Boneh–Lynn–Shacham signatures (BLS signatures) are digital signatures that provide a different set of tradeoffs compared to the ECDSA and schnorr signatures currently used in Bitcoin.
Probably the most interesting property of BLS signatures is that they allow non-interactive signature aggregation. For example, if Alice and Bob both independently sign the same transaction, a third party can combine their signatures into a single signature that proves both of them signed. By comparison, using a Bitcoin’s existing schnorr signatures, a single signature proving both of them signed can only be produce through an interactive protocol like MuSig2 where at least one of them receives the other’s partial signature before producing their own partial signature and producing the complete signature.
In theory, if other changes to Bitcoin were made and if support for BLS signatures was added, miners could aggregate all signatures together before producing a block, allowing blocks to contain only a single signature, which would moderately improve onchain capacity and might speed block verification when cached verification was unavailable (e.g. during initial block download).
BLS signatures are not directly compatible with the elliptic curve used by Bitcoin and are not as well studied as schnorr signatures. It would not be possible to use signature adaptors and technology based on them (such as PTLCs) with BLS signatures.
Optech newsletter and website mentions