Not from a medical background, but I spent some time writing hospital infection surveillance software so can share a few anecdotes. What he's saying isn't inac

Not from a medical background, but I spent some time writing hospital infection surveillance software so can share a few anecdotes. What he's saying isn't inaccurate. The idea is antibiotic resistance generally carries some metabolic cost -- the bacteria disable or find some other way around whatever pathway is being targeted by the antibiotic. This means they operate "less efficiently" than their non-resistant peers. When the antibiotic disappears from the environment, the resistance no longer offers an advantage, so they get out-competed by their non-resistant "more-efficient" peers. This is in fact one of the ways hospitals fight outbreaks of resistant bacteria -- they cycle antibiotics. More subtleties of course, but basically the head infection surveillance person says "alright docs, this month we're blacklisting these antibiotics, next month we're blacklisting these." (This is one of the interventions our infection software helped with.) On the other hand, there are places where the widespread use of antibiotics extends far beyond the reaches of a hospital building. Commercial agriculture is infamous for overusing antibiotics (basically mixing them into the lot's feedstock so every cow gets it). Outbreaks like SARS and avian flu weren't antibiotic-resistance outbreaks, but they WERE caused by sizable quantities of people working in close proximity with livestock and then spreading the disease through their normal travel and social interaction vectors. India is known for having antibiotics widely available over-the-counter for people. There are stories of people coming back from India with an infection resistant to literally everything (there's only about 50 antibiotics out there) -- when that happens, it's like the outbreak movies where the CDC comes in with plastic sheets and wearing bunny suits to quarantine the room. So will the lower-metabolic-efficiency attribute of antibiotic resistance be self-limiting enough to prevent it from turning into another 1918 spanish flu global pandemic? Maybe. But as long as there are large city-scale pockets of heavy antibiotic over-use with people traveling in-and-out, there are going to be constant transmission vectors that will still kill people in largely preventable ways.