AI 版 Google？Perplexity AI 每月查詢量達 7.8 億次，AI 瀏覽器快將推出
Perplexity AI logo is seen in this illustration taken January 4, 2024. REUTERS/Dado Ruvic/Illustration
黃仁勳也曾公開表示喜歡使用的 AI 搜尋引擎 Perplexity，CEO Aravind Srinivas 在 Bloomberg Tech Summit 上透露他們的服務在 2025 年 5 月達到 7.8 億次查詢，同時正以每月 20% 的增長速度快速發展。他預測若維持此增長速度，Perplexity 將在一年內達到每週 10 億次查詢。
Perplexity 的發展正正是以 Google 為模板，因為在 AI 搜尋引擎之後，他們正開發一款名為 Comet 的全新瀏覽器，旨在將 AI 從提供答案升級至執行完整操作。例如，用一條指令完成整個瀏覽流程。Srinivas 稱 Comet 不僅是一個瀏覽器，而是「認知操作系統（cognitive operating system）」，讓 AI 成為用戶生活中的個人化助手。
Srinivas 認為在瀏覽器上的用戶會是有無限的留存率，因為所有發生在搜尋列、分頁、網頁之上的瀏覽、互動，都會成為活躍用戶的一個額外查詢，同時能吸引那些厭倦了傳統瀏覽器（點名 Google Chrome 瀏覽器）的新用戶。
Comet will have a native virtual meets recording, transcription and searches over them. Won't be part of the first release, but very fast follow up. As for release date: it's going to take a min of three weeks and a max of five weeks. Reliability and latency have improved over…
— Aravind Srinivas (@AravSrinivas) May 13, 2025
Srinivas 在上個月曾經透露過 Comet 瀏覽器將會追蹤用戶活動以支持高效廣告投放，類似 Google 的盈利模式。首個版本的 Comet 預計將會在 6 至 7 月面世，會具備虛擬會議記錄、轉錄及智能搜索等功能。
更多內容：
Perplexity received 780 million queries last month, CEO says
ChatGPT 唯一對手？主打 AI 搜尋的 Perplexity 是什麼？
緊貼最新科技資訊、網購優惠，追隨 Yahoo Tech 各大社交平台！
🎉📱 Tech Facebook：https://www.facebook.com/yahootechhk
🎉📱 Tech Instagram：https://www.instagram.com/yahootechhk/
🎉📱 Tech WhatsApp 社群：https://chat.whatsapp.com/Dg3fiiyYf3yG2mgts4Mii8
🎉📱 Tech WhatsApp 頻道：https://whatsapp.com/channel/0029Va91dmR545urVCpQwq2D
🎉📱 Tech Telegram 頻道：https://t.me/yahootechhk
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
New York Post
30 minutes ago
- New York Post
Google's AI is ‘hallucinating,' spreading dangerous info — including a suggestion to add glue to pizza sauce
Google's AI Overviews, designed to give quick answers to search queries, reportedly spits out 'hallucinations' of bogus information and undercuts publishers by pulling users away from traditional links. The Big Tech giant — which landed in hot water last year after releasing a 'woke' AI tool that generated images of female Popes and black Vikings — has drawn criticism for providing false and sometimes dangerous advice in its summaries, according to The Times of London. 3 Google's latest artificial intelligence tool which is designed to give quick answers to search queries is facing criticism. Google CEO Sundar Pichai is pictured. AFP via Getty Images In one case, AI Overviews advised adding glue to pizza sauce to help cheese stick better, the outlet reported. In another, it described a fake phrase — 'You can't lick a badger twice' — as a legitimate idiom. The hallucinations, as computer scientists call them, are compounded by the AI tool diminishing the visibility of reputable sources. Instead of directing users straight to websites, it summarizes information from search results and presents its own AI-generated answer along with a few links. Laurence O'Toole, founder of the analytics firm Authoritas, studied the impact of the tool and found that click-through rates to publisher websites drop by 40%–60% when AI Overviews are shown. 'While these were generally for queries that people don't commonly do, it highlighted some specific areas that we needed to improve,' Liz Reid, Google's head of Search, told The Times in response to the glue-on-pizza incident. 3 Google AI Mode is an experimental mode utilizing artificial intelligence and large language models to process Google search queries. Gado via Getty Images The Post has sought comment from Google. AI Overviews was introduced last summer and powered by Google's Gemini language model, a system similar to OpenAI's ChatGPT. Despite public concerns, Google CEO Sundar Pichai has defended the tool in an interview with The Verge, stating that it helps users discover a broader range of information sources. 'Over the last year, it's clear to us that the breadth of area we are sending people to is increasing … we are definitely sending traffic to a wider range of sources and publishers,' he said. Google appears to downplay its own hallucination rate. When a journalist searched Google for information on how often its AI gets things wrong, the AI response claimed hallucination rates between 0.7% and 1.3%. 3 Google's AI Overviews, was introduced last summer and is powered by the Gemini language model, a system similar to ChatGPT. AP However, data from the AI monitoring platform Hugging Face indicated that the actual rate for the latest Gemini model is 1.8%. Google's AI models also seem to offer pre-programmed defenses of their own behavior. In response to whether AI 'steals' artwork, the tool said it 'doesn't steal art in the traditional sense.' When asked if people should be scared of AI, the tool walked through some common concerns before concluding that 'fear might be overblown.' Some experts worry that as generative AI systems become more complex, they're also becoming more prone to mistakes — and even their creators can't fully explain why. The concerns over hallucinations go beyond Google. OpenAI recently admitted that its newest models, known as o3 and o4-mini, hallucinate even more frequently than earlier versions. Internal testing showed o3 made up information in 33% of cases, while o4-mini did so 48% of the time, particularly when answering questions about real people.
Miami Herald
39 minutes ago
- Miami Herald
Tesla takes drastic measures to keep robotaxi plans secret
Tesla (TSLA) has big plans for Austin, Texas, this summer. The company is expected to roll out its long-anticipated robotaxi program in mere days. However, the company has also been fighting tooth and nail to keep those plans under wraps, according to a new report. Tesla has teased its robotaxi program since CEO Elon Musk first mentioned it in 2016. While the program has advanced at a snail's pace since then, it is days away from getting its first real-life test run. Related: Elon Musk says he is 'paranoid' about this issue; he's right to be The program will use Tesla's full self-driving technology to allow Teslas to operate autonomously as ride-hailing vehicles. According to Musk, Tesla plans to test only about 10 vehicles during this initial pilot run. Still, the ultimate plan is to have every Tesla on the road capable of serving as a robotaxi. Earlier this year, Tesla said that its FSD system has driven a cumulative total of 3.6 billion miles, nearly triple the 1.3 billion cumulative miles it reported a year ago. While Tesla has been open about some of the progress FSD and its robotaxi program have made, it has given scant details on many other issues and is currently fighting to keep one media organization from learning more about it. On Friday, June 6, Reuters published a story detailing its efforts to get public information about robotaxis from the city, specifically communications between Austin city officials and the company over its plan to launch autonomous vehicles this summer. Reuters requested access to communications from the last two years between the company and city officials in February, after Musk announced in January that robotaxis were coming to Austin. The city's public information officer told the news agency that "third parties" asked the city to withhold those records to protect their "privacy or property interests." The dispute went to the Texas Attorney General's office, as he adjudicates Freedom of Information Act (FOIA) request disputes. Related: Elon Musk, Tesla finally get some good news out of China Tesla's attorney wrote to the AG on April 16, objecting to releasing the "confidential, proprietary, competitively sensitive commercial, and/or trade secret information" contained in the emails. The attorney says the emails contain information that would reveal "Tesla's deployment procedure, process, status, and strategy" and "irreparably harm Tesla." Reuters sent a letter to the AG stating that Tesla's plan to deploy unproven technology on Texas streets "is an issue of enormous importance to Texas and the public at large," while underscoring the public's right to know about the program. While Tesla recently killed its Cybercab concept, at least for now, the company plans to test Model Ys already on the road as part of its robotaxi program. "It's prudent for us to start with a small number, confirm that things are going well, and then scale it up," Musk told CNBC's David Faber. Once it proves its concept in Austin, Tesla plans to expand the robotaxi program to Los Angeles and San Francisco soon after. California was Tesla's old stomping grounds before Musk moved the company's HQ to Austin in 2021 due to what he said were arduous regulatory practices, which may have to do with the company's operation during the Covid pandemic. With Tesla's plan to expand in the state, Musk will be heading back into that regulatory environment, except now the rules governing autonomous driving are much stricter. In April, the California Department of Motor Vehicles announced that it is seeking public comment on proposed regulations for self-driving vehicles. The agency told CNBC that it wants California to "offer the nation's most comprehensive rules for the operation of autonomous vehicles and underscore the DMV's commitment to enhancing public safety, fostering innovation, and establishing a robust framework for AV technology testing and deployment." Related: Elon Musk confirms Tesla's summer plans The Arena Media Brands, LLC THESTREET is a registered trademark of TheStreet, Inc.
CNET
42 minutes ago
- CNET
Google's AI Mode Now Creates Interactive Stock Charts For You
Google's AI Mode now can create interactive charts when users ask questions about stocks and mutual funds, the company said in a blog post Thursday. Users might ask the site to compare five years of stock performances for the biggest tech companies, or request to see mutual funds with the best rates of return over the past decade. Gemini, Google's AI engine, will then create an interactive graph and comprehensive explanation. I created a sample by going to the webpage for the new experiment (if you try it at work, you might learn that your admin has banned it, but it should work on a personal computer). Once there, I told it exactly what I wanted, "make me an interactive chart showing MSFT stock over the past five years." It produced the chart, and I was able to move the slider from one date to another, showing the stock price on that date. It's the same kind of chart you can probably get at your financial advisor's site, but it did work. Tell Google AI what chart you want, and it will create one that you can interact with. CNET But be warned: AI has accuracy issues, and users need to be extra-careful with financial information of any kind. "AI has historically struggled with quantitative reasoning tasks," said Sam Taube, lead investing writer at personal finance company NerdWallet. "It looks like Google's AI mode will often provide links to its data sources when answering financial queries. It's probably worth clicking those links to make sure that the raw data does, in fact, match the AI's output. If there's no link to a data source, proceed with caution; consider manually double-checking any numbers in the AI's output. I wouldn't trust any AI model to do its own math yet." The feature is a new experiment from Google Labs. At its I/O conference last month, Google announced AI Mode's ability to create interactive graphics for complex sets of data. The feature is now only for queries about stocks and mutual funds, but it will be expanded to other topics eventually. "I'd avoid asking AI any 'should I invest in XYZ' type questions," Taube told CNET. "The top AI models may be smart, but they aren't credentialed financial advisors and probably don't know enough about your personal financial situation to give good advice. What's more, AI doesn't have a great track record at picking investments, at least so far. Many AI-powered ETFs (funds that use AI to pick stocks to invest in) are underperforming the S&P 500 this year."
