web search

Anthropic’s new AI search feature digs through the web for answers

Caution over citations and sources

Claude users should be warned that large language models (LLMs) like those that power Claude are notorious for sneaking in plausible-sounding confabulated sources. A recent survey of citation accuracy by LLM-based web search assistants showed a 60 percent error rate. That particular study did not include Anthropic’s new search feature because it took place before this current release.

When using web search, Claude provides citations for information it includes from online sources, ostensibly helping users verify facts. From our informal and unscientific testing, Claude’s search results appeared fairly accurate and detailed at a glance, but that is no guarantee of overall accuracy. Anthropic did not release any search accuracy benchmarks, so independent researchers will likely examine that over time.

A screenshot example of what Anthropic Claude's web search citations look like, captured March 21, 2025. — A screenshot example of what Anthropic Claude’s web search citations look like, captured March 21, 2025. Credit: Benj Edwards

Even if Claude search were, say, 99 percent accurate (a number we are making up as an illustration), the 1 percent chance it is wrong may come back to haunt you later if you trust it blindly. Before accepting any source of information delivered by Claude (or any AI assistant) for any meaningful purpose, vet it very carefully using multiple independent non-AI sources.

A partnership with Brave under the hood

Behind the scenes, it looks like Anthropic partnered with Brave Search to power the search feature, from a company, Brave Software, perhaps best known for its web browser app. Brave Search markets itself as a “private search engine,” which feels in line with how Anthropic likes to market itself as an ethical alternative to Big Tech products.

Simon Willison discovered the connection between Anthropic and Brave through Anthropic’s subprocessor list (a list of third-party services that Anthropic uses for data processing), which added Brave Search on March 19.

He further demonstrated the connection on his blog by asking Claude to search for pelican facts. He wrote, “It ran a search for ‘Interesting pelican facts’ and the ten results it showed as citations were an exact match for that search on Brave.” He also found evidence in Claude’s own outputs, which referenced “BraveSearchParams” properties.

The Brave engine under the hood has implications for individuals, organizations, or companies that might want to block Claude from accessing their sites since, presumably, Brave’s web crawler is doing the web indexing. Anthropic did not mention how sites or companies could opt out of the feature. We have reached out to Anthropic for clarification.

Archie, the Internet’s first search engine, is rescued and running

alan emtage, anonymous ftp, archie, Biz & IT, FTP, search engines, serial port, Tech, web search / Mike M. / May 16, 2024

Search for the Lost Searcher —

A journey through busted tapes, the Internet Old Farts Club, and SPARCstations.

Kevin Purdy – May 16, 2024 5: 44 pm UTC

Screenshot from The Serial Port's Archie project showing an Archie prompt with orange text on a black screen. — The Serial Code/YouTube

It’s amazing, and a little sad, to think that something created in 1989 that changed how people used and viewed the then-nascent Internet had nearly vanished by 2024.

Nearly, that is, because the dogged researchers and enthusiasts at The Serial Port channel on YouTube have found what is likely the last existing copy of Archie. Archie, first crafted by Alan Emtage while a student at McGill University in Montreal, Quebec, allowed for the searching of various “anonymous” FTP servers around what was then a very small web of universities, researchers, and government and military nodes. It was groundbreaking; it was the first echo of the “anything, anywhere” Internet to come. And when The Serial Port went looking, it very much did not exist.

The Serial Port’s journey from wondering where the last Archie server was to hosting its own.

While Archie would eventually be supplanted by Gopher, web portals, and search engines, it remains a useful way to index FTP sites and certainly should be preserved. The Serial Port did this, and the road to get there is remarkable and intriguing. You are best off watching the video of their rescue, along with its explanatory preamble. But I present here some notable bits of the tale, perhaps to tempt you into digging further.

The Serial Port notes the general loss of the Internet’s FTP era, including the recent shutdown of the Hobbes OS/2 Archive. Emtage, interviewed at length by the team, sent a tape copy of Archie to the Computer History Museum in Mountain View, California, but it was unrecoverable. Emtage’s company, Bunyip Information Systems, last sold version 3.5 of Archie’s server software for $6,000 in the mid-1990s (almost $12,000 today), and yet you can’t find it anywhere on the web. The Internet Archive wasn’t really running until 1996, just as Archie was fading from the web and, likely, memory.

The Serial Port team works dozens and dozens of resources to find a working copy of Archie’s code, including the Internet Old Farts Club on Facebook. I won’t give away the surprising source of their victory, but cheers (or na zdrowie) to the folks who keep old things running for everyone’s knowledge.

Not only did The Serial Code rescue the last working version of Archie (seemingly a 3.5 beta), but they posted its docs and now run an actual Archie server on an emulated Sun SPARCstation 5. It’s currently indexing its own mirror of the Hobbes archive, along with the FTP sites for FreeBSD, Adobe, and D Bit emulation. Searching for “word” in Archie found me a bunch of files, including the classic “Antiword” app and password managers and generators for OS/2.

Emtage, who would later help define the Uniform Resource Locator (URL) standard, gave his blessing to The Serial Port’s efforts to recapture and preserve the code of Archie’s server. It’s a happy ending to a story about archiving the early Internet in a way that’s relevant to today, with hopefully more to come.

Listing image by The Serial Port/YouTube

Archie, the Internet’s first search engine, is rescued and running Read More »