Edit away your flaws with new DNA-editing CRISPR-technique, powered by AI
Have you dreamed about what it would’ve been like to be rid of all your bad habits and features? A disease that runs in the family, or perhaps your inability to digest milk without letting the whole room know about it.
Embryo selection
I recently published another note about embryo selection and gene editing and how the company Orchid currently is undergoing trials to allow you to select what specific embryo you want to turn into a real baby. This allows for reducing the risks of the child to have certain weaknesses, like allergies and likelihood of getting different types of cancer, blindness, etc.
Gene editing and AI
So far, however, the only intervention made is picking between naturally developed embryos and scanning them for “defects”. The next step toward the “brave new world” is to actually go into the DNA of the cell and copy paste the kinds of genes you want to promote in there. In what might be our last step, as humanity, is to let an AI, or LLM do it for us: “Our LLMs were trained on massive scale sequence and biological context to generate millions of diverse CRISPR-like proteins that do not occur in nature, thereby exponentially expanding virtually all known CRISPR families at-will.”
Should we let AI write the code in humans?
This seems like a bad idea on the face of it. It’s one thing to use AI to enhance us by assisting us with tasks we’re not built for. It’s a completely different thing to actively let AI decide what our children should look like, behave like, and what illnesses they should be susceptible to.
I’m not convinced it’s a good idea for us to go about editing genes to our liking. AI is not very good at the moment, despite being very impressive at summarising large pdf’s and articles. It will get better, however, and the better it’ll become, the less we will understand its capabilities.
Letting AI write some code for a small project today works just fine and is pretty harmless because any given programmer understand what it does. The better it become, the fewer people will understand, until that time it outperforms and outsmart even the brightest of humans.
The same will probably be true for editing our genetic code, not just computer code.