Why ai drug discovery isn't enough to reach longevity escape velocity

#1

We’re not getting in silico testing until we have enough data that AI can confidently derive correct molecular structures that address the issues want to address.
Even worse, we likely don’t even know what the causal issues are when it comes to controlling biological age.

AI is helping us scale drug discovery. This is great, but until we can scale data on what these molecules do in living organisms we’ve simply moved the bottleneck up to trials.

AI Powered Robotic Wet Labs

A project I consider essential for humanity to undertake is the construction of AI powered robotic wet labs at mass scale with access to a library of all molecules and model organisms.

We need to be performing lifespan, healthspan and performance enhancing (physical, mental, creative, etc.) studies on these individual and combined molecules in all model organisms and build the greatest modern database of biological knowledge.

This dataset will be recursively improved upon and give us far more information that what effect these molecules have individually.

A sufficiently powerful artificial intelligence can then:
• Begin to map patterns humans are likely to miss that will allow it to formulate new synthetic molecules that never before existed.
• Find out why certain molecules in combination work better than others, and this rule will give it the ability to automatically find the best combinations.
• It will figure out more accurate (and likely dynamic) allometric scaling between species and eventually this will help us scale anything to humans with near perfect accuracy.
• Ultimately this data will begin true in silico testing.

I wrote this today. It’s an idea I’ve been thinking over for a while.

If you’re interested in seeing the rest of the article check it out here: Travis Cox on X: "WHY AI DRUG DISCOVERY ISN'T ENOUGH TO REACH LONGEVITY ESCAPE VELOCITY We're not getting in silico testing until we have enough data that AI can confidently derive correct molecular structures that address the issues want to address. Even worse, we likely don't even know what the… / X

4 Likes
#2

This feels like an important safety threshold. Once AI can operate wet labs, it moves from manipulating information to taking actions that affect the physical world. That creates obvious risks from both mistakes and intentional misuse. The opportunity is huge, but autonomous capability should come with equally powerful guardrails, human oversight and hard limits on what systems are allowed to do.