The Deep Feed

For the last two years, we have been obsessed with the conversational interface. We treat Large Language Models like digital companions, typing long prompts and waiting for a polite, wordy response. But this interaction is fundamentally inefficient. If you want to move a grocery item from a voice command to a digital cart, you do not need a philosopher; you need a router. This is where the shift from generative AI to decision engines begins. The goal is no longer to produce text, but to produce action with zero latency.

The Speed of Certainty

John Lindquist’s work with Jev illustrates a move toward what might be called 'system-one' thinking for machines. In psychology, system-one is fast, instinctive, and emotional. In software, it is the ability to take a messy, unstructured input—like a spoken sentence—and instantly map it to a specific function. When Lindquist demonstrates a real-time voice to-do app, the magic isn't in the chat; it is in the absence of the pause. The system classifies the intent and executes the command before the user has even finished their thought. This is the end of the 'typing' era.

The output is not the product. Customers are not buying the lines of code; they are buying the resolution of their intent.

This requires a different kind of architecture. Instead of one massive, expensive model trying to do everything, the new pattern uses small, lightning-fast models to act as routers. These models don't need to know the history of the French Revolution; they only need to know whether the user wants to 'add milk' or 'check the weather'. Once the decision is made, the system can then call a more specialized tool. It is a hierarchy of intelligence that prioritises cost and speed over breadth of knowledge.

The implication for agency owners is clear: stop trying to build 'smarter' chatbots and start building faster decision engines. The value is moving away from the ability to generate content and toward the ability to navigate complexity. The winner in this space won't be the one with the most parameters, but the one with the lowest latency between intent and execution.

When a doctor says the word 'cancer', the air in the room changes. For many patients, it is a moment of sudden, sharp terror—a feeling that the floor has dropped away. Yet, for a significant portion of those diagnosed, the word is a mismatch for their reality. Modern medicine has reached a point where we can identify tumours that are clinically indolent, meaning they pose almost no threat to a person's lifespan or quality of life. We have the technology to treat them, but we lack the vocabulary to describe them without causing unnecessary trauma.

The Betrayal of the Self

Most diseases are external invaders: a virus, a bacterium, a broken bone. They are things that happen *to* us. Cancer is different because it is a betrayal from within. It is your own cells behaving badly. This distinction is not just philosophical; it changes how the body responds to the diagnosis. The psychological weight of knowing that your own biological building blocks have turned against you creates a specific kind of existential dread that antibiotics or simple surgery cannot address.

The noun, the proper noun with the capital letter, still looms, menacing.

In the clinical setting, doctors often try to soften the blow with modifiers: 'low-grade', 'slow-growing', or 'non-threatening'. But these are mere adjectives clinging to a heavy noun. They rarely succeed in neutralizing the terror. As we move toward more precise, molecular-level understandings of biology, our language remains stuck in a blunt, Victorian era of categorisation. We are using a sledgehammer to describe a scalpel's worth of precision.

The consequence of this linguistic failure is real. Patients, gripped by the 'C-word' anxiety, often rush into aggressive treatments—radiation, surgery, chemotherapy—that may be unnecessary. They trade long-term side effects for short-term psychological relief. We are treating the word, rather than the pathology. To fix this, we need a new taxonomy that separates the biological presence of abnormal cells from the clinical necessity of intervention.

Updating our medical vocabulary is not about being 'soft' or avoiding hard truths. It is about accuracy. If a patient is told they have a 'grade group 1' prostate tumour, they should understand that this is a biological observation, not a death sentence. Precision in language is as vital to patient outcomes as precision in the operating room.

There is a specific kind of person who exists more as a collection of anecdotes than as a coherent human being. In Belgrade, this person was Zoran. To some, he was an exceptional, intelligent man; to others, he was a person who simply never felt shame. When filmmaker Maja Penčič set out to find him, she wasn't just looking for a man; she was looking for a version of the truth that could survive the conflicting testimonies of those who knew him.

The Fragmentation of Identity

The documentary uses a home-video aesthetic to mirror the way memory works: it is intimate, slightly surreal, and deeply imperfect. As the cast of acquaintances rotates, Zoran's character shifts. He is not a fixed point. Instead, he is a shape defined by the people standing around him. This reveals a fundamental truth about human social structures: we do not remember people; we remember our interactions with them. We curate our memories to fit our own narratives of who we were when we were with them.

The film never settles on a single version of events, instead watching the details rearrange themselves around a few fixed points.

This process of reconstruction turns a living person into a cautionary tale. The more people speak, the less we actually know about the man himself. The details become a comedy of errors, a tragicomedy where the 'truth' is merely the average of a dozen different biases. Zoran becomes a vessel for the observers' own judgements, their own frustrations, and their own need to categorise the unclassifiable.

In the end, the search for Zoran suggests that the 'truth' of a person is an impossible metric. We are all, in some way, a composite of the stories told about us. The man who 'destroys everything' might have been a man who simply lived too loudly for those around him. The documentary forces us to confront the fact that our social reality is built on these shifting, unreliable foundations.

There is a specific kind of identity crisis that arrives with grey hair and a sudden, unexpected role in the lives of others. For many men, the transition into elderhood is marked by a withdrawal from the active world—a move toward the role of the observer. But there is another path: the 'full-service grandad'. This is not a role of leisure or storytelling, but one of labour. It is the decision to engage in the messy, unglamorous work of childcare, from changing diapers to managing nap times.

The Labour of Presence

The author notes a sharp distinction between being a 'present' parent and being a 'full-service' grandparent. One can be physically there, providing stability and structure, while still being emotionally distant. The full-service role requires a different kind of surrender. It requires the willingness to be useful in the most basic, biological ways. It is a rejection of the 'grandfather as sage' archetype in favour of the 'grandfather as caretaker'.

The full-service grandad does all the routine labours of care, but must also ask what distinctive contributions he brings.

This shift is often accompanied by a humbling of the ego. A man who has spent decades building a career, perhaps as a professor or an executive, must suddenly find value in the repetitive, often exhausting tasks of a three-year-old's life. The 'project' of a child is not something to be observed from a distance with academic curiosity; it is something to be lived through, side-by-side, in the dirt and the rain.

Ultimately, the value of this role lies in what it allows the elder to see. By engaging in the routine, the world is viewed through a new lens. The sparrow, the puddle, the worm—these are not just objects of interest for the child, but reminders for the adult of the immediacy of existence. The clock is ticking on this period of service, and the goal is to use that time to reconnect with the fundamental textures of life.

We are entering an era of unprecedented abundance in software production. When a model can generate functional code in seconds, the cost of building a product approaches zero. This creates a dangerous paradox: it has never been easier to build something that nobody wants. In this new landscape, the traditional skills of product management—writing requirements, managing roadmaps, and coordinating engineering sprints—are being devalued. The bottleneck is no longer production; it is selection.

The Rise of the Judge

If anyone can ship a feature, then the value of a Product Manager shifts from 'how do we build this?' to 'should we build this?' This is the transition from builder to judge. The most successful product leaders of the next decade will not be those who can manage the most complex workflows, but those with the most refined taste. They will be the ones who can look at a thousand AI-generated possibilities and identify the one that actually solves a human problem with elegance.

The output is not the product. Customers are not buying the lines of code.

This requires a move away from the 'feature factory' mentality. In the past, success was often measured by velocity—how many things did we ship this quarter? In the AI era, velocity is a commodity. High-speed shipping of mediocre ideas is just a way to create noise. The new playbook emphasises quality, intent, and the ability to anticipate user needs before they are articulated. It is about the 'clever touches in the details' that signal a deep understanding of the user's life.

The winners will be those who treat AI as a factory for ideas, but maintain a strict, human-led gatekeeping process. We must resist the urge to ship everything we can imagine. Instead, we must focus on the hard work of deciding what is worth our attention. The moat is no longer the code; the moat is the standard of excellence you refuse to compromise.