January 07, 2026
Over the past few years, AI has transitioned from what was once a cutting-edge technology reserved primarily for researchers and engineers into something that even our grandparents now use and discuss around the holiday dinner table. As with any foundational technology shift, adoption has happened faster than expected, and familiarity has set in quickly.
Despite this rapid normalization, we are still early in understanding how AI will ultimately reshape work, organizations, and entire industries.
The pace of innovation at the model layer, or intelligence layer, is unlike anything we have seen in technology history. But model capabilities have skated far ahead of our collective ability to harness them into a broad set of real-world applications that are deeply embedded into how we work and live.
In 2026, we are entering a phase where AI is no longer just a new tool, but a co-creator, collaborator, and partner embedded directly into our systems, workflows, and environments. This is a transition toward embedded intelligence; systems where AI agents are built directly into products, workflows, and physical environments, capable of executing actions on behalf of humans rather than merely responding to prompts.
Most AI products in use today still live inside standalone applications with primarily chat or Q&A-based interfaces such as ChatGPT. While these products have led to a paradigm shift, they represent an intermediate step rather than the end state. A deeper transition is now underway:
From AI assistants to autonomous collaborators embedded directly into how work gets done
From data infrastructure to governed, reliable AI systems that can be trusted in production
From single-model, narrow use cases to multi-modal embedded intelligence capable of tackling complex problems
I tend to agree with Jensen Huang’s framing that “AI won’t take your job, but someone using AI will.” And contrary to what some have suggested, venture capital is not immune to this transformation of work, especially given how competitive and information-dense our market is. I’m actively applying this mindset in my own work and in how we are evolving systems and processes at Crosslink. The opportunity is not to react to AI adoption, but to build, adapt, and upskill ahead of the curve. Organizations that do so early have a meaningful advantage. Read rest here.