The Deep Feed

01 — Lenny's Newsletter

The Death of the App: Designing for Conversation

How Meta’s Muse moves away from buttons and towards relationships

By Lenny Rachitsky · 12 min read

Editor's note: A look at how the next generation of software will feel less like a tool and more like a person.

The era of clicking through menus is ending. For decades, software design has been a series of commands: click this, drag that, select from a dropdown. It is a transactional relationship. But as artificial intelligence moves from a novelty to a utility, the interface is shifting from a dashboard to a dialogue. Meta’s Muse is a primary example of this transition. Instead of building a complex suite of features, the designers focused on a single, persistent relationship. The goal was not to build an 'AI app' but to build something that feels like messaging. This is a fundamental shift in how we interact with machines. We are moving from being operators of tools to being participants in conversations.

The Three Layers of Understanding

One of the hardest problems in AI design is the spectrum of user expertise. A casual user wants a simple answer; a power user wants to see the logic. Muse solves this through progressive disclosure. It offers three distinct levels of interaction. At the top, there is a simple emoji that communicates the agent's status. Below that, a task list provides more detail for those who want to know what is happening. At the deepest level, there are raw logs for engineers. This prevents the interface from becoming cluttered for the average person while ensuring the technical user is never left in the dark. It is a way to manage complexity without making it visible to everyone.

AI agents should be designed for conversations with people, not interactions with software.

This approach extends to how the agent handles data. Most software requires you to input information manually. Muse, however, can ingest existing digital lives. Alex Cornell, the design director, demonstrated how the agent could take years of saved Instagram posts and transform them into a structured shopping catalog. It turns forgotten data into immediate utility. This isn't just about convenience; it is about the agent becoming a continuous extension of the user's own memory and preferences. The machine isn't just responding to prompts; it is anticipating needs based on a long-term context that a standard app could never maintain.

Design Principles for the Agentic Era

Prioritise persistent relationships over isolated threads
Use progressive disclosure to manage technical complexity
Treat personal data as a continuous context, not a one-off input
Design for conversation, not command execution

There is also a psychological dimension to this design. When an agent works, it shouldn't feel like it is 'calculating'. Many products show 'thinking' steps as a way to hide latency, but Muse aims for speed that allows the agent to work quietly. The interaction is meant to be seamless. Even the way users communicate with the agent matters. Cornell noted that polite, specific prompting often yields better results than frustration. We are, in effect, learning how to manage a new kind of social entity that exists entirely within our screens.

Ultimately, the success of these agents will not be measured by how many features they have, but by how much they disappear into the background of our lives. If the interface is perfect, you don't notice it. You only notice the person—or the intelligence—on the other side of the chat. The transition from software to agent is the transition from tool to companion.

Key Takeaway

The future of software is not a better menu, but a better conversation.

02 — Lenny's Newsletter

The Lab Mindset

Why the most successful companies are becoming experimental engines

By Lenny Rachitsky · 10 min read

Editor's note: In an age of rapid technological change, the ability to fail fast is more valuable than the ability to plan perfectly.

The traditional corporate structure is built for efficiency and scale. You find a product that works, you optimise it, and you defend its market share. But in the era of generative AI, this model is dangerous. The speed of change means that what is a market leader today could be an antique by next year. To survive, companies must stop acting like factories and start acting like laboratories. This means shifting the focus from execution to experimentation. A lab mindset accepts that most ideas will fail and builds the infrastructure to find the few that work.

The 'Almost Possible' Framework

Josh Woodward of Google Labs argues that the best ideas come from exploring the 'almost possible'. This is the space where technology is just on the edge of being useful. It is not about what is currently easy to build, but what will be inevitable once the technical hurdles are cleared. Working in a lab requires a specific type of intuition—the ability to recognise genuine product-market fit before the market even knows it wants the product. It requires looking at early, messy signals and seeing the potential for a massive shift in human behaviour.

The most important skill in a lab is knowing when to quit.

This brings us to the hardest part of innovation: the kill switch. In a standard company, killing a project is seen as a failure. In a lab, killing a project is a successful part of the process. If you spend six months pursuing a dead end, you have wasted resources. If you spend two weeks identifying it as a dead end, you have succeeded. The goal is to increase the velocity of learning. The faster you can cycle through hypotheses, the faster you reach the truth. This requires a culture that rewards the identification of failure as much as the achievement of success.

How to Build an Internal Lab

Decouple the team from standard quarterly KPIs
Prioritise speed of learning over speed of shipping
Empower small, autonomous units to test radical ideas
Create a formal process for killing unviable projects

This shift also changes the nature of roles. The lines between engineer, designer, and product manager are blurring. The new essential role is the 'builder'—someone who can navigate the entire stack of idea, prototype, and test. It is less about specialised mastery and more about the ability to move through the entire cycle of creation. In a lab, a specialist who cannot iterate is a bottleneck. The value lies in the ability to bridge the gap between a theoretical possibility and a functional prototype.

As AI continues to lower the cost of creation, the barrier to entry for experimentation will drop. The companies that win will not be those with the most resources, but those with the highest rate of experimentation. They will be the ones that treat every product launch not as a final destination, but as a single data point in a much larger, ongoing experiment.

Key Takeaway

Efficiency is for maintaining the past; experimentation is for building the future.

03 — Psyche

The Weight of Being

Racial trauma and the misinterpretation of Black male distress

By Yamonte Cooper · 15 min read

Editor's note: A necessary examination of how structural racism manifests as psychological injury and is often misread by society.

In clinical settings, the symptoms of distress are often misdiagnosed when they are actually responses to environmental reality. A Black man may present with irritability, social withdrawal, or hyper-vigilance. To an untrained observer, these look like general anxiety or personality flaws. However, when these symptoms are viewed through the lens of sustained exposure to racialised threat, a different picture emerges. This is not a series of isolated incidents; it is a chronic psychological injury caused by a world that frequently treats Black men as threats rather than citizens.

The Gendered Nature of Anti-Blackness

The experience of Black men is uniquely shaped by the intersection of race and gender. Stereotypes often link Black maleness with danger, aggression, and criminality. This creates a specific kind of pressure: the need to constantly perform a version of oneself that is non-threatening to survive. This is not an individual choice but a survival strategy. When a Black man is viewed as inherently dangerous, his natural reactions—anger, defensiveness, or even silence—are interpreted through a lens of suspicion. His attempts at self-protection are seen as evidence of the very threat he is trying to navigate.

What looks like emotional shutdown is often a survival strategy.

This is a structural problem, not an episodic one. Traditional understandings of PTSD often focus on a single, traumatic event. But racial trauma is often cumulative. It is the cumulative effect of being followed in a store, being questioned by a supervisor, or seeing violence against others replayed on a loop. It is a constant, low-level state of threat that never truly resets. This constant state of vigilance takes a massive cognitive and emotional toll, leading to exhaustion and burnout that is frequently mislabelled as simple depression.

Misinterpretations of Racial Trauma

Anger is often misread as aggression rather than grief or exhaustion
Vigilance is misread as suspicious behaviour
Withdrawal is misread as lack of engagement rather than a need for safety

To address this, we must move beyond the idea that trauma is only about acute violence. We have to recognise the damage done by the daily, structural constraints placed on Black lives. If we continue to misread these coping mechanisms as character flaws, we fail to provide the actual support required. True clinical and social support requires acknowledging that the environment itself is often the source of the injury.

Recognising racial trauma is not about ranking suffering, but about accuracy. It is about understanding the specific ways that society imposes psychological costs on certain groups. Until we can see the difference between a personality trait and a survival response, our attempts at healing will remain superficial.

Key Takeaway

Survival strategies are often misread as character flaws when they are actually responses to structural threats.

04 — Aeon

The Panpsychist Argument

Does the demand for reason imply a conscious universe?

By Sam Kimpton-Nye · 18 min read

Editor's note: A deep dive into the philosophical bridge between the need for logic and the possibility of universal consciousness.

Humans possess an innate drive to make the world intelligible. When we see a pattern, we seek the cause. When we encounter a fact, we ask why it is so and not otherwise. This is the Principle of Sufficient Reason (PSR)—the idea that everything must have a reason for its existence. Once we understand the reason for a phenomenon, the mystery evaporates. One plus one equals two; it is not just a fact, it feels like a necessity. Philosophy seeks to bring the entire universe into this light, turning the arbitrary into the inevitable.

The Logic of Monads and Substance

This demand for intelligibility has led philosophers to radical conclusions. Leibniz proposed that reality is composed of 'monads'—tiny, self-contained universes that are the fundamental building blocks of existence. Spinoza, on the other hand, argued that there is only one infinite substance, which he called God or Nature. Both were trying to solve the same problem: how can a world that seems chaotic and arbitrary actually be a coherent, reasoned system? If everything has a reason, then everything must be connected to a single, underlying logic.

The demand for intelligibility can lead to the strange and fascinating world of panpsychism.

Panpsychism is the view that consciousness is not a biological accident that happened to humans, but a fundamental property of all matter. It suggests that electrons, protons, and even planets possess some degree of experience. While this sounds like science fiction, it is a serious philosophical response to the 'hard problem' of consciousness. If consciousness is a fundamental part of the universe, then we no longer have to explain how dead matter suddenly 'wakes up' to become sentient. It was always there, just in different degrees of complexity.

The Core Pillars of Panpsychist Thought

Consciousness is a fundamental property, not an emergent one
The universe is inherently intelligible through its mental properties
The divide between 'conscious' and 'non-conscious' is a matter of scale, not kind

If panpsychism is true, the universe is not a cold, mechanical clockwork. It is a medium of experience. This changes our relationship with the physical world entirely. We are not observers looking in from the outside; we are participants in a continuous field of being. The distinction between the 'self' and the 'object' begins to dissolve into a spectrum of awareness.

Whether or not one accepts this view, the investigation itself is valuable. It forces us to confront the limits of our own logic. If we insist that everything must have a reason, we must be prepared for the reasons to be stranger than we ever imagined.

Key Takeaway

If the universe is truly intelligible, consciousness may be its most basic ingredient.

05 — Aeon

The Fragile Majesty

The struggle to preserve the Buganda royal tombs

By Aeon · 8 min read

Editor's note: A reflection on how cultural identity is tied to physical structures that are inherently temporary.

On Kasubi Hill in Kampala, the resting places of four Buganda kings stand as a testament to a monarchy that remains central to Ugandan life. These are not the stone monuments of Europe; they are structures built from vegetal materials. They are living buildings, requiring constant, meticulous upkeep to survive. This choice of material creates a profound tension: the structures are deeply connected to the earth and the spirituality of the Baganda people, but they are also incredibly vulnerable to the elements and to catastrophe.

The Architecture of Impermanence

In 2010, a fire destroyed much of the tomb complex, a devastating blow to the cultural and spiritual heart of the kingdom. The reconstruction process was not merely a technical challenge but a spiritual one. It required the involvement of architects, artisans, and the kingdom's leadership to ensure that the essence of the tombs was preserved even as they were rebuilt. The fragility of the material is not a flaw; it is a feature of their relationship with the community. To keep the tombs alive, the people must remain active in their care.

The tombs embody the spirituality and values of the Baganda people through their very vulnerability.

This creates a unique form of cultural preservation. Unlike a museum piece that is kept behind glass to prevent decay, these tombs require a continuous cycle of renewal. The act of rebuilding is an act of devotion. It ensures that the traditions and skills required to maintain the structures are passed down through generations. The physical decay of the building is countered by the social cohesion of the people who maintain it.

Key Aspects of the Kasubi Tombs

Built from organic, vegetal materials
Central to the spiritual life of the Baganda people
Require constant communal maintenance to prevent decay

The film documenting this restoration captures the intersection of ancient tradition and modern preservation techniques. It shows how a community navigates the loss of its physical landmarks while reinforcing the intangible values those landmarks represent. The tombs are more than just graves; they are the physical manifestation of a continuous history.

In a world that often seeks to build for eternity using steel and concrete, the Buganda tombs offer a different model. They suggest that importance is not found in durability, but in the depth of the relationship between a structure and the people who hold it sacred.

Key Takeaway

Cultural permanence is found in the continuity of care, not the durability of stone.

06 — Lenny's Newsletter

The Agentic Shift

Why the next decade of tech will be defined by autonomy

By Lenny Rachitsky · 7 min read

Editor's note: A synthesis of the emerging trends in AI product design and business strategy.

We are witnessing a fundamental change in the nature of digital interaction. For the last twenty years, we have been building tools that wait for our commands. We have been the masters, and the software has been the servant. But the rise of AI agents like Meta's Muse signals the beginning of a new era: the era of the autonomous partner. These systems do not just follow instructions; they anticipate, they remember, and they act on our behalf. This is not a minor upgrade; it is a complete reordering of the human-computer relationship.

From Tools to Teammates

This shift requires a new set of skills and a new way of thinking about product development. As discussed by leaders at Google Labs, companies must move away from rigid, long-term planning and toward a lab-based model of rapid experimentation. The goal is no longer to build a perfect product, but to build a system that can learn and adapt. In this new landscape, the most valuable asset is not a stable codebase, but a high rate of learning.

The transition is from software as a tool to software as an agent.

This transition also brings new psychological and social challenges. As we begin to converse with machines, we will have to navigate new forms of social interaction. We will have to learn how to prompt, how to correct, and how to maintain boundaries with entities that feel increasingly human. The design of these interfaces—whether they are simple chat windows or complex generative UIs—will determine how successfully we integrate these agents into our lives.

The Three Pillars of the Agentic Era

Autonomy: Systems that act on behalf of the user
Continuity: Agents that maintain long-term context and memory
Adaptability: Interfaces that change based on the task and user

The implications for business are equally vast. The traditional SaaS model, based on seats and features, may give way to a model based on outcomes and agency. If an agent can perform a job, the value shifts from the software itself to the results it delivers. This will disrupt entire industries, from customer service to professional services, as the cost of intelligence continues to plummet.

We are standing at the edge of a massive shift. The tools of the past are becoming the companions of the future. How we design, build, and interact with these agents will define the next chapter of human history.

Key Takeaway

The value of technology is shifting from what it can do to what it can achieve for us.