> Cloud makes sense when elasticity matters; bare metal wins when baseload dominates.
This really is the crux of the matter in my opinion, at least for applications (databases and so on is in my opinion more nuanced). I've only worked at one place where using cloud functions made sense (keeping it somewhat vague here): data ingestion from stations that could be EXTREMELY bursty. Usually we got data from the stations at roughly midnight every day, nothing a regular server couldn't handle, but occasionally a station would come back online after weeks or new stations got connected etc which produced incredible load for a very short amount of time when we fetched, parsed and handled each packet. Instead of queuing things for ages we could instead just horizontally scale it out to handle the pressure.
We've switched over our libraries at $WORK to use `tsdown` and it's mostly been a very boring journey, we switched from `tsup` and the DX gains have been massive. Running our `dev` process in the frontend monorepo compiles and bundles all the libraries in less than a second on a cold start compared to `tsup` which was far slower. The biggest gain however was in our CI/CD pipeline where the build servers are much weaker than our developer machines, the `build` step in the quality gate for example went down by over a minute. We've also switched to the new native `tsgo` [0] for type checking, saving us another minute on CI/CD and have migrated a few things from ESLint to Oxlint, which was another easy minute saved. And we switched from Prettier to Biome, and checking the formatting on CI went from ~15s to ~1s. Massive gains are being had in the JS-world from gradual oxidation. Can't wait for Vite with rolldown, we tried that but have a few libraries that depend on SWC which made it a show stopper.
I used Org Mode for all mandatory assignments [0] when I was a student, it worked super well and I could create well formatted LaTeX documents/PDFs from it as well. The supervisors were very impressed that I did my assignment in LaTeX, and I could never be bothered to correct them :-)
I would imagine it wouldn't be advertised as a product developed solely for Darknet related markets but rather a marketplace for people who value their privacy and such. A few of the torrent sites I use have their tracker source code available on GitHub.
I have a friend who quite frequently used to buy from online markets (I have since lost touch with him). From what he told me what he and his friends used to do was to buy a new post box and set it up somewhere away from where they lived under a fictitious address and name. So they'd be John Doe living at Copper Street 45 (when it only actually goes to 44 or whatever).
Then he'd track how long the package took, whether or not it was delivered normally by the postal service and even then wait a week or so and get it in the middle of the night. Nowadays they apparently just use drones instead to deliver stuff.
This really is the crux of the matter in my opinion, at least for applications (databases and so on is in my opinion more nuanced). I've only worked at one place where using cloud functions made sense (keeping it somewhat vague here): data ingestion from stations that could be EXTREMELY bursty. Usually we got data from the stations at roughly midnight every day, nothing a regular server couldn't handle, but occasionally a station would come back online after weeks or new stations got connected etc which produced incredible load for a very short amount of time when we fetched, parsed and handled each packet. Instead of queuing things for ages we could instead just horizontally scale it out to handle the pressure.