agents you're building into your business.

What they did:
Researchers ran the most comprehensive behavioral economics test on AI to date—16 experiments originally designed to document human irrationality, applied to 12 frontier models across GPT, Claude, Gemini, and Llama families.

Same tests Kahneman and Tversky used to prove humans violate Expected Utility theory. Loss aversion. Probability weighting. Hyperbolic discounting. The classics.

What they found:
Advanced LLMs exhibit a split personality.

5 Lessons from Uber's Airport Forecasting (That Apply to Every Org)

Uber has a problem at airports:
Too many drivers = wasted time waiting.
Too few = riders can't get cars.
Both kill the marketplace.

At peak times, drivers can wait 45+ minutes in queue. During slow periods, riders wait 20+ minutes for pickup. Neither side wins.

They published research on predicting demand and managing driver queues. Buried in the technical details were lessons every Product and Ops leader needs.

New study on AI use in 41.3M scientific studies: Scientists using AI publish 3x more papers, get 4.8x more citations, become team leaders 1.4 years earlier. But the breadth of subjects studied shrank 5%. Collaboration dropped 22%.
Everyone is sprinting toward the same finish line.

Now your innovation team has the same problem.

What are we actually selling? Working on pricing for a CPG client the past few weeks, and the math keeps breaking the same way. Cost of goods is too high. Can't price competitively. Every model we run shows we're 30-40% above competitors.

Classic consultant answer?
"We need to lower COGS or we're dead."

But that's not the important question.
The question is: What are we actually selling?

Developers are 12 months ahead on AI adoption. 
Here's what they learned the hard way:

UC San Diego studied 112 experienced developers using AI coding agents in real work. The lessons translate directly to every desk job.

What the hell are RLMs? What’s Ralph Wigguming? Welcome back. If you’re lucky, you only have to catch up on 2700 unread emails AI’s entirely new jargon layer.

While everyone was arguing about Claude vs ChatGPT over the holidays, engineers were quietly building infrastructure that actually matters. Here's the cheat sheet…

Professional developers don't vibe code. They control code.
New research shows exactly how…and why it matters for everyone else.

UC San Diego researchers studied 112 experienced developers (3-25 years) using AI coding agents in real work. The headline contradicts the hype: Pros don’t delegate. They direct.

Science says you're terrible at giving gifts.

Researchers have spent decades studying gift-giving. The findings flip everything you think you know…

Adding "AI skills" to your resume might tank your interview odds.
Researchers sent 1,185 identical resumes to employers. Half randomly included AI qualifications.

The result?
No overall effect on callbacks.

But the breakdown reveals something strange…

Good news: Artificial Super Intelligence won’t want to replace you, it’ll want to hire you.

New economics paper flips the script on AI doom. 
Not because ASI will be aligned. Because it will be rational.
It’ll understand you're more valuable thriving than unemployed.

Think about Dune's spice economy. 
The Emperor doesn't glass Arrakis—he needs the spice to flow. 
When your wealth depends on someone else's productivity, you invest in their success.

ASI will run the same calculation.
You’re worth more to it employed.

Three forces could create surprisingly good outcomes:

New MIT research ran a simple experiment with 667 people:

First, answer questions alone (math, physics, moral reasoning).
Then, answer similar questions with AI assistance.

Everyone got the same AI.
But got huge differences in their performance boost.
Some people’s accuracy jumped from 55% to 85%.
Others barely moved.

What separated them wasn't technical skill.

My feed is painting an uncomfortable picture:

UC Berkeley research on 306 production AI agents: Everyone’s building agents in “safe mode” to drive efficiency.

Calvin Cheng data on what drove valuation for 5,700+ companies, 2022-2025: To create value you need to balance investments in efficiency and growth. 

Before you open your mouth for your next presentation, your deck already told them whether you're winning or losing.

Researchers just analyzed 12,000 European paintings from 1600–1820 and discovered that color composition predicted economic growth as reliably as satellite imagery does today.

Bright, saturated palettes dominated boom times. Muted earth tones appeared during wars and recessions. The pattern held across five countries for 220 years.

Here's the color code:
Want to signal prosperity? Use blue. Green. High brightness. Saturated color.
Signaling contraction? Use muted browns. Desaturated palettes. Low brightness. Earth tones dominate.

Now look at today's brands through that lens:

Stripe's saturated purple gradient? Prosperity signal. "We have resources. We're confident."
Figma's electric palette? Same energy. Slack's original bright hashmark said "growth mode." But watch what happened when Slack got acquired. The palette got muted. Softer purples. Less pop.

The Jaguar rebrand everyone dissed last year? They ditched the leaping cat for abstract pastels and earth tones. The internet called it "a spa logo." The color research says something sharper: they accidentally signaled retreat.

Meanwhile, Robinhood stayed electric green through every crisis. Cash App kept its bright palette when crypto collapsed. The colors said "we're still here" even when the headlines didn't.

—Now check 2025's design forecasts—
Pantone's color of the year is Mocha Mousse—a warm brown. Benjamin Moore chose Cinnamon Slate—deep plum. Design blogs are calling it "cozy" and "grounding."

The historical data calls it something else:
A contraction palette.

But there's a counter-trend. Dopamine colors. Hyper-saturated gradients. Optimistic yellows. The design world is split down the middle—half bracing, half betting on growth.

How can you use this in your next deck?
→ Frame the problem in earth tones. Muted. Serious. "Here's the challenge."
→ Shift to saturated colors for the solution. Brightness says "we have a way forward."
→ Use blue when you need trust. Green when you're talking growth.
→ If your whole deck is muted, you're accidentally saying "we're in survival mode."

A forecast for 2026?
The brands that push toward saturation now are betting on expansion. The ones that stay in earth tones are hedging.

Historically, the color shifts preceded the economic data. Artists' palettes changed before GDP caught up.
If the dopamine colors win the next 18 months, the market is pricing in optimism before it arrives.

Your deck is already making a statement.
Make sure it's the one you want.

—
Source: Colors of Growth, Boerner et al, Nov 25 2025

Most businesses review performance quarterly.
Most marriages never do.

Joel Peterson thought that was backwards. The JetBlue chairman and Stanford professor raised seven kids with a radical idea: run your family like a business. 

When I first heard him say that it sounded…cold.
But, it wasn’t robotic.
It was intentional.

Data backs Peterson up.
Couples who prioritize weekly time together are 15% more likely to be "very happy."
Yet 52% of couples rarely or never have date nights. 

Families eating dinner together 3+ nights have healthier kids. 

Maybe it’s time to track our families, like we track our steps.