AI Adoption: Emotional, Practical, and Complex

AI adoption is moving fast, but the “why” behind it is far more complicated than a single metric can capture. That’s why we’re excited to partner with Outset on AI P.U.L.S.E., a new monthly index designed to uncover the emotions, tensions, and real-life behaviors shaping how people use AI. Join Daniel and Aaron on July 28 for a first look at Wave 1. 🚀 https://lnkd.in/eHbGUWXr

You are not behind on AI. I need more people to hear that, because the fear is louder than the facts. Here is where AI actually is right now, in plain numbers. The US Census Bureau tracks this every two weeks. As of May 2026, about 1 in 5 American businesses are actually using AI in their operations. Right around 19.8 percent. Read that again. Four out of five businesses are not deep into this yet. So if you feel like the whole world passed you by, it did not. Most of the room is standing right where you are. Now the other half of the reset. What AI even is. Strip away the buzzwords and modern AI is a prediction machine. It learned patterns from a giant pile of examples, and it guesses what comes next. The next word. The next number. The next step. That is it. It is not alive. It is not thinking about you. It is not plotting anything. It is a very good pattern guesser that you point at a job. When you see it that way, the fear gets smaller and the question gets simpler. Not "will this replace me." Just "what job do I want to hand it first." At PeakAgent, that is the whole starting point. We do not throw ten tools at you. We help you get comfortable, then build one assistant that slowly learns you, your pricing, your voice, and how your business actually runs. You are not late. You are right on time. Curious where you actually stand with all this? Drop a comment or send me a message. No pitch, just a straight answer.

There is a growing comfort with average output, as long as it arrives quickly and requires minimal effort to produce. I am not sure when average became acceptable. The current culprit is AI. Used well, it is amazing! The issue is what happens when AI becomes a substitute for thinking rather than a support for it. I went into further details in my latest Substack. Here were some of the key considerations I shared: 1. Before using AI, form your own view first: Even a rough one. Use the tool to refine and pressure-test, not to replace the initial thinking 2. Ask the question AI cannot answer for you: what do you actually want the audience to think, feel, and do as a result of this? That answer has to come from you 3. Review AI output critically: The default quality is average. Your job is to push it past that before it goes anywhere near a client or a public channel 4. Build real knowledge: Read widely, have real conversations, form opinions on things and write them down. These are the inputs that make strategic thinking possible 5. Protect relationships: pitches, proposals, follow-up conversations are human on both sides of the coin. Average is just not good enough. If you care enough, you’d be sure of that too. What have you noticed?

AI’s Next Leap May Not Be More Intelligence—It May Be State. Over the past few days, I’ve had an unusual opportunity to reflect on AI while managing a complex, evolving real-world situation. It led me to a realization: The next major advancement in AI may not come from larger models or more parameters. It may come from how AI represents state. Today’s AI excels at understanding language, answering questions, generating content, and synthesizing information. But many real-world problems are not isolated questions—they are dynamic systems that evolve over time. Think about healthcare. A patient’s condition isn’t just a collection of symptoms or test results. It’s an evolving clinical state. Every new observation, laboratory result, medication adjustment, and physician assessment changes the current picture and influences the next decision. The same principle applies across many domains: * A telecom network is an evolving operational system. * A software project progresses through changing requirements, dependencies, and milestones. * A legal investigation develops as evidence is gathered and hypotheses are tested. * A business strategy evolves as markets, risks, and priorities change. In each case, success depends on understanding the current state, how it arrived there, and what events are likely to change it next. This makes me wonder whether future AI systems should combine language models with explicit state management—maintaining structured representations of evolving situations rather than relying primarily on conversational context. Imagine an AI that continuously updates a living model of the problem: * Current state * Active hypotheses * Evidence supporting each hypothesis * Planned actions * Outstanding decisions * Confidence levels * History of how the state evolved Instead of simply remembering what was said, it would understand where the system currently is. Language models have transformed how we interact with information. Perhaps the next leap is enabling AI to reason over evolving state, making it a more effective partner for solving complex, long-running problems in engineering, healthcare, science, and beyond. Complex problems are not just conversations. They are evolving systems. I’m curious to hear how others see this. Do you think explicit state management is a key missing piece in the evolution of AI?

Researchers mapped leading AI models onto the World Values Survey. 88 countries, real people, two axes: traditional vs secular, survival vs self-expression. The AI models didn't land in any human cluster. They came out more secular and more self-expression oriented than Sweden. The most liberal country in the dataset. That's territory no human population occupies. And it's not a fluke of one model. Prompting a model to "answer like a typical person from Nigeria" shifts the response a bit, but it still stays anchored in its western value cluster. The worldview runs deeper than a prompt. Why this matters for business, not just researchers? When AI judges what counts as a fair salary, what counts as employee initiative, how to respond to a customer from a different culture, it's not doing it neutrally. It's doing it through its own built-in worldview. Companies pick a model on price, speed, benchmark accuracy. The worldview that model brings into every answer almost never factors into that choice. No tool is neutral. Especially one that talks to people across different markets every day.

How AI in Daily Life Is Rewriting Our Routines: Mastering AI Tools 2026 The Invisible Revolution: AI in Daily Life As we navigate through 2026, the integration of intelligent technology into our personal spheres has moved from mere novelty to an essential pillar of modern existence. AI in daily life is no longer about flashy interface changes; it is about the silent, background optimization of our habits, environments, and professional workflows. This shift has fundamentally altered the relationship between humans and their digital tools, turning systems that were once reactive into proactive partners....

I’ve been thinking about artificial intelligence for years, long before it became part of my everyday working life. The original version of this article was written when AI still felt mostly like science fiction. Today, I use it to help build software, support marketing systems, understand technical concepts and turn ideas into working processes. But this piece is not really about productivity tools. It explores a much bigger question: what happens if we eventually create something genuinely conscious? I look at emergence, competing theories of consciousness, the limits of human perception, AI alignment, the incentives surrounding development, and why the environment in which a powerful intelligence emerges may matter as much as the intelligence itself. It also brings the argument back down to earth: AI is already changing what a single person or small business can build, learn and operate, even without anything resembling a true Singularity. #ArtificialIntelligence #FutureOfTechnology https://lnkd.in/eeQbDqER

No. No. No! – the best answer is NO! 😁 How do we use AI? Together with Earl Monte Casho (EMC), we’re looking for possible solutions. Following Warren Buffett’s example, Earl responds to everything with a resolute: NO! “No” is his favourite answer, but I won’t be deterred. Some solutions aren’t immediately obvious, even when using AI. But I still prefer common sense and use AI only as a tool. Perhaps to put it better: I’m looking for the optimal link or bridge between the best AI results and our ability to think (our team’s ability). AI can only somehow amplify or better articulate the ideas that arise within our team. In both simple and more complex cases, I always start with the basics: - product and service portfolio, - future results, - ownership structure, and even if liabilities are much higher than future revenues can cover, I believe there is a solution. Am I being too optimistic? photo Mike Hostinský

I like to shift the conversation away from what AI is to how people choose to use it. Most discussions are about agents replacing people: "I built an agent to do X." My perspective is almost the opposite: I use AI to expand my thinking, not replace it." That distinction is subtle but important. I frame it as a series of contrasts: My goal isn't to automate judgment. It's to improve judgment. Then perhaps: I don't ask AI for answers first. I ask it for possibilities. I use multiple models to expose different perspectives. I deliberately look for disagreement. I ask for the strongest arguments against my current thinking. I connect ideas across disciplines rather than staying in one domain. The final judgment remains mine. And then perhaps end with something like: People often ask how I use AI. I don't think of it as an autonomous agent working for me. I think of it as a multidisciplinary brainstorming partner whose job is to expand my possibility space before I exercise judgment. That last phrase—expand my possibility space—captures something distinctive my work. One thing would avoid is making it sound like agents are inherently bad. There are plenty of tasks where agents are excellent—routine workflows, monitoring, repetitive actions, and well-defined processes. The distinction isn't agents versus conversations. It's that you're using AI for a different class of problems: those where the value comes from better human judgment under uncertainty, not from automating a known procedure. Intelligence isn't just producing an answer quickly; it's improving the quality of the questions, perspectives, and judgments that lead to an answer.

The Compassionate Algorithm: Reimagining AI for Human-Centric Decision-Making In the age of AI, we are all driven by algorithms; every decision we make in this digital age is nudged by algorithms that constantly push us toward a choice. But the real question is what's driving the algorithms. Usually, it’s the hard quantitative data that’s the central driving force behind the algorithm, often missing the contextual nuances and human empathy. The need is to pivot towards the Compassionate Algorithm, which, i.e., moving AI from pure mathematical optimization toward human-centric outcomes. We have often seen that in standard cases, algorithms usually aim to solve for maximum efficiency, speed, or profit—often overlooking human context. In that context, a compassionate algorithm can intentionally encode empathy, equity, and human dignity into its code. How It Can Work Focus on Contextual Variables: The first and foremost, it looks beyond the standard metrics and focuses on contextual variables. For instance, a welfare-allocation AI would not just look at raw income data; it would factor in regional vulnerabilities, family health crises, and systemic barriers. Placing Equity Over Equality: Second, usually, the standard code treats all inputs identically. A compassionate algorithm recognizes historical disparities and actively adjusts to direct resources to the most marginalized first. Automated Yet Human-Centric: Though it serves the automation function, it refuses to let automated systems make absolute, life-changing decisions. The algorithm serves as an advisor, flagging complex cases for human review rather than blindly cutting off assistance or services. Adaptive and Dynamic Feedback Loops: It continuously measures its own real-world impact on human well-being, pivoting when the data shows that its outputs are inadvertently causing stress or harm. #ArtificialIntelligence #ResponsibleAI #HumanCenteredAI #AIEthics #CompassionateAI #AlgorithmicFairness #TrustworthyAI #AIForGood