Amazon Files Lawsuit Against AI Startup Over Automated Shopping Features in Browser

On Tuesday, Amazon filed a lawsuit against a well-known artificial intelligence startup over a feature in its browser that enables users to automate purchases. Amazon alleged that Perplexity AI had illicitly accessed customer accounts and disguised the AI’s actions as human browsing.

“The misconduct by Perplexity must cease,” Amazon’s legal representatives stated. “Perplexity has no permission to act where it is forbidden. The intrusion involves a code rather than a lockpick, rendering it equally illegal.”

Perplexity, which has experienced significant growth in light of the AI assistant boom, previously accused Amazon of leveraging its dominant market position to suppress competition and dismissed Amazon’s allegations.


“Bullying occurs when larger companies employ legal threats and intimidation to stifle innovation and negatively impact people’s lives,” the company expressed in a blog post.

This dispute underscores new conversations regarding the regulation of the increasing use of AI agents, AI-powered autonomous digital assistants, and their interactions with websites.

In its legal action, Amazon accused Perplexity of secretly accessing Amazon’s private customer accounts via the Comet browser and associated AI agents, misrepresenting automated actions as human browsing. Amazon asserted that Perplexity’s systems endangered customer data and ignored repeated calls to shut them down.

“Instead of being transparent, Perplexity deliberately configures its CometAI software to mask Comet AI agent activity on Amazon’s platforms,” the company stated.

Amazon’s complaint also claimed that Perplexity’s Comet AI agent undermined the shopping experience for customers and hindered Amazon’s ability to guarantee that users benefiting from the agent receive the personalized shopping experience it has developed over decades.

In a previous statement, Amazon indicated that third-party applications making purchases on behalf of users should operate transparently and respect companies’ preferences for participation.

Perplexity had earlier revealed that it received legal threats from Amazon aimed at preventing Comet AI agents from shopping on its platform, asserting that this action poses a wider threat to user choice and the future of AI assistants.

Skip past newsletter promotions

Perplexity is among several AI startups that are restructuring web browsers to incorporate artificial intelligence, aiming to enhance user autonomy and simplify everyday online tasks, from composing emails to completing purchases.

Amazon is also developing similar functionalities, including Buy For Me, which enables users to shop across various brands within the app, and Rufus, an AI assistant that recommends products and manages shopping carts.

The Comet browser’s AI agent from Perplexity acts as a purchasing and comparison assistant for users. The company contends that user credentials are stored locally and not on its servers, asserting that users have the right to select their own AI assistant and framing Amazon’s actions as an attempt to safeguard its business model.

“Simplified shopping leads to more transactions and greater customer satisfaction,” Perplexity remarked. “However, Amazon is less focused on that and more on serving ads.”

Source: www.theguardian.com

Perfect Timing for Firefox: Developing an AI Browser and the Future of the Web

Need an assistant for your online activities? Several major artificial intelligence companies have moved away from chatbots like ChatGPT and are now focusing on new browsers with deep AI integration. These could take the form of agents who shop for you or ubiquitous chatbots that follow you, summarizing what you’re looking at, looking up related information, and answering related questions.

In the last week alone, OpenAI released the ChatGPT Atlas browser, while Microsoft showcased Edge’s new Copilot mode, both heavily utilizing chatbots. In early October, Perplexity made its Comet browser available for free. Mid-September saw Google rolling out Chrome with Gemini, integrating its AI assistant into the world’s most popular browser.

Following these releases, I spoke with Firefox General Manager Anthony Enzor-DeMeo to discuss whether AI-first browsers will gain traction, if Firefox will evolve to be fully AI-driven, and how user privacy expectations may change in this new era of personalized, agent-driven browsing.

Guardian: Have you tried ChatGPT Atlas or other AI browsers? I’m curious what you think about them.

Anthony Enzor-DeMeo: Yes, I’ve tried Atlas, Comet, and other competing products. What do I think about them? It’s a fascinating question: What do users want to see? Today, users typically go to Google, perform a search, and view various results. Atlas seems to be transitioning towards providing direct answers.

Guardian: Would you want that as a user?

Enzor-DeMeo: I prefer knowing where the AI derives its answers. References are important, and Perplexity’s Comet provides them. I believe that’s a positive development for the internet.

Guardian: How do you envision the future of the web? Is search evolving into a chat interface instead of relying solely on links?

Enzor-DeMeo: I’m concerned that access to content on the web may become more expensive. The internet has traditionally been free, mostly supported by advertising, though some sites do have subscriptions. I’m particularly interested in how access to content might shrink behind paywalls while aiming for a free and open internet. AI may not be immediately profitable, yet we have to guard against a shift towards a more closed internet.

Guardian: Do you anticipate Firefox releasing an AI-integrated or agent-like browser similar to Perplexity Comet or Atlas?

Enzor-DeMeo: Our focus remains on being the best browser available. With 200 million users, we need to encourage people to choose us over default options. We closely monitor user preferences regarding AI features, which are gradually introduced. Importantly, users retain control; they can disable features they do not wish to use.

Guardian: Do you think AI browsers will become popular or remain niche tools?

Enzor-DeMeo: Currently, paid AI usage is about 3% globally, so it’s premature to deem it fully mainstream. However, I believe AI is here to stay. The forthcoming years will likely see greater distribution and trial and error as we discover effective revenue models that users are willing to pay for. This varies widely by country and region, so the next phase of the internet presents uncertainties.

Guardian: What AI partnerships is Firefox considering?

Enzor-DeMeo: We recently launched Perplexity, akin to a search partnership agreement. While Google search is our default, users have access to 50 other search engines, providing them with options.

Guardian: Given your valuable partnership with Google, what financial significance does the Perplexity partnership hold?

Enzor-DeMeo: I’m unable to share specific details.

Guardian: Firefox has established its reputation on user privacy. How do you reconcile increasing demands for personalization, which requires more data, with AI-assisted browsing?

Enzor-DeMeo: Browsers inherently have a lot of user context. Companies are developing AI browsers to leverage this data for enhanced personalization and targeted ads. Mozilla will continue to honor users’ choices. If you prefer not to store data, that’s entirely valid. Users aren’t required to log in and can enjoy completely private browsing. If it results in less personalized AI, that’s acceptable. Ultimately, the choice lies with users.

Guardian: Do you think users anticipate sacrificing privacy for personalization?

Enzor-DeMeo: We’ve observed a generational divide. Younger cohorts prioritize value exchange—will sharing more information lead to a more tailored experience? In a landscape with numerous apps and social media, this expectation has emerged. However, perspectives vary between generations; Millennials often value choice, while Gen Xers prioritize privacy. Many Gen Z users emphasize personalization and choice.

Guardian: What are your thoughts on the recent court decision regarding Google’s monopoly?

Enzor-DeMeo: The judge acknowledged the influx of competition entering the market. He deliberately avoided delving into the browser engine domain. We support search competition but not at the cost of independent browsers. The ruling allows us to keep receiving compensation while monitoring market evolution over the next few years. The intersection of search and AI remains uncertain, and a prudent stance is to observe how these developments unfold.

Guardian: Firefox’s market share has been steadily declining over the past decade; what are your realistic goals for user growth in the coming years?

Enzor-DeMeo: Every user must decide to download and use Firefox. We’re proud to serve 200 million users. I believe that AI presents us with significant growth opportunities. We want to provide choices rather than lock users into a single solution, fostering diverse growth possibilities for us.

Source: www.theguardian.com

ChatGPT Atlas: OpenAI Introduces Chatbot-Focused Web Browser | Tech News

On Tuesday, OpenAI unveiled an AI-driven web browser centered around its renowned chatbot.

“Introducing the revolutionary browser ChatGPT Atlas” Tweet from the company stated.

This browser aims to enhance the web experience with a ChatGPT sidebar, enabling users to ask questions and engage with various features of each site they explore, as demonstrated in a video shared with the announcement. Atlas is currently accessible worldwide on Apple’s macOS and will soon be released for Windows, iOS, and Android, according to OpenAI’s announcement.

With the ChatGPT sidebar, users can request “content summaries, product comparisons, or data analysis from any website.” Website for more details. The company has also begun presenting a preview of its virtual assistant, dubbed “Agent Mode,” to select premium users. Agent Mode allows users to instruct ChatGPT to execute a task “from start to finish,” such as “travel research and shopping.”

While browsing, users can also edit and modify highlighted text within ChatGPT. An example on the site features an email with highlighted text along with a recommendation prompt: “Please make this sound more professional.”

OpenAI emphasizes that users maintain complete control over their privacy settings: “You decide what is remembered about you, how your data is utilized, and the privacy settings that govern your browsing.” Currently, Atlas users are automatically opted out of having their browsing data employed to train ChatGPT models. Additionally, similar to other browsers, users can erase their browsing history. However, while the Atlas browser may not store an exact duplicate of searched content, ChatGPT will “retain facts and insights from your browsing” if users opt into “browser memory.” It remains unclear how the company will handle browsing information with third parties.

Skip past newsletter promotions

OpenAI is not the first to introduce an AI-enhanced web browser. Companies like Google have incorporated their Gemini AI models into Chrome, while others such as Perplexity AI are also launching AI-driven browsers. Following the OpenAI announcement, Google’s stock fell 4%, reflecting investor concerns regarding potential threats to its flagship browser, Chrome, the most widely used browser globally.

Source: www.theguardian.com

Kosmik: A Visual Canvas Equipped with PDF Reader and Web Browser Functionality

In recent years, tools like Figma, TLDraw, Apple’s Freeform, and the Easel feature in the Arc browser have tried to sell the idea of ​​using an “infinite canvas” to capture and share ideas.french startup cosmic builds on that general concept with knowledge acquisition tools that don’t require users to switch between different windows or apps to retrieve information.

Kosmik was founded in 2018 by Paul Rony and Christophe Van Deputte. Prior to that, Ronnie worked as a junior his director at a video production company, but instead of files and folders on which he could place videos, PDFs, websites, notes, and drawings, Ronnie used a single whiteboard as his type of canvas. was needed. And that’s when he started building his Kosmic, Ronnie told his TechCrunch. He draws on his background in the history and philosophy of computing.

“It took us almost three years to create a working product that included baseline features like data encryption, offline-first mode, and built a spatial canvas-based UI,” Rony explained. “We built all of this on top of his IPFS, so when the two of us collaborate, everything is peer-to-peer instead of relying on a server-based architecture.”

Image credits: cosmic

Kosmik offers an infinite canvas interface where you can insert text, images, videos, PDFs, and links, which can be opened and previewed in the side panel. It also has a built-in browser, so users no longer have to switch between windows to find relevant links on his website. Additionally, the platform also features a PDF reader that allows users to extract elements such as images and text.

This tool helps designers, architects, consultants, and students to build information boards for various projects. This tool is useful because it doesn’t require you to open numerous Chrome tabs and enter details in documents. Documents are a less visual medium for many different types of media. Some retail investors use apps to monitor stock prices, and consultants use apps for project boards.

Image credits: cosmic

Ronnie emphasized that bringing these different tools together in one place is a core selling point for Kosmik.

“I think it all revolves around the idea that we don’t have the best web browser or text editor or the best thing. It’s a PDF reader,” Ronnie said. “But being able to have them exist together in the same place, and being able to drag and drop items between them, makes this tool very powerful.”

Available via the web, Mac, and Windows, Kosmik comes with a basic free tier, which has a limit of 50 MB of files and 5 GB of storage with 500 canvas “elements.” For more storage and unlimited elements, the company offers a monthly subscription of $5.99, and eventually he offers a “one-time” subscription for those who only want to use the software on one device. We are planning to offer a “pay-as-you-go” model.

double down

Cosmic also announced today that it has raised funds. $3.7 million in seed round of funding led by Creandum. Alven, Kima Ventures, Betaworks, Replit and Quizlet founders participated.

Hanel Baveja, a principal at Creandum, told TechCrunch that the company decided to invest in Kosmik because Kosmik is a bit like Notion or Miro, and the potential to build something that completely changes an organization’s workflow. He said this is because there is. But Babeja said that like any consumer tool in this space, the startup needs to create immediate value for users.

“The time to value for any product must be immediate. Especially if it aims to become a commodity, you only have one chance to attract users,” Babeja said. “Finding a balance between a rich feature set and ease of deployment is certainly one of the challenges and is an area where the Kosmik team continues to strive.”

This cash injection is also timely given the product iterations in the pipeline.As expected of Cosmic is consolidating its codebases and Kosmik 2.0 will bring feature parity. The new app will be web-based and the desktop client will essentially be a wrapper app.

Additionally, the new version includes features such as multiplayer collaboration and AI-powered automatic tagging of items in images.

Ronnie said that in multiplayer mode, you can collaborate with someone on just a portion of the canvas using “cards,” which are like folders with objects dropped into them, rather than sharing the entire board. .

Kosmik opened to users in March and currently claims to have around 8,000 daily users, but the product can work completely offline, making it difficult to determine exactly how many people are actively using it. said it was difficult.

It’s worth noting that Kosmik isn’t the only startup active in the personal whiteboard space. Berlin-based Deta is building a new cloud OS for this problem and solution. sanity Building a social knowledge sharing platform. These companies must compete in some way to capture users’ attention and persuade them to try new paradigms for acquiring knowledge.

Source: techcrunch.com

Firefox’s Android browser now features 450+ new extensions after 3-year renovation

The experience has gotten even better for some markets who choose Firefox instead of Chrome, the default browser on Android devices. Mozilla, the maker of Firefox, today announced the release of over 450 new extensions (also known as browser add-ons). These are currently available on his website at Mozilla. Firefox browser add-ons page. These extensions allow users to customize their mobile browser to suit their needs. Tracking protection privacy tools, content blockerproductivity tools, or other features that introduce new experiences. streaming music, or anything that allows users to personalize their browser’s user interface – etc. Switch all websites to dark mode or provide A better way to manage tabs.

The lack of extensions has been a problem for Firefox Android users for years. 2020 launch of rebuilt version of mobile browser This replaces the app’s previous codebase with a new, faster, more customizable browser engine called ‘GeckoView’. At the time, the company said he had made a decision Limits supported extensions to those in the “Recommended Extensions” program, that is, extensions that are commonly installed by end users. This choice allowed Mozilla to quickly get the new browser into consumers’ hands, but it also killed the long tail of extension development and opportunities for software developers focused on this market. .

Nightly builds of Firefox have since enabled many more extensions, but hundreds of these extensions were not accessible in the publicly available Firefox browser for Android. This meant that most mainstream users of Firefox also had no access to extensions.

In August of this year, Mozilla said We finally have the infrastructure needed to bring the open extension ecosystem back to Firefox for Android. We then began testing hundreds more extensions for Firefox Android users, culminating in today’s news that we now have over 450 extensions available.

The company highlighted the importance of having an open ecosystem, noting that nearly half of all Firefox desktop users install extensions to customize their experience.

Many of the recommended extensions for Android browsers have six-figure user numbers or more, but the apps themselves only hold a small portion of the mobile browser market due to the traction that default browsers Chrome and Safari have gained in the mobile browser market. I have not. Android and iOS respectively. On mobile devices, Firefox has a market share of only 0.5% as of November 2023. According to data from StatCounter. In comparison, Chrome’s share is 64.23%. Safari, Samsung’s browser, Opera, etc. are also ahead of Firefox in the mobile browser competition.

Still, the app has a small but loyal following, including those looking for alternative options outside of big tech companies. One of its key selling points is automatic tracker blocking, which is appealing to those who value privacy. According to Google Play Store data, Firefox for Android has surpassed 100 million installs to date.

“The opportunities for innovation are huge,” said Giorgio Natiri, Director of Engineering at Firefox. announcement About launching extensions. “It’s exciting to see extension developers capitalizing on this moment and creating novel browsing experiences and features for Firefox for Android users. You don’t have to browse the mobile web with Extensions, which give you the freedom to change the way Firefox for Android looks and behaves. As more developers innovate in this exciting new field, things get even better. It will be,” he added.

Source: techcrunch.com

Beta Version of Windows Client Released by Arc Browser

The Browser Company today announced that it has onboarded its first set of users for its Windows client, launching it in beta today. Users and observers have been demanding a Windows version of Arc, especially after the company made its Mac client available to everyone after being in invite-only mode since 2022.

The company said it will bring more beta users on board this month and “rapidly” increase the rollout of invites in January. The company said it has more than 500,000 users I’m on the Windows waiting list.

The Thrive Capital-backed startup did not say whether the Windows client will have the same full functionality as the Mac client. However, he also mentioned that he is currently working on porting features such as Peek (a preview of his web page when hovering) and Mini Arc. Therefore, it is expected that some features will be missing at launch.

Last year, the founders josh miller and Harsh Agrawalhas talked about working on a Windows version and shipping it in 2023. And with only a few days left in the year, the company launched its Swift-based browser on the platform.

The company said its core team building on Windows included Alexandra Medway, product engineering lead, Darin Fischer, former vice president of Chrome, and Saleem Abdulrasool, a former Facebook engineer and Swift expert on Windows. I did. As the company develops, I’m trying to open source my code For Windows.

The Arc browser has had a busy year.The company introduced it in April. iPhone companion app A browser that allows users to save web pages in various workspaces and easily access them later from a desktop client. In May, it released a tool that lets you adjust web pages and remove some elements, such as the user suggestion box on X (formerly Twitter). In October, it finally rolled out AI-based features like renaming files and pinned tabs, page summary previews, and accessing ChatGPT from the command bar. Later that month, the company released a feature that allows users to share their folders and spaces and split views with non-Arc users.

Source: techcrunch.com