The tech industry once had its roots in counterculture and bold thinking. But you wouldn't think that today.

In 2025 everyone is talking about AI and Agents. I believe we will look back at the word "agentic" as being distinctly 2025. But what does that mean? What type of software are we shipping? So far I see a lack of imagination in this space despite the incredible engineering going into AI tooling. 

Most of of the applications built with agent frameworks, like LangGraph, have been examples of incremental change. For instance, LinkedIn created an internal tool that translates natural language into SQL queries. While Elastic created some insight tools to enable security analysts to better respond to alerts generated by Elastic Security.

Given that LangGraph or LangChain are suddenly everywhere and marketed as a tool for developing "next generation interactive applications" I would expect more. Even though  these case studies provide some utility they all boil down to automating some tasks. I don't think this is the fault of the product or engineering teams but rather the fault of the industry's lack of imagination.

Today's tech industry is widely motivated by surveillance, optimization and automation. Most of the "game changing" value propositions of this technology is based on these three precepts. Very rarely are motivations behind AI applications aligned with amplifying human ability's to solve problems. Instead the focus, so far, has been on automating tasks, optimizing a workflow, or monetizing user behavior. I believe that this focus, more than anything, is why we haven't seen AI's killer app or really any killer app in quite some time.

My favorite ideas in tech were born from the era of 1960s and 1970s counterculture. Groups like Xerox Parc designed the first personal computer. The ideas that went into that machine required thinking beyond the current landscape and paradigms. Back then the vision for computing was all about empowerment, creativity and collaboration. Some ideas, such as overlapping windows and using a mouse to control and manipulate a GUI were revolutionary. While other ideas such as the chorded keyboard have long been (mostly) forgotten but were still interesting bets.

None of these bets would have been possible if Xerox Parc merely copied the best practices and opinions of that day. In the case of Xerox Parc those ideas were mostly ignored by their parent company and Xerox lost the future of computing to Apple and eventually Microsoft.

New mediums throughout history result in a reproduction of the status quo in the beginning. Film, for its first years, was theatre with cameras pointed at the actors before the medium evolved into tracking shots and creative edits. Likewise, the internet was initially a reproduction of existing print media before Web 2.0 and 3.0 made the medium ubiquitous. 

In 2025 we are probably at the same point with AI. That might even be cliche to admit. However, I am worried that the current status quo will never support bold new thoughts. Currently big tech controls 70-80% of the market. This wasn't the case with personal computing in 1975. At that time no company dominated the market. It was only the weirdos and the hippies who could conceive of the potential innovation that was on the horizon. 

Today big tech is very much aware of AI and is focused on dominating the direction of its development. Increasingly it's harder for a small player to enter the scene and innovate with a radically new idea. This is especially true when we all use the same tools and patterns used by the big tech companies.  

I believe there might be something else that is in plain sight that none of these companies are looking at. Sometimes that next big thing might even be an expansion or a re-evaluation of some old ideas. The history of computing has many forks in the road. Some of the paths I am interested in exploring in 2025 are: 

Innovation, when it happens, will happen in the least expected places. To get there I believe we need to start looking beyond the hype flowing through our social media feeds, take a few steps backwards and walk a different path.