> While examining the DNA surrounding a reverse transcriptase gene in a giant virus, the agents spotted the same short sequence of DNA letters over and over again. The agents decided the find was worth a closer look and ultimately found similar patterns in the genomes of other viruses.
> The arrays of repeated DNA they found are somewhat like those seen in CRISPR systems.
It comes across to me that noticing this sort of thing is way more a matter of throwing raw compute at the problem than any sort of "intelligence" (or, heaven forbid, "creativity" or "insight") from a frontier model. Surely you could answer "does this sequence of letters contain any short sequences that repeat many times?" with simpler tools.
And since we also (per the article) apparently aren't any closer to knowing what the sequences mean, well.
zahlman•1h ago
> The arrays of repeated DNA they found are somewhat like those seen in CRISPR systems.
It comes across to me that noticing this sort of thing is way more a matter of throwing raw compute at the problem than any sort of "intelligence" (or, heaven forbid, "creativity" or "insight") from a frontier model. Surely you could answer "does this sequence of letters contain any short sequences that repeat many times?" with simpler tools.
And since we also (per the article) apparently aren't any closer to knowing what the sequences mean, well.