Pretty fascinating development for a few reasons: 1) COBOL is what a gigantic portion of legacy federal govt systems (including IRS, DoD, etc) run on 2) The average COBOL programmer/expert is ~58 years old (<5 years from retirement) 3) Almost no one is being taught COBOL or learning it. 4) Previous efforts to modernize COBOL systems were all deemed far too expensive (not enough expert engs to actually perform the upgrades without breaking everything) It’s probably a net-good if Claude can do this? We’re just a couple years out from no long having the humans that would be able to manage these systems