It’s my birthday today!
And because I am apparently incapable of celebrating like a normal person, I have decided that the best possible gift is to provoke my nerd friends and network into arguing with me about my fav topics: technology, philosophy, and the question that has been bothering me A LOT lately:
As intelligence becomes abundant, what remains human?
Please consider your thoughts, disagreements, obscure philosophical references, and unnecessarily long comments my birthday presents.
There is also another gift, made by my Team I cannot resist talking about: the new www.wiserwork.ai . Please try it for free & give me feedback; this is the most useful gift!
So when people ask why WiserWork feels different from AI note takers, they usually expect a technical answer.
A better model? Yes; a mix of models for different cognitive levels of tasks. Better orchestration. Sure. Better context. Yes! The best one. The tacit knowledge context, the structured way that we store the collective judgement, not only the data, information and the documented knowledge.
Yes we do all of that.
But the truth is much simpler. The biggest advantage is not technical.
It's philosophical.
Before asking what AI could automate, we asked a much older question:
What should remain human, as intelligence becomes abundant?
Aristotle: Relationships of Utility vs. of Virtue
Aristotle distinguished three kinds of friendship in “Nicomachean Ethics.” Friendships of utility: people who help each other get things done. Friendships of pleasure: people who bring each other delight. And friendships of virtue: people who wish the best for each other and are changed by the relationship. The third kind was, for Aristotle, the only complete form, and the only one capable of transforming the people inside it.
Some things exist because they help us achieve an outcome. Others exist because they change who we become. GenAI has become extraordinarily good at utility. It retrieves information. Writes. Summarizes. Plans. Coordinates. These are remarkable achievements.
But organizations don't become exceptional because they automate utility. Already this now is commoditised. They become exceptional because people develop judgment together.
That distinction became one of the foundations of WiserWork. AI should optimize utility. It should never optimize away human growth.
Do we want to remove all friction with AI?
Two Types of Friction
AI removes enormous amounts of friction. That is good. But not all friction is waste. Some friction creates meaning. We realized exactly the same thing happens inside organizations. There is friction that wastes people. And friction that forms people. And organizations.
When I say organizations I mean the collective judgement and taste on which the entire organization aligns to.
a. Utility Friction
This friction wastes human potential: Searching, Scheduling, Remembering, Writing follow-ups, Finding documents, Repeating context, Rebuilding decisions.
This should disappear.
b. Formative Friction
This friction creates better organizations.
Disagreement. Judgment. Taste. Leadership. Trade-offs. Shared principles. Vision. Commitment. This should remain.
Not because it is inefficient. Because it is precisely how collective intelligence develops.
What kind of company are we? Choice Decisions Vs Optimization Decisions.
The Bottleneck Is No Longer Answers. It Is Choices.
AI is exceptional at optimization: finding a faster route, improving a process, comparing alternatives, or generating thousands of plausible answers.
But the most important company decisions are not optimization problems. They are choice problems.
Even the most optimised companies in the world; the AI-labs themselves, select answer differently when making their Choice Decisions, even if they play with a similar Optimization SkillSet --> similar Optimization Decisions: pretty similar model capabilities; every next model has a shelf life of one month, they feed their models with similar data sets from the same data companies; Mercor, Micro1, Handshake and 20 others, they fundraise gigantic Series C,D, E, F.... rounds from similarly structured Valley Growth-stage VCs and the similarity list on the Optimization capabilities goes on.
but they are making distinctively different Choice Decisions on what future they are currently building:
Anthropic is laser-focused on entrerprise solutions serving the different departments; finance, operations, customer success, legal and multiple professional roles, focusing on governance, safety and business value,
OpenAI is a b2c commercially focused company that has even entered the hardware space for indiciduals & households, aiming to provide everyone with their personal AI agent.
Grok/xAI is more opinionated, real-time (due to its x feed) AI, which resembles its creator mindset!
All of those are quite different futures that are the products of similar Optimization Decisions but quite different Choice Decisions.
What product should we build? Which customers should we prioritize, and primarily which ones are we willing to disappoint? What pricing model reflects the company we want to become? Which opportunity deserves years of our lives? Which mountain is worth climbing?
AI can show us more believable futures. It cannot decide which future should be ours.
A mistake that a lot of professionals make is that we think that this is the strategic decisions that we make once a quarter or once a year.
Wrong.
With AI we have Choice Decisions every day
Every day leaders make dozens of micro-decisions.
Product. Every week we have a battle on feature prioritization; we have to remain laser focused on how every next micro-feature, which bug we fix next, what improvement we make, especially now that execution became so fast.
Growth. Every day we decide what customers we say yes, what customers we say no to. We are not good for everyone, e.g. we have so many individuals coming to us and say we want to try wiserwork but I am an individual contributor. We recommend other tools or stay with Google Gemini. If they become part of a team or they are solo-preneurs yes, we accept them.
Hiring. Every month and lately every few weeks we do make the most important decision: who is the next member in our team? Are we hiring for skills or for extreme agency, ownership and ethics? The latter. Of course if candidates excel on both fronts, they make our life easier!
Every day leaders make dozens of micro-decisions. AI doesn't reduce them. It multiplies them.
Optimization problems Vs Choice problems. AI dominates optimization. Humans remain responsible for choices.
Human in the loop, because the model is not yet good enough…?
In many AI narratives, especially here in the Valley, human participation is treated as a temporary inconvenience.
The human is “in the loop” because the model is not yet good enough.
Once the model improves, the assumption goes, the human should disappear.
We reject that assumption. We argue that participation can be a permanent source of capability because human beings possess evolving knowledge, goals, and judgment that the model needs.
Human participation becomes more valuable as AI capability rises.
Because people no longer need to spend as much time performing the mechanical parts of the work. They can focus on the parts where human participation actually matters.
For example, in a meeting today, much of the cognitive load is wasted on remembering what was said, taking notes, converting notes into tasks, formatting the output, writing the follow-up, and assembling a first draft.
Our AI performs these functions (most of the times better than any competitors :) , but the meeting does not necessarily disappear. Quite the opposite; the meeting becomes more human. Here is how we/”AI-first teams run meetings”: https://www.linkedin.com/pulse/how-ai-first-teams-run-meetings-yiorgos-nikoletakis-cafwf/
People can concentrate on formative disagreement, prioritization, alignment commitment & direction, judgment & taste
This allows humans to return to the center of the system at a higher level. AI removes humans from the machinery of execution. It returns them to what humans are best at:
Judgement and taste on the big strategic macro-decisions and the daily micro-decisions
and then, when we speak at the team level
finding alignment and direction on the team’s top performers.
You want to have your best experts in the room. And when you have experts they are opinionated. A good meeting is not when everyone agrees but when disagreement is formative friction and leads to collective judgement & institutional wisdom (as opposed to just individual intelligence and institutional knowledge).
Viktor Frankl: Meaning Comes Through Encounter. Not All Friction Should Disappear
Frankl argued that meaning cannot simply be delivered. It emerges through encounters: with work, with love, with responsibility, with suffering, with realities that resist us.
Organizations work the same way. Great companies are not built because everyone agrees.
They are built because people wrestle with difficult trade-offs until they discover better answers together. The purpose of AI is not to remove those encounters. It is to remove everything that prevents them from happening. And then provide the technology that makes this friction formative.
Michael Polanyi’s “Tacit Knowledge”: We Know More Than We Can Tell
Polanyi introduced the idea of tacit knowledge. The most valuable knowledge inside an organization is rarely written down. It lives inside people’s judgment. Their intuition. Their sense of taste. Their understanding of context. Ironically, this tacit knowledge appears most clearly inside conversations. Rarely inside documents. Meetings contain something documentation almost always loses: why a decision was made. What assumptions were rejected. Which principle ultimately mattered. That’s why searching Google Drive is not enough. The company’s living intelligence exists inside its conversations - inside its non documented knowledge and wisdom, at the pinnacle of this “Theory of Knowledge” pyramid model:
Frontier models have become extraordinary at explicit knowledge. They are rapidly absorbing everything documentable. But the highest levels of the pyramid remain deeply human. Wisdom. Collective Judgment & Taste that form the company's Decisions. Of course augmented with AI & domain-specific Agents.
Stelios Ramfos: Formation Happens Through Friction
Ramfos often returns to a simple but profound idea. Human beings are not formed by comfort. They are formed by encounter. By resistance. By the other person. By realities that refuse to bend completely to our will.
AI changes this equation. For the first time in history, we are creating systems designed to adapt almost perfectly to us.
That is extraordinary. It is also dangerous. Because a world with no resistance may become a world with less formation.
"Your wife was not designed to agree with you."
Neither is your son. Neither is your best engineer. Neither is your co-founder.
Their resistance is not a product bug. It is part of how they change you. The question for AI isn't how to remove resistance. It's how to remove everything except the resistance that makes people & organizations wiser, more experienced (an experienced person, might be a very young one who has 2-3 years of experience in high growth, high formative friction environments, Vs an old professional in a low paced working environment).
In different words: “Struggle is the mother of innovation & learning”.
We need to develop the willingness to be changed by people who are not designed for us.
That Changed Everything
At the moment we made this realization, we stopped thinking we were building an AI note-taker or simply an assistant.
We realized we were building a system that acts as a mentor who protects the organization’s public space & time for formative friction,
the AI agent that empowers AI-First teams form strong alignment & direction so that they execute very fast,
not for the sake of speed but for immediacy that increases the pace of iterations. Of aligned iterations, which make our users outcompete anyone in their markets.
Our job is not to eliminate meetings. Our job is to eliminate everything that prevents meetings from producing better thinking. The deliverable is not the transcript. Many times the deliverable is the consensus (see again our “How AI-first teams run meetings”: https://www.linkedin.com/pulse/how-ai-first-teams-run-meetings-yiorgos-nikoletakis-cafwf/
The deliverable is better judgment.
A Different Philosophy of AI
Technology did not lead us here. Philosophy did.
The future of AI is not about making humans unnecessary.
It is about deciding which parts of being human should never be automated. Of course they will be augmented by AI. Everything else is implementation.
Most AI note takers treat the meeting in an admin way as something that happened in the past and ask: "How can we summarize this meeting for every individual?"
WiserWork asks: "What is this organization learning?"
Most AI retrieves information. WiserWork retrieves evolving judgment. Our agent's soul/character is to trigger the team to form collective judgement. Very fast. Faster than your competitors.
Most AI remembers what was said. WiserWork remembers what the team currently believes.
Remove the friction that wastes people.
Preserve the friction that forms people.
This makes us human.
