Indirect costs, or F&A, bundles infrastructure, compliance costs, payroll systems, and so on. In effect, the federal government subsidizes research universities. If the limit for indirect costs is limited to 15% or less, these universities will charge sponsored research awards for lab fees, network and computer infrastructure (including security, backups, reliable power), salary/benefit systems, and so forth. In effect, an unbundling. This change will have a positive side by itemizing the costs so that external bean-counting optimizers can select what they consider to be excessive, much as health insurance picks and chooses what it will pay for. Among the negative aspects, there is overhead of all this fine-grained accounting, a less predictable research environment, and likely less risk-taking in research proposals.
Probably more HN contributors are on the software or engineering side of things, so these moves may not seem concerning. However, I often see scientific results, including links to papers with government-sponsored research, in HN posts.