@scottearleIn case (like me) you prefer text to video: AI companies scrape Web sites for training data. Scraping is so a...

@scottearleIn case (like me) you prefer text to video: AI companies scrape Web sites for training data. Scraping is so aggressive that it acts as a denial of service attack. Web sites naturally block #AI scrapers. AI companies bypass the blocks by routing their traffic through hundreds of thousands of domestic devices. These devices act as #ResidentialProxies — they use domestic IP addresses and make network requests on behalf of whoever rents capacity from the people who installed the proxies. Each #ResidentialProxy by itself imposes a manageable load but, put together, the fleet of proxies is as aggressive as the original AI #scraper and much harder to block.Customers don't, in general, understand that they're signing up for this: it's mentioned in vague jargon on page 73 of terms of service that nobody reads. Customers think they're getting a TV app, a "dodgy stick", or some other goody. The residential proxy sneaks in like a Trojan horse.Residential proxies can also be used for oth...

Read Original

Related