Since joining Worksome, I have been engaged in intense experimentation with (agentic) AI. My usage pattern of models in Cursor shows it: from early curiosity to “all-in” when the models have delivered consistently (hello Anthropic’s Claude 4.5 Opus ❤️🔥).
This era of software engineering is the perfect moment to reconnect with our skills. The human connections become more important than ever (“s̶o̶f̶t̶ ̶s̶k̶i̶l̶l̶s̶ ➝ 𝐜𝐨𝐫𝐞 𝐬𝐤𝐢𝐥𝐥𝐬), because work can be more “outsourced” to agentic colleagues.
Which means that some of the skills I have been preaching to my product-building teams become more relevant than ever: capture context in a written, accessible form, and explain what you want to achieve as much as what you want to prevent. This has been useful for communication in software development teams (product/engineering/design/QA…), and it becomes top-priority again as we incorporate team members that can’t really “read the room” (our agentic colleagues).
For software developers, it’s especially relevant to master clear written communication, so our prompts achieve the expected result (and prevent issues we should anticipate).
As my experimentation progressed, I’ve seen how I have incorporated new use cases for AI - thanks in big part through MCP connections, to “connect the dots” between tools, and use Cursor as a “way to work”, instead of just being “a way to write code”.
I’m not sure exactly where AI will take us, but for now, I’m definitely having fun experimenting with it and 𝐬𝐡𝐢𝐩𝐩𝐢𝐧𝐠 𝐯𝐚𝐥𝐮𝐞 to our clients. So far, it has helped me reduce time spent on menial tasks and spend more time on connection-building, strategic thinking, and human connection.
This points to an exciting 2026! #cursor #agenticAI