Compute power, Pythagoras, and run-length encoding
Last week Wayne wrote about DeepSeek and computing power. He gives examples that show that it is plausible that a breakthrough can lead to using much less computational power than previously assumed necessary.
Wayne’s latest post pairs well with this first post on the blog, The Million Dollar Matrix Multiply.
This morning I wrote a post about a family of Pythagorean triples that have a simple representation in binary. The binary representation of the triples segues into a discussion of sparsity, one’s complement, and run-length encoding.
Thanks for reading.
If you’d like to hear from me more often, you could follow one of my X accounts. The most popular accounts are @CompSciFact, @DataSciFact, and @AlgebraFact.