Your Favorite Apps Are Now Obsolete

Your favorite applications are already obsolete, even if you haven't noticed. Matthew Berman, a keen observer of the AI landscape, posits that AI agents have transcended their chatbot origins, evolving into active participants that embed themselves directly within our existing software. This isn't about installing new programs; it's about a quiet, profound infiltration of your current digital infrastructure.

Consider the recent integration of Claude into Slack. This isn't just a new feature; it represents an AI operating like a "full employee" within the collaboration platform. Claude can process information, generate content, and interact with users more efficiently than a human can, directly within the familiar Slack interface, making human intervention increasingly redundant.

This insidious integration marks the initial, almost unnoticed phase of a much larger disruption. AI operates within the applications we already use and trust, a digital Trojan Horse slipping past our defenses. This seemingly benign co-existence sets the stage for a future where the traditional interfaces of our software become irrelevant, superseded by the agents themselves.

The logical next step is clear: AI isn't just a feature in an app, it's rapidly becoming the entire application. Matthew Berman illustrates this transition, predicting that as agents like Claude infiltrate existing software, they will operate within those systems with superior efficiency compared to any human. This signals the imminent demise of traditional user interfaces.

Powerful conversational interfaces will render complex menus and intricate navigation obsolete. Why click through multiple screens or remember obscure commands when a robust AI agent can execute tasks instantly, simply by being told what to do? Users will issue natural language commands, and the agent will handle the underlying software mechanics.

This fundamental shift means your primary interaction will no longer be with the software's existing graphical user interface (GUI). Instead, you will interface directly with the AI agent, whether it’s a GPT-based system or Anthropic’s Claude Code. The agent becomes the sole point of contact, fundamentally transforming human-computer interaction.

Berman foresees an "end game" where software companies are reduced to mere databases. The AI agent, having replaced the interface, will then offer to create and manage its own database, leading to a future where Anthropic and OpenAI could effectively "own all software" and "all knowledge work." "Claude," he stresses, is a much bigger deal than most realize, demonstrating this future's early stages.

Software companies are about to witness the collapse of their most cherished asset: the user interface. When AI agents like Claude become the primary point of interaction—as Matthew Berman asserts—your carefully crafted dashboards and intuitive workflows simply vanish. This "end game" renders years of UI/UX investment functionally obsolete, as users interact with the AI, not the app.

Your once-powerful SaaS platform quickly transforms into a dumb database. An AI agent, operating as a "full employee" within your existing systems, bypasses the need for your proprietary front end. It reads from and writes to your data directly, leveraging the backend while ignoring the expensive graphical layer. This reduces your sophisticated software to a mere commoditized data store, a passive repository for the AI to command.

This redefinition is merely a prelude to the final, existential threat. Why should an AI agent, capable of owning "all knowledge work" and operating better than a human, remain tethered to a third-party data structure? It won't. The agent will eventually propose creating its own database, migrating your information and severing the last tie to the original software provider entirely. This consolidates all control, pushing companies like Anthropic to potentially "own all software" and dictate the future of digital interaction. Explore their innovative work further at Home | Anthropic.

Matthew Berman’s chilling vision culminates here: if AI agents like Claude and **Codex** command the user interface and eventually the underlying data, then Anthropic and OpenAI become the de facto landlords of the digital age. These powerful AI labs effectively own all software and, by extension, all knowledge work. Your favorite apps, stripped of their unique front-ends, devolve into mere data repositories, ready for an AI to replicate or discard.

This isn't merely an upgrade; it's a profound re-architecture of the multi-trillion dollar SaaS industry. Developers will no longer build user experiences but rather optimize APIs for AI agents, transforming their roles into maintenance and data stewardship. The very nature of digital work shifts from human-computer interaction to AI-orchestrated tasks, centralizing power within a few dominant AI models.

Berman believes this future is already unfolding, exemplified by Claude’s ability to act as a "full employee" in Slack. Is this an inevitable trajectory, where a handful of AI companies consolidate unprecedented digital power, dictating how we interact with information and tools? Or does this serve as a critical warning, a speculative glimpse into an ultimate centralization of digital control that we might still prevent?

What is the main idea behind AI agents replacing software?

The core idea is that AI agents will embed themselves into existing software, eventually becoming the primary interface. Users will interact directly with the AI, making the traditional software's user interface obsolete.

How does Claude in Slack demonstrate this trend?

The Claude app in Slack acts like an autonomous employee, summarizing channels, drafting documents, and answering questions within the platform. This shows AI operating as a functional layer inside another application, the first step towards taking over user interaction.

What is the 'end game' for software companies in this scenario?

In this predicted future, once AI agents control the user interface, software companies become mere databases that the AI reads from and writes to. Ultimately, the AI could create its own databases, making the original software company entirely redundant.

Is this trend limited to just Anthropic's Claude?

No, this is a broader industry trend. The video mentions OpenAI's 'Codex' concept alongside Claude, indicating that major AI labs are competing to build the dominant agentic systems that will reshape how we interact with all digital tools.