The next singularity may come from AI and biotech

I've been noticing more AI health products showing up lately, and honestly, it makes sense. Health feels like one of the clearest fits for AI, especially for keeping track of your own health and making sense of the questions that come up between appointments. The approaches are already starting to look different. ChatGPT Health can bring records into a conversation, and I've used Theta Wellness to keep my own health information together. Samsung Health uses AI to make wearable data easier to read, while Google Health Coach uses Gemini to offer personalized wellness guidance. They are not replacements for medical care, but they show how AI is moving closer to the health data people already collect.

The part I find most promising is probability. When someone first looks at a health problem, the most likely explanation naturally gets the most attention. The risk is that an unusual case can initially look like a very ordinary one. In theory, AI can work across a huge amount of medical knowledge and keep more possible connections in view. It might notice that one detail doesn't quite fit, even if all that does is give the patient or doctor another thing to check. But this only helps if it knows what it still needs to ask. If I say I've been unusually tired, does it ask when that started and what else changed around then? If a wearable shows a sudden change, does it ask about travel, a new routine, or whether I even wore the device correctly before treating the number as meaningful? A confident answer based on missing context isn't much of an improvement.

This is where I start thinking beyond health apps. For one person, the job is connecting records over time and noticing what does not fit. In biotech, the data is vastly more complicated, but the useful part may be similar: finding relationships across biological data and helping researchers decide which questions or hypotheses are worth testing. That would not prove a treatment works or replace experiments. It could change where research starts and how quickly weak ideas are ruled out. The part that starts to feel closer to a singularity is the possible feedback loop: models help decide what to test, experiments produce new data, and that data shapes the next round of models and experiments. If that cycle keeps getting faster, the result could be more than another generation of health apps.

For now, I still see AI as a tool. It can help organize what happened and give a doctor a clearer starting point. An experienced doctor still has to examine the patient and make the diagnosis. Researchers still have to test what the models suggest. I do wonder how far this goes, though. Today it helps us make sense of records. The much bigger leap could come when AI starts changing which biological questions researchers know to ask.