Category: News

Google

Google: “people have been finding ads within AI overviews helpful”, so we are expanding them

Posted on by Martin Brinkmann

When you run a search on Google, you may get an AI Overview at the top of the results. Google announced an expansion of the AI-powered feature to roll out in more markets and for more languages.

Google revealed information about its AI plans yesterday on its Ads & Commerce blog. There, the company noted that AI Overviews is “one of the most successful launches in Search in the past decade”. Google saw a 10 percent increase in usage of Google for queries that displayed the overview generated by Google’s Gemini AI.

As a consequence, Google is expanding ads across AI Overviews. In particular, Google says that it is expanding ads in AI Overviews in English to more countries on mobile and desktop, and Search and Shopping ads in AI Overviews for desktop users in the United States.

Ads will also be tested in AI Mode, which Google calls its “most powerful AI search” feature. AI Mode is limited to Google Search users from the United States at the time of writing.

In other words: expect (more) advertisement in AI Overviews and other AI tools in the coming months and years.

Clearly, running AI is expensive. While companies like Google, Meta, or Microsoft have the means to sustain development, infrastructure and operations, the clear goal is the direct or indirect generation of revenue from these services.

Direct revenue may come from subscriptions or ads, indirect revenue from using data for training or other purposes.

The consequences are clear: users should expect to see more and more ads baked into AI services. This looks like a testing ground currently, but it looks similar to how Google introduced ads in search. First, some ads, separated clearly from search results. Now, Google users get more ads than organic search results in some queries and there is no longer a clear distinction between the two.

Now You: do you use AI services? What is your take on ads being integrated into those?

Gemini Google Chrome

Gemini in Chrome: Google integrates AI directly into the browser

Posted on by Martin Brinkmann

Google announced Gemini in Chrome yesterday. The integration of the AI into Google’s browser is the next step in Google’s masterplan to spice up its products using artificial intelligence.

Chrome users in the United States who are subscribed to Google AI Pro or the new Google AI Ultra plan, will be the first to gain access to the AI.

The initial version closes the gap to other browsers that use AI already to summarize webpages for users or allow users to ask questions about the content of a website.

Gemini for Chrome will do the same initially. You activate the AI tool with a click on its icon in Chrome. From there, you can ask questions or give it instructions. The first iteration is limited to the active webpage. You can, for instance, ask it to explain certain concepts to you or provide a summary of the key points.

Google included the following examples in a demo video:

  • Make a regular recipe gluten-free.
  • Helping a student understand the differences between chemical bonds.
  • Adding a reminder to calendar.
  • Asking Gemini which plant is best from a selection of plants in open tabs.
  • Asking Gemini which poses one should do for a racing podium.

Note: it is a good idea to verify information generated by AI to make sure it is accurate.

Gemini in Chrome: goal is full access to all tabs and agentic functionality

Google is working on improving this base functionality. The company plans to give Gemini access to all open tabs. This enables more features, including the ability to compare different webpages or products or taking everything into account when generating a response.

Gemini will also be able to open websites on behalf of the user, according to Google’s announcement. Google did not explain why users would want the AI to open webpages on their behalf.

Closing Words

Like it or not, AI is going to be integrated into the majority of web browsers. Not everyone will see the use of this, but this will surely expose AI tools to more users. This could change how users use web browsers, searches, and the Web significantly.

Now You: do you use AI features or tools in your browser of choice already? Let me know in the comments below.

Opera GX

Opera GX update introduce new usability features, including one that is a first

Posted on by Martin Brinkmann

When was the last big usability feature introduced in your favorite browser? One that is new and unique? Not speaking of tab groups and vertical tabs support in Firefox. While great and new to users of the browser — if you ignore extensions and Mozilla’s decade-old attempts, they are not particularly new features.

Google and other Chromium browsers supported tab groups for quite some time, and some browsers, like Vivaldi, supported vertical tabs as well. Opera introduced AI-powered tab commands recently and an AI browser operator that acts as the user’s agent.

Opera announced three new features for its gaming browser Opera GX the other day. Not all of them are new as well, but some are. It is a browser for gamers, featuring gaming news and all kinds of gaming related features, including themes and sounds that Opera thinks appeal to gamers.

It is not my preferred browser, but Opera continues to introduce features to it that I find interesting. This particular update introduces three:

  • Split tabs — this enables you to display two open websites side by side. Opera says this is ideal for keeping an eye on the game you are playing while using another tab to chat or look up information.
  • Tab traces — if you have lots of tabs open, you know that it is sometimes difficult to switch to a tab was active recently. While most browsers on the desktop support the shortcut Ctrl-Shift-Tab or Ctrl-Tab to go back to the previous active tab. Tab traces in Opera GX highlights the five tabs that were active previously, making it a tad easier to switch back and forth between tabs. The minimum number of open tabs for the feature in Opear GX can be configured in the settings.
  • Automatic tab islands (tab groups) — Opera GX will create tab groups automatically, if you open links from a page in a new tab. This bundles all links together, so that you get handy options, such as collapsing all tabs at once.

You can check out the blog post on the official Opera blog for additional information.

Split tabs is a useful feature, provided that the browser window is large enough to display two websites side by side without side-scrolling. Tab traces sounds like a useful feature, as it may enable you to locate the previous tabs quickly. Good option for users who do not use shortcuts.

What is your take on these features? Good additions to the browser? Do you use similar features in your browser of choice?

Security

Firefox 138.0.3 fixes two crashes and some other issues

Posted on by Martin Brinkmann

Mozilla plans to release a new point update for its open source Firefox web browser later today. Firefox 138.0.3 is a non-security update that fixes three issues in earlier versions of the web browser, including one crash issue.

If you use Firefox on the desktop only, you may ask yourself if you missed the Firefox 138.0.2 release. You may have, as this special version was only released for Firefox on Android.

Firefox 138.0.3: the fixed issues

Firefox 138.0.3 fixes a WebGL-related crash. Mozilla notes that Firefox could crash when websites that use WebGL were opened. WebGL, which stands for Web Graphics Library, is used to render 2D and 3D graphics using JavaScript.

Web games and other interactive or animated content may make use of WebGL. So, if you ran into crashes while running the recent version of Firefox, there is a good chance that this was caused by the WebGL bug.

The update fixes a second crash. This one occurred on websites where certain SVG effects were applied to very small areas.

Firefox 138.0.3 fixes two additional issues:

  • Videos looked washed out on Linux under Wayland, if HDR was not supported on the system.
  • The keyboard shortcut Alt-C did not work as expected in the sidebar search.

Mozilla has not published the official release notes at this time. The release notes will be published shortly after the Firefox 138.0.3. You can access them here afterwards to check them out.

The new update is only available for the desktop versions of the Firefox web browser. Furthermore, Firefox ESR are not updated, as they appear to be unaffected by the issues.

Mozilla plans to release the next major version of Firefox and Firefox ESR on May 29th, 2025. The following versions will be released on the day:

  • Firefox 139
  • Firefox 115.24 ESR
  • Firefox 128.11 ESR

Now you: where you affected by one or several of the fixed bugs? How do you handle point updates usually? Do you install them right away, postpone their installation, or skip them? Feel free to leave a comment down below.

Flickr restricting free account downloads to push Pro subscriptions

Posted on by Martin Brinkmann

The photo hosting and community site Flickr has been a popular option for free and paying users from all over the world to back up their photos to online storage, share them with others, find new photos to download, and to interact with other community members.

The site changed its owner years ago and since have been getting worse for free users ever since. First, new owners SmugMug limited free users to 1000 photos, down from 1 terabyte of photo storage, and to 50 non-public photos.

Now, starting May 15th, 2025, free users face another limitation on the site. Their ability to download photos is limited to 1024 pixels and less in resolution. Means: free users can’t download the original resolution of a photo anymore nor the large version of it, including their own uploaded photos.

Why Flickr is making the change: Flickr says that it is implementing the change to address “the misuse of free accounts as cloud storage for original files”. This voilates Flickr’s terms of services and affects the performance for paying customers negatively.

The change does not affect the uploading of photos “of all accepted file sizes to a free account”. Flickr notes that creative commons photos can continue to be downloaded in all available sizes, provided that they are not set to private. Additionally, Flickr Commons members are also exempt from the change.

Photos will display in high quality on the site and “embeds and external links” won’t break either.

Flickr recommends upgrading to a paid Pro account to keep the downloading functionality.

Closing Words

Is Flickr still a popular option for photographers and still going strong community-wise? Many of the recent changes limit free users on the site, likely to push more users into subscribing to a Pro account. I can’t really say how well the strategy works, but the limitations are certainly driving some Flickr users to different photo hosting platforms.

Speaking of which, do you upload photos to online photography sites or photo storage sites? Or do you prefer to keep photos and images on your local devices and systems only? Feel free to leave a comment down below.

Google Maps is getting a new feature that you either love or dislike

Posted on by Martin Brinkmann

Google Maps is quite the handy helper-app in many parts of the world. While it is not excellent everywhere, many use it to find restaurants, shops, places of interest, or for directions.

Soon, Google Maps is getting a new feature that is using Google’s Gemini AI to identify places in screenshots you take so that you can save them to a list.

Here is how Google describes the feature:

If you ever have trouble keeping track of all of the screenshots you take of travel blogs, news articles or social media posts when you’re researching places to go for an upcoming trip, you’ll want to try out this new Google Maps feature. It uses Gemini capabilities to identify places mentioned in your screenshots and helps save them to a list for you, making travel planning a breeze.

Taking screenshots? Identifying the locations of place? Adding them to Google Maps? It may sound like a niche thing for many.

Good news is that you need to enable the feature to start using it. If you do not, none of your screenshots get analyze by Gemini.

Here are Google’s instructions on using the new feature:

  1. Navigate to the You tab.
  2. At the top of the You tab, you’ll see a Screenshots list with a badge that says “Try it out!” Tap the badge. This will open a video showing you how the feature works — you’ll even be able to test it out alongside the video.
  3. You’ll see a request to allow Google Maps access to your photos. Choose when it has access. (There’s also a manual option if you don’t want to give Maps full access, more on that later.)
  4. Next time you take a screenshot that includes location information, head to Google Maps after.
  5. If Maps recognizes a place, a message will pop up telling you that it has places ready for you to review.
  6. Tap “review,” and decide if you want to save the image to your screenshots list or not. (You can add these images to other lists later, too, if you want.)
  7. You also have the option to upload screenshots manually: When you’re in the app select the Screenshots list under the You tab. From there, you can upload screenshots manually. The rest of the process is the same!
  8. You’ll see the saved places directly on your map or can get to the list through You tab to access while you’re on the go.

It sounds as if the image needs to have location information for the feature to work, but Google also says that the feature uses Gemini to identify locations. Google added several safeguards to the process to block users from using the feature unknowingly.

So, who is this feature for? Google gives the answer in the introduction. It is for people how take screenshots of travel blogs, news articles, or social media. It may also work for users who save images that they find, as they can upload them to Google Maps to see if Gemini can identify the locations.

It can be handy, if you do your research mostly in apps like Instagram or TikTok, especially if location information is not revealed by the poster.

Those who do not can safely ignore the feature, as it needs manual activation.

Note: Google may gain access to all screenshots that you take on your devices, if you enable the feature.

Now You: what is your take on the feature? Handy and you would use it, or something that you would not want to use no matter what?

Expect more ads in AI chats soon, courtesy of Google

Posted on by Martin Brinkmann

How do many of the AI startups make money? Many have launched products that are freely available and without ads. There are multiple answers to the question, depending on the business’ strategy. Some want growth at all costs to sell the business later on. Others plan to introduce ads or subscriptions at a later point in time.

Soon, users from all over the world will see more ads in chats with AIs, as Google is expanding its Adsense program to chats.

Bloomberg reports that Google is expanding Adsense to keep its edge in the advertising vertical. Tests with select startups such as iAsk and Liner have begun already reportedly.

New types of AI-powered services spring up virtually everywhere on the Internet. Liner, for example, is an AI-powered search engine, while iAsk promises answers to questions that users may have.

Most services have in common that they are not page-based websites. Blogs or regular websites publish static pages most of the time, whereas interactions with AI are usually dynamic.

While companies may display Adsense on the pages, they were not particularly optimized for the queries of users. This changes with the rollout of the new Adsense format.

Integration means that ads may show up in chats. Where and how depends on the AI service, but users should expect to see an increase in advertisement, especially on AI sites that are not operated by billion Dollar companies.

Good news is that content blockers should be able to deal with those ads, just like they are regarding ads in search or on websites.

Now You: do you use AI services regularly? If so, for what purpose and what do you like or dislike in particular when you compare the AI service to other means of acquiring the information?

Waterfox Private Search engine

Waterfox Private Search: a first look at the privacy-friendly search engine

Posted on by Martin Brinkmann

The developer of the Waterfox web browser announce Waterfox Private Search earlier this week. The new search engine is in open beta currently, which means that you can give it a try if you want to.

Waterfox Private Search pulls data from other search engines privately, unlike Brave Search, which uses its own engine for search results.

Here are the details:

  • Waterfox Private Search is a meta-search engine. Means, it will support several search engines and not just one. During the beta, results come only from Google Search though.
  • A proxy is used for communication with the supported search engines. Means, the search engines won’t see your IP address or other information.
  • The search engine won’t feature any AI content, e.g., AI summaries.
  • There will be two tiers in the long run.
    • An ad-supported tier, that shows “privacy-friendly advertisement” that won’t track users or create user profiles. These will be contextual ads based on the search.
    • A subscription-based tier that removes all advertisement.

You can point your browser to Waterfox Search to give it a go.

The search engine works as expected. You type in a query and get results. These come from Google, but it is refreshing to see that the main focus is on the search results and not added content. While you do get some, such as a “People Also Ask” module, the main focus is clearly on returning web links to the user.

A click on the options menu displays just a few. You can change the location there to get results from a different region and use the domain blocklist to exclude certain websites from the results.

You may also switch search engines, with Bing, Brave, and Mojeek listed, but not yet selectable.

A click on the themes icon in the main interface displays a good dozen or so themes that you can switch between. There should be something for everyone, from very light to dark, colorful and even cyberpunk.

As far as search options are concerned, there are just a few available as menus. A click on the settings icon shows an option to disable autocorrect and to set a specific time period for the search.

You may also switch from web search results to images, videos or news. Other options, including shopping or books are not available.

The loading of search results takes a bit longer, but not too long to be unpleasant. It remains to be seen if the loading time will change when traffic increases.

Initial verdict: Based on a preliminary test, Waterfox Private Search is a refreshingly focused search engine. I did not spot any ads during my tests, but these will come in the future. Privacy-friendly ads are not anything new, but they are more than welcome and if they are implemented in an unobtrusive way, likely not a problem for the majority of Private Search users.

There is certainly the question of sustainability. Ads play a key role here and so does the subscription-based tier. This only works if enough users sign up or see ads when they use the search engine.

Now You: which search engine is your favorite currently and why? Feel free to leave a comment down below to join the discussion.

Yahoo is working on a web browser and eying a Chrome acquisition

Posted on by Martin Brinkmann

Yahoo, once an Internet powerhouse, is only a shadow of its former self. This is not keeping company officials from working on new products and services to restore some of the former glory of the company.

Yahoo, apparently has been working on a browser prototype for the past couple of months and expressed interest in acquiring Google Chrome in a “from zero to hero” type of scenario.

The information was revealed on the fourth day of Google’s antitrust trial. Yahoo is not the only company that expressed interest in acquiring Chrome. The AI companies OpenAI and Perplexity have also expressed interesting in buying the world’s most popular web browser.

General manager of Yahoo Search, Brian Provost, testified before the court that about 60 percent of all search queries were made through web browsers. Chrome, the leading browser on mobile and desktop, is responsible for the bulk of these searches.

On desktop, Google Chrome has a market share of over 66 percent according to Statcounter. Since Google owns Chrome and has made Google Search the default search engine, it is fair to say that the bulk of searches that originate from browsers is flowing through Google Search. Google has also struck deals with Apple and Mozilla to make its search engine the default.

While Yahoo is prototyping a browser, it is in discussions with “other companies about buying a browser”. Provost did not reveal the names of the companies. An acquisition would speed up the entire process.

Yahoo would replace the default browser with Yahoo Search to push its own search engine to new heights. Interestingly enough, Yahoo Search is powered by Microsoft’s Bing search engine. It still has a market share of about three percent.

Now You: If Google is forced to sell Chrome, which company would be your favorite to buy the browser?

AI

Why you may not want to say “thank you” or “please” to AI

Posted on by Martin Brinkmann

What is the proper etiquette when chatting with AIs? Do you say “thank you”, “please” or “see you next time”, or do you prefer the direct approach and avoid any form of politeness when communicating with AI tools?

OpenAI CEO Sam Altman was recently asked about the electricity costs that these pleasantries would generate. According to him, processing these cost “tens of millions of dollars”.

If everyone would stop being nice to AI, at least in this regard, AI companies electricity bills would go down noticeably. While that has no immediate benefit to the user, it is at the very least beneficial to the environment.

Altman ended the answer with a joke, stating that the money was well spend as “you never know”, referring to the possibility of an AI uprising and the potential saving of people who have been nice to AI in the past.

You might say that AI causing havoc to bring doom to the human race might not be persuaded to let humans live based on their use of “please” or “thank you”, and you are probably correct here.

Where does this leave the statement? When you avoid please or thank you, you will help AI companies save power, which is beneficial to them, but also the environment, at least potentially.