AI JUST SOLVED ANOTHER SCIENTIFIC DOMAIN

The hypegeist missed it. But something big just emerged in AI-driven biopharma. A new domain was solved.

It's another example of digital intelligence expanding by filling the gaps it doesn't yet occupy. First chess in 1997. Then Go in 2016. Protein structure prediction fell with AlphaFold in 2020. Language generation has worked since 2022.

Each time, the domain space was considered too complex - or too vast - for AI to assist. Until digital intelligence evolved and finally cracked it.

And what the hypegeist missed was a paper landing in Nature Communications.

AI framework CAMPER designed a 12-amino-acid peptide to kill MRSA in living mice. It works against the bug's dense bacterial colonies. And the dormant bacteria that survive conventional antibiotics and seed new infections.

The sequence space for CAMPER's new peptide was larger than ten trillion possible arrangements - more than traditional screening can cope with. But CAMPER navigated it by combining machine learning with biophysical scoring - membrane interaction energy, structural stability and synthesisability. It then ranked and prioritised candidates.

The ultimate result is a molecule that reduces MRSA bacterial load by a factor of 300 in living mice.

Why's this worth paying attention to?

Most obviously because the industry is trying to prevent 39m projected deaths from antibiotic resistance. And CAMPER signals many thousands of other discoveries to come.

But it's also another tide mark in AI's progression more generally. Another domain succumbing to the evolution of digital intelligence.

Ten trillion possible sequences. One working molecule. Found by a machine. The gaps are filling ever-faster - and now include antimicrobial drug design.

What space gets filled next? There's plenty to go at.