Telecom’s automated abuse is genuinely old, older than the web crawler. Automatic dialers were harassing households by the 1970s and 80s, and the Telephone Consumer Protection Act tried to rein them in in 1991, before the first search engines had indexed much of anything. The robocall did not wait for the internet.

What VOIP changed later was not the *existence* of the robocall but its economics. It made originating a call, and stamping it with any number you liked, nearly free. When a call costs a fraction of a cent, a one-in-a-thousand hit rate is still profitable, so cheap origination in the 2000s turned a nuisance into a flood. (This is worth stating precisely: large-scale web scraping predates cheap, spoofable VOIP by roughly a decade. Telecom’s primacy rests on the analog autodialer, not on VOIP.)

What makes telecom instructive is that it was forced to respond at the protocol layer. Caller ID was trivially spoofable, so the industry built STIR/SHAKEN: a way to cryptographically attest that a call really originated where it claims. That is human verification by another name: not "is this person real?" but "is this signal what it says it is?"

• The attack surface was the identifier itself, the phone number, not the content.
• The fix was attestation at origin, not filtering at the edge.
• The residue is "Scam Likely" labels: a probabilistic warning, not a hard block.

Every other corner of TMT is now re-learning this lesson. Media has spoofed page views; tech has spoofed sessions and accounts. Telecom just got there first because its unit, a call to a universal, addressable number, was the cheapest thing to automate and the most directly convertible to money, on a network that authenticates nothing about a call’s origin. For the full ordering, see The Order of Collapse.