OpenAI joining the seed round of Merge Labs is less about neurotech novelty than about where risk now sits.

As systems span signal, interface, AI and decisions, credibility is decided at the seams — where assumptions break and evidence rarely transfers cleanly.

In many real decisions, uncertainty isn’t a temporary lack of data but a permanent condition. Teams often act as if one more analysis will remove it. At some point the question shifts from reducing uncertainty to designing decisions that remain stable despite it. Most systems are not built for that.

Over the past few years, multiple studies in cognitive neuroscience have converged on the same result:

Beyond a threshold, adding information and options degrades decision quality — even for experts.

The hard part is not adding more, but deciding what to ignore, and this rarely feels comfortable