We should have a rule/button for rejecting links that talk about technology, but don't contain any technical details at all. For all that I know these guys could have just trained some logistic regression that is surprisingly useful in aligning the fragments. Or maybe their system is some transformer that is larger than GPT-3. Or maybe it computes directly on the DNA sequences using some custom exotic enzymes that contain convolutional layers. Who knows?
We should have a rule/button for rejecting links that talk about technology, but don't contain any technical details at all. For all that I know these guys could have just trained some logistic regression that is surprisingly useful in aligning the fragments. Or maybe their system is some transformer that is larger than GPT-3. Or maybe it computes directly on the DNA sequences using some custom exotic enzymes that contain convolutional layers. Who knows?
There is a link to the paper in Nature: https://doi.org/10.1038/s42256-025-01003-z
Now that the paper is posted, what are your thoughts on the research?