BLAKE3 is a newly announced cryptographic hash function (9th of January, 2020) and a SHA-3 finalist designed by a team of cryptographers among which Zooko Wilcox from Zcash (the others being Jack O'Connor, Samuel Neves and Jean-Philippe Aumasson). BLAKE3 is significantly faster than MD5 and SHA-1/2/3 and additionally has some very useful properties as it is itself a Merkle tree on the inside. This makes it highly parallelizable and capable of verified streaming and incremental updates, something especially useful to distributed computing. And Zcash itself has always been more of a testing ground for the research of such cryptographic applications in the wild than a cryptocurrency geared towards immediate real life applications (something similar can actually be said about most legitimate, genuine crypto projects out there and there's absolutely nothing wrong with that; also, many of these things/applications cannot be tested/verified any other way, so in a way, investing in crypto is also bound to make you part of the ongoing research and experimentation - as the second law of thermodynamics goes, there's no free lunch and even luck is a function of calculated risk and intelligent exposure).
The Merkle tree is a data structure on the basis of which blocks in the sequence of the chain are structured. It's basically a tree of hashes, starting from a set of items at the bottom until reaching the top (header, root) and it's a way of allowing for the efficient and secure verification of large data sets.
The chart below shows BLAKE3's performance on modern server hardware (an Intel Cascade Lake-SP 8275CL processor):
BLAKE3 builds upon the established BLAKE2, and on the original Bao tree mode. The specifications and design rationale are available in the BLAKE3 paper and the current version of Bao implements verified streaming with BLAKE3. The official Rust implementation of the method is available at the repository (including a Rust crate and a simplified reference implementation) and there's also a version in Go available on Github.
BLAKE3 was at the top of Hackernews recently.
BLAKE3 can be seen two ways. On one hand, it appears to be the fastest, cheapest and most parallelizable secure hash function. On the other hand it's also a general-purpose Merkle tree. So its potential usefulness covers quite a bit of territory. And the emerging domain of the Internet-of-Things (which is something different from, if parallel to, the existing human-centric Internet) would require the fast and efficient computing of hashes on limited and low-end performance hardware.
Even though much more efficient, BLAKE3 wouldn't lead to less energy intensive PoW as difficulty would adjust as well.
The applications/implications of something like BLAKE3 to blockchain-based systems are as important as those of zero-knowledge proofs, but likely much more so. And BLAKE3 is no doubt going to be introduced to Ethereum and other chains in future upgrades.
The 2014 edition of the comprehensive description of BLAKE and BLAKE2.
This is an example of the value distributed ledger tech and collaborative peer-to-peer networks can bring/produce. And how value isn't necessarily about market price valuation - matter of fact, in the age of financial capital, value does not drive price and the two are actually fundamentally divorced in the process of abstraction (derivatives, etc.) But results cannot always be produced through simulations and in sterile, predictable laboratory environments (sometimes, if you want to make an omlette..) I think it's about time we abandon the notion that price, in the current arrangements of how markets "work", is an accurate or reliable signal for anything meaningful and recognize that the time where one would "trade" in between jerking off and make profits off that are over. From now on, there needs to be more active engagement, actual education and adequate understanding of wtf (therefore, the informational pollution and common deliberate misinformation for purposes of manipulation and mobilizing the bandwagon effects of public opinion need be countered and personal beliefs, wishes, biases and appeals to/blind trust in authority must eventually have gradually less of an effect/consequence).