Lawmakers questioned #Google#'s CEO #SundarPichai# for over three hours, looking for answers on alleged anti-conservative bias, plans for a censored #searchengine# for #China#, and #data# collection #dataprotection# #dataprivacy# #Censorship# #accountability#

AI search explained ranking on chatgpt vs claude vs gemini Claude = Brave Search Gemini / AI Overviews = Google ChatGPT = a hot mess of bing + its own index + google + other feeds Brave is on Anthropic's subprocessor list, and there's a BraveSearchParams field inside Claude. Studies found about 79–87% of the URLs Claude cites come from Brave's top 10 organic results, and Claude barely re-ranks them. If you rank on Brave, you get cited with that said Claude only searches on about 37% of prompts. ChatGPT searches on about 90%. Claude looks things up when a query needs fresh info, rankings, a location or a comparison. Otherwise it answers from training data gemini page one on google, your in gemini. pretty simple chatgpt, high on the pretty crazy scale It started on Bing, and some of that is still there. But OpenAI now has its own index, built by OAI-SearchBot, plus third-party scrapers and partners like Yelp and TripAdvisor. How closely it matches any one search engine's top results is low and keeps changing. When you ask Claude and ChatGPT the same prompt, only 8–20% of their citations overlap So being dominant in one engine doesn't mean you show up in the others. What to do about it: Track Brave rankings separately. If you want Claude citations, this is your most direct lever. Submit your URLs to Brave directly so they get picked up faster. For ChatGPT Don't skip Bing. It's cheap insurance for ChatGPT: set up Bing Webmaster Tools, turn on IndexNow and allow OAI-SearchBot in robots.txt. also go past rankings. Get brand mentions, write structured and authoritative pages, and get lots of other sources saying the same things about you. Write for the prompts that trigger search. Claude mostly searches on comparisons, "best X," recent news and local queries. Build pages for those. Remember when people optimized for Google and Bing separately? That's basically back

Better search may come less from smarter indexes than from giving agents a richer way to touch text. Shows that AI agents using basic terminal tools like grep, file reads, and shell commands to search raw data perform far better than conventional retrieval systems on multiple benchmarks. On BrowseComp-Plus, swapping semantic retrieval for terminal search raised accuracy from 69% to 80% while lowering cost. The deeper point is not that grep is magically smarter than embeddings. It is that retrieval is usually treated as a model problem, when it is also an interface problem. A conventional retriever turns the corpus into a narrow ritual: ask once, receive a ranked list, reason over whatever survived. That works when the question is close to a document’s semantic center, but it breaks when the answer depends on exact phrases, faint clues, document structure, or a chain of small discoveries. Direct Corpus Interaction changes the shape of the task. The agent can search an exact string, inspect nearby context, notice a new entity, constrain the search again, and keep testing its hypothesis against the raw files. Here’s the part most people miss: the gain was not mainly from finding more gold documents, but from extracting more usable evidence once a promising document was reached. That makes DCI less like a better search engine and more like giving the model fingers. The limitation is real: as the corpus grows, the cost of finding the first useful anchor rises quickly, and blunt terminal search will not replace indexes for every large, static collection. But the paper’s lesson still lands cleanly. For capable agents, the bottleneck may no longer be only what they know, or even how they reason, but how much of the world their tools allow them to touch. ---- Paper Link – arxiv. org/abs/2605.05242 Paper Title: "Beyond Semantic Similarity: Rethinking Retrieval for Agentic Search via Direct Corpus Interaction"

A search engine has many moving parts, and they constantly change with scale of index and traffic. In this blog post, we dive into the code-level abstractions we've built to develop search infrastructure that is both performant and a joy to work with.

GM to the Unstoppable Search Engine. Tether P2P team is developing a peer-to-peer search engine that can scale across thousands of nodes and self-organize data in fault-tolerant structures to maintain high availability and censorship resistance. Work progress is impressive. Here you can see the full Wikipedia archive stored and searchable across 100 distributed nodes in our test environment. Performance is insane.