Prepare for a glimpse into the future, where technology takes a bold new turn! As a tech business founder, I approach technological advancements with a healthy dose of skepticism, especially when they seem more like clever marketing ploys. My approach has always been straightforward: identify the problem and find the most effective solution. While technology often plays a role, it's just one piece of the puzzle, alongside innovative business models and streamlined processes.
So, while I've kept an eye on AI developments, I've been cautious about jumping on the bandwagon. I've explored its immense potential and remained skeptical of doom-and-gloom predictions. However, everything I'd encountered until recently seemed overhyped or simply not feasible. That is, until the release of Claude Cowork on January 12th, 2026, which offered a glimpse into a new era of possibilities.
But let's rewind a bit. At the start of the year, I embarked on my first 'AI project,' curious about the hype surrounding platforms and apps built using Claude Code or OpenAI's Codex. Some claimed we'd no longer need programmers (a bold statement!), while others boasted about 'vibe coding' apps with AI. While these experiences were undoubtedly valuable, they didn't seem to revolutionize the game.
My challenge was managing an ever-growing sea of data: emails, slide decks, articles, reports, and notes. I needed a personalized knowledge base, a large language model (LLM) tailored to my context, to easily access information and connect new ideas.
I began working with Claude, crafting a detailed plan. Three steps in, I realized the suggested tools fell short, so I asked Claude to refine its recommendations. It apologized and updated my scope, and I started afresh.
One step involved installing my LLM on my laptop, which crashed due to its demanding hardware requirements. Claude acknowledged that while technically possible, it might not deliver optimal results and could cause further issues.
Then, in mid-January, Anthropic introduced Claude Cowork, a non-technical version of their programming productivity tool. Intrigued, I gave it a shot, tackling practical tasks. I chose to organize my expenses, a task I'd been putting off. I instructed Claude Cowork to sort my scanned receipts by month, rename them consistently, and generate an Excel report. It delivered an impressive, detailed report, categorizing expenses by month, all without explicit instructions.
I connected Claude Cowork to Chrome, Google Drive, and Gmail, and tasked it with planning my 'Grown-Up Gap Year' driving trip in France. It navigated the web for accommodation, sourced files, and suggested hidden gems, even creating a map of the proposed route.
I integrated Claude Cowork into my knowledge base project, connecting Readwise (for web articles, podcasts, and videos) and Granola (for meeting recordings) with my Google Drive reports and presentations. Now, I can ask Claude Cowork for meeting follow-ups, feedback on my performance, and suggestions for improvement.
This brings us to the future and OpenClaw. OpenClaw is an operating system or wrapper for agents, allowing users to build automated tasks on top of LLMs like Claude or OpenAI. It's a Claude competitor, but it lacks Claude's built-in guardrails and permission controls. However, with OpenAI's recent acquisition, expect improvements to Codex and ChatGPT soon. An intriguing aspect is Peter Steinberger's requirement that OpenClaw remain open-source, potentially attracting more development resources and becoming the standard for agent management.
This step-change is significant. With proper safeguards and reliable information sources, we'll move beyond apps to a world of automated workflows, skills, and tasks. We can envision a future where mundane tasks are automated, allowing us to focus on our unique human value.
Imagine a digital assistant offering impartial advice to prospective students, potentially disrupting the business model and eliminating agent commissions. How might we re-engineer the admissions process for transparency and effectiveness? What training do we need to manage digital coworkers?
I'm more confident than ever in the potential of emerging technology to provide useful answers to these questions and more. Some of these changes may not just automate existing processes but fundamentally alter how we operate.
This is the future I envision, and I'm excited to see how it unfolds.
