Mastodon Feed
Boosted by soatok@furry.engineer ("Soatok Dreamseeker 🔜 Megaplex"):
joepie91@slightly.tech ("Sven Slootweg, ("still kinky and horny anyway")") wrote:
jon schlinkert wrote a relatively widely-used (but small) JS library that claimed to 'the fastest' as shown in benchmarks. it took years before someone noticed that it was cheating those same benchmarks
google puts out libraries like a JS protobuf implementation that claim to be "more performant" because they do code generation, except when you actually test them they perform worse than the non-codegen options on every real-world case
deno's *entire list of marketing points* upon its release, the full 100.00% of the list, was somewhere between a lie and a misdirection, and nobody ever wrote up how all of these are false; people widely believe those claims to be true to this day
cloudflare made repeated claims of "the biggest ddos attack ever", even though it took less than an hour of googling to find bigger historical attacks. nobody called them on it
the majority of "highly performant" libraries I've evaluated, across languages, come without any benchmarks or evidence, and nobody has ever requested it, despite repeating the claims. for most of these, the author cannot explain *why* it is faster. it often turns out to be because it outright ignores edge cases
and let's not forget amazon AWS, which for years marketed its services on the premises that it's "cheaper than traditional hosting". all the while charging between $50-$100 per TB of traffic while $1-$5 was standard in the industry, and I have literally seen companies go bankrupt over their AWS bill