Hello, Lovely Human,
This is probably the most common problem I have with experts who join our community. For some reason, all the versions of them are meshed into one. And it gets confusing. They then wonder why they are not getting hired as advisors.
I do this exercise with many of them. I go to their LinkedIn profile and I ask, "Tell me exactly what is the one thing, screaming out loud, that you know and do like nobody else, that I would be willing to pay for an hour with you?" We go through their profiles and they all seem to have at least 5 to 10 things. Choice overload is something proven in behavioral science. Offer shoppers 6 flavors of jam and 30% of them buy. Offer them 24 flavors and only 3% do.
I completely understand many of us Gen Xers have three, four and even five decades of scars and knowledge. And while we are all eager to show how much we know about everything, basic behavioral science says focus on just a few. I often say 1 or 2 max, and within the same skill or domain, so nobody has to think too much.
Why is it we have such an issue standing tall in the current, limited version of ourselves? Even if it's a work in progress.
I recently had this show up in another way with our data lake. Some of you may know a data lake is just a large collection of data, with pipelines constantly feeding it automatically. I have been building and tagging ours for over 7 years. It is where the exit value of GLEAC sits: 400k pieces of unique, extremely high-judgment human skills behavioral data. It's what every AI model or agent does not have. Human behavior, or tacit knowledge, on why we do what we do on the job in any role, when exceptions apply, what the edge cases are, and so on. You cannot scrape this off the internet, since good judgment, like common sense, is not common. It differs by job, geography, level of human skills and more.
Over the last 30 days I have been having calls with all the major frontier labs and their data suppliers to license our dataset. I could not do it before now, since the market was just not ready. It is now.
On a recent call with a Google CXO, I excitedly shared that we have over 400k pieces of data across 10 human skills, everything it could do, and I listed it all.
After I was done with my spiel, the very generous Google CXO was extremely helpful and candid. He advised I should not talk about the data the way I just had with him. He suggested I segment each piece and talk about it specifically in terms of domain. Good judgment datasets for healthcare. Good judgment datasets for banking. He said that without getting very specific, nobody would really understand our very premium dataset, and they would undervalue it.
I've given that exact advice to hundreds of people about their profiles, and here I was getting it myself… I love these moments.
So the team spent the last 2 weeks repackaging our data lake. Then we tested it. On a call with the largest law firm in India this past week, we spoke about what gets missed when lawyers use Legora, Harvey or Claude: the tacit knowledge, the WHY behind legal decisions. We demoed examples in our playground… and they got it. On another call, with a data supplier to the largest US mutual funds, we demoed the dilemmas investment analysts face, and without us asking, they said, "Let's get you in front of Bridgewater."
Being everything to everyone means being nothing to anyone in particular. So simple, yet we all forget it… even me.
Just remember, you can know and be a lot of things. Be proud of them, since everything else stands on those other versions of you. But you can only be KNOWN for one version of you at any given moment.
🤖 AI + Frontier Updates
1 - Anthropic and Accenture are each putting $1 billion into human judgment.
Accenture's team becomes Anthropic's first embedded evaluator, red-teaming models and testing safeguards from inside the lab over five years. Outside human judgment just got a billion-dollar line item. Somebody credentialed has to sit in those chairs.
2 - A training data company just raised $350 million at a $3.5 billion valuation.
Snorkel AI stopped selling software and started selling ready-to-use datasets built with tens of thousands of human experts. Revenue grew 18 times and passed a $375 million run rate. Expert data is no longer a side business. It is the business.
3 - Google now pays publishers only when their content actually changes an AI answer.
Pay per value, not per use, across Gemini and AI Overviews. The catch: nobody can see how value is scored, and early payouts are being called peanuts. The supplier who can prove exactly what their data changed sets the price. Everyone else takes it.
4 - Retraining on licensed music didn't get Suno off the hook.
Universal and Sony sued again over 60,202 recordings, arguing that a new model trained on an old model's outputs "launders" the infringement. Exposure runs past $9 billion. Data now has a family tree, and buyers are going to ask to see it.
5 - Canadian business? Canada will co-fund your AI project.
Mitacs AI Advantage pairs companies with students and researchers to adopt or build AI, and covers half the cost. Open to SMEs, nonprofits, hospitals and municipalities. The cheapest AI talent you will ever hire is already in a Canadian university.
Watch the full system run live on Sep 30. Walk away ready to do it yourself.
Most founders have LinkedIn traction with nothing to show for it in the CRM. On Sep 30, Maria Gharib (Mindstream) and Valerie Chapman (Ruth AI) walk through the exact system live.
From AI-assisted content creation to sequenced outreach to booked meeting. You'll leave with a process you can run the same day.
Eligible startups also get the LinkedIn-to-Leads Toolkit: ad credits, Apollo, Captions, and HubSpot's Prospecting Agent.
Lovely Humans in our Community
How Can I Stay Ahead
in Learning & Development?
📚
1 - The book I'd give anyone who can't explain what they do.
April Dunford's Obviously Awesome was written for products. It works frighteningly well for people, and for datasets. The Google advice in this week's essay is chapter one: decide what you are the best in the world at, and for whom.
2 - Your tacit knowledge is your next competitive moat.
Two Accenture leaders in Berkeley's California Management Review on why the reasoning your people carry in their heads is worth more to AI than your data or your tech stack, and how to map it before it walks out the door.
3 - Thinking of licensing your data to an AI lab? Read this first.
Mozilla's plain-English checklist: define exactly which models can use it, restrict redistribution, demand deletion, build in audits, and get paid fairly. Eleven clauses that separate a partnership from a donation.
4 - Who actually supplies the humans behind AI? A map.
From Scale, Surge and Mercor to the new reinforcement learning environment builders, this is a clear picture of how the data supply chain is tiered, and why the big labs deliberately spread their work across many vendors. Neutrality is becoming a selling point.
5 - The science of asking better questions.
A comprehensive review of the psychology behind why, when and how we ask questions. Useful for anyone whose value lies in the question they ask before the answer anyone else gives.
Religious Focus this month: Ubuntu
Every month we pick one religion or way of living. It allows us to open our minds and be mindful of how we and others live better, healthier, joyful and meaningful lives.
Last week we looked at ukubuyisa, bringing someone back into the fold. This week, the place where Ubuntu becomes poetry: the praises a community calls out loud about who you are.
The school hall in Pietermaritzburg is decorated with streamers the Grade 9s clearly argued about. It is Zanele Mthembu's last day as principal after nineteen years.
There have been speeches. The district official read from a sheet. The deputy principal cried twice. Zanele has smiled through all of it and not recognised herself in a single word.
Then her nephew Lwazi, twenty-two and studying engineering in Durban, walks to the front in his university hoodie and starts to call her praises.
His voice changes. Faster, louder, the rhythm of the imbongi, the praise singer.
He calls her the teacher who arrived with one suitcase and a box of chalk. The one who stood at the gate the year of the taxi strike and walked the smallest children home herself. The one who fought the department for a science lab, won, and then fought again for someone who could teach in it.
Then he says a line nobody has heard before.
"Daughter of the gate, you are closing it behind you now. Open the next one."
The hall goes quiet. Zanele laughs, and then she doesn't.
These are izibongo, praise poetry in the isiZulu tradition. A person's praises are not a CV. They are the handful of lines a community uses to hold who you are: your deeds, your character, the moments that made you. They are given by others, performed out loud, and added to as a life moves on. A young person's praises are short. An elder's grow long, verse by verse, one chapter at a time.
That last part is what Zanele hadn't expected. The praises didn't mesh all her versions into one. They named each one, in order, and then they let the last one end.
For months she had been dreading the question everyone asks at a retirement: so what will you do now? She had been answering it as the principal. Lwazi had just answered it for her, as the next version.
Izibongo work because they are brief enough to be remembered and repeated when you are not in the room. That is the whole point. A praise nobody can repeat is not a praise.
On Monday, Zanele signs up to tutor matric maths at the library on Saturdays. She tells the librarian she is "a teacher who used to run a school." It is the first time the sentence has felt true.
Guiding question:
If the people who know you best had to call your praises today, which version of you would they name, and is it the one you are living now?
This Week's Practice
Three steps. Your versions.
1. Draw your chapters.
On one page, list the versions of you, each with a rough start and end. Name them the way someone who loves you would, not the way your CV would.
2. Circle two.
The version you are living now, and the version most people still introduce you as. If they are different circles, that is the gap you have been feeling.
3. Write your praise line.
One sentence, the way someone would call it out in a room. Send it to two people and ask, "Is this who I am to you now?"
Zanele spent nineteen years as the principal. It took one line from her nephew to let the next version begin.
🍀 What Can I Do to Impact, Live Sustainably and Make the World a Better Place?
1 - New York has a food waste problem. South Korea already solved it.
Korea recycles almost all of its food waste by making households pay by weight for what they throw away. One simple rule changed the habits of a whole country. Sometimes the solution already exists, it just lives somewhere else.
2 - Kosovo has finished clearing its landmines. It took 27 years.
In the first month after the war, around 150 people were killed or injured by mines, most of them under 24. Today that land can be farmed and walked again. Proof that patient, unglamorous work eventually finishes the job.
3 - Every A&E in England now has to listen when a family says something is wrong.
Martha's Rule gives patients and families the right to an urgent second opinion. It is named after Martha Mills, who died of sepsis at 13 after her parents' concerns were dismissed. One grieving family changed a whole system.
4 - A quiet meeting in Muscat may have kept one more front from opening.
US officials met Houthi leaders in Oman, heard that the ceasefire on American shipping still holds, and decided against backing Saudi Arabia militarily in Yemen. Oman keeps being the room where enemies talk. In our region, that room matters more than most people realise.
5 - Meet the first new wild cat species named in over 100 years.
Your delight for the week. Leopardus tilcayo was found living in a Bolivian sanctuary after a local family took it in. It was hiding in plain sight until someone looked closely enough to see it was one of a kind.
