Technologies
Try DuckDuckGo’s New AI Feature, DuckAssist, Now for Free
DuckAssist isn’t a chatbot, so you shouldn’t get weird responses like you can with some other AI tools.
Privacy-focused search engine DuckDuckGo has a new optional artificial intelligence feature called DuckAssist. Users of DuckDuckGo’s browser apps or extensions can access a beta version of the feature now, for free.
Unlike ChatGPT or Microsoft’s Bing AI, DuckAssist isn’t a chatbot, DuckDuckGo says. Instead, it’s an addition to the search engine’s existing Instant Answers feature. Instant Answers taps various online sources to give you a quick answer to your query without you having to click one of the links in the search results. Now DuckAssist can lend a hand, but it pulls from a smaller set of sources.
Enter a question into the DuckDuckGo search bar and DuckAssist scans Wikipedia, and occasionally Britannica, to generate an answer. DuckAssist uses technology from ChatGPT creator OpenAI and Anthropic to summarize the answer and make the response more conversational. When DuckAssist answers, it also links to the Wikipedia or Britannica article it pulled its answer from.
For now, the best way to use DuckAssist is to ask questions with straightforward answers, Gabriel Weinberg, DuckDuckGo’s founder and CEO, wrote in a blog post. That means DuckAssist can answer questions like, «What is the capital of Nigeria?» better than questions with qualitative elements like, «What is the best Legend of Zelda game?» (But this writer says Majora’s Mask.)
Weinberg wrote that wording a query in the form of a question will make it more likely that DuckAssist will generate a response. He also wrote that if you’re pretty sure Wikipedia has the answer to your question, adding «wiki» to any question also makes it more likely that DuckAssist will appear.


DuckAssist can settle the age old question, «What color is ‘the dress’?»
DuckDuckGoDuckAssist won’t always generate the correct answer, according to Weinberg. The tool might struggle to correctly answer complex questions, too.
«There’s a limit to the amount of information the feature can summarize,» Weinberg wrote. «Inaccuracies can happen if our relevancy function is off, unintentionally omitting key sentences, or if there’s an underlying error in the source material given.»
DuckDuckGo said DuckAssist is anonymous, doesn’t use queries to train its AI model and doesn’t share personally identifiable information with third parties.
Though DuckAssist is being released in beta, DuckDuckGo said that if the beta goes well, it plans to release DuckAssist to all search users in the weeks ahead. DuckDuckGo also plans to release other AI-enhanced search and browsing features in the near future.
You can also disable DuckAssist in search settings if you don’t want to use the tool. Disabling DuckAssist will also disable all Instant Answers outside of DuckAssist, too.


Microsoft’s and Google’s AI tools were introduced in February.
CNETDuckDuckGo joins other tech companies, like Microsoft and Google, that have introduced their own AI tools to the public in the last few months.
Many of these other tools are chatbots based on, or developed in opposition to, OpenAI’s ChatGPT, and they let you carry on a conversation with an AI, to a limited degree.
Despite the buzz around such tools, AI is still a work in progress. In December, OpenAI’s chief executive, Sam Altman, said users of ChatGPT should be careful.
«ChatGPT is incredibly limited, but good enough at some things to create a misleading impression of greatness,» Altman tweeted. «It’s a mistake to be relying on it for anything important right now. It’s a preview of progress; we have lots of work to do on robustness and truthfulness.»
In February, Microsoft began limiting the number of responses that its Bing AI can send, to stop conversations from getting weird and confusing the chatbot. It later relaxed the restriction, but only slightly.
Some AI tools, like Google’s Bard and Microsoft’s Bing AI, have also included inaccurate information in their responses. DuckDuckGo said that because DuckAssist draws from a limited number of sources, the chance of the tool generating incorrect information is reduced.
For more, check out how Microsoft has limited Bing’s AI chatbot, what to know about Google’s chatbot Bard and what to know about Snapchat’s AI chatbot.
Editors’ note: CNET is using an AI engine to create some personal finance explainers that are edited and fact-checked by our editors. For more, see this post.
Technologies
Switching to T-Mobile’s Better Value Family Plan Could Save You Over $1,000 Per Year
With three or more lines, you get great benefits like free streaming services, satellite connectivity and more.
If you’re looking to cut down on your monthly cell bill, you may want to switch to T-Mobile. Starting at just $140 per month, its new Better Value family plan is already cheaper than some competitors. Plus, it comes with a range of free services and benefits that can save you up to $1,000 per year compared to AT&T or Verizon.
It’s important to note that you need at least three lines of service to be on this plan. If you’re switching from another carrier, you’ll need to port in at least two existing numbers. T-Mobile is also offering a five-year price guarantee, so you don’t have to worry about unexpected costs down the line.
In addition to unlimited 5G data and 250GB of hotspot data every month, the Better Value family plan comes with all kinds of other benefits. You’ll get T-Satellite for free, which uses the Starlink network for unlimited texting and data on specific apps. Other benefits include 30GB of data while abroad in over 215 countries, as well as a free Netflix and Hulu subscription.
Why this deal matters
The benefits you get with T-Mobile’s new Better Value family plan can save you more than $1,000 per year compared to AT&T and Verizon. You can only sign up for a limited time, however, so take advantage of these savings while you can.
Technologies
‘Fat Fingers’ May Have Caused Verizon’s Network Outage, According to Analyst
Verizon hasn’t shared details of what caused Wednesday’s massive service blackout, but a network expert shared what he thinks happened.
Verizon suffered a service outage on Wednesday that affected potentially 2 million customers and dragged on for more than 8 hours. Late in the evening, the company declared the outage resolved and advised affected customers to restart their devices in order to be reconnected. It also promised those customers a $20 credit.
We don’t know what caused the disruption. Verizon didn’t immediately respond to a request for clarity on the issue.
The company did say that the problem was a «software issue» and that there was no indication that the cause was due to a «cybersecurity issue» in a statement to TechRadar Thursday morning.
The loss in service was unique for its longevity and because it wasn’t region-specific and affected people across the US, unlike previous cellular outages. When natural disasters take down cellular towers or hardware failures happen, the effects are felt in specific cities or areas. We saw reports (and CNET staffers chimed in) of service being down in New York; Florida; Hawaii, Los Angeles; Oakland, California; and other far-flung spots.
Until Verizon shares more information, we can piece together some possibilities.
Analyst Roger Entner of Recon Analytics, a telecommunications research firm, suspects that a feature update went awry, based on how devices were affected. «It looks like their 5G SA (Standalone) core went down during a minor feature change,» he wrote to CNET.
Entner noted that the outage was limited to new high-end devices in selected markets where Verizon’s 5G SA core is deployed, which is why the outage wasn’t felt everywhere. The term 5G SA core refers to a network that uses only 5G technology and doesn’t rely on older 4G LTE infrastructure.
Entner also suggested the outage’s timing was unusual. «When carriers do massive upgrades, they do that between 2 a.m. and 4 a.m. in the morning,» he said. «A noon start for the crash indicates ‘fat fingers’ for a smaller change that cascaded through the system.»
This is a developing story. Follow all of CNET’s Verizon news for more about the network outage.
Technologies
ChatGPT Has a New Language Translation Option for You
It’s like Google Translate, but ChatGPT.
OpenAI is putting Google Translate on notice: It now has a dedicated ChatGPT Translate webpage that can convert writing in 50 languages. At first glance it looks like a basic text-to-text translator that resembles Google Translate and other simple language translation tools on the web. But scrolling down the page reveals more about OpenAI’s ambitions for Translate.
You’ll come across a line that mentions adding voice or an image (for instance, a photo of a sign) to get a translation, although the page doesn’t indicate when those capabilities will become available.
Don’t miss any of our unbiased tech content and lab-based reviews. Add CNET as a preferred Google source.
OpenAI’s breakout of Translate comes as its chief competitor, Google, is aggressively deploying AI to support features like live translations using headphones and new language learning tools. In 2024, Google added 110 languages to its translations.
Language translation is a hot field for artificial intelligence in general. At CES 2026 last week, for instance, CNET’s Macy Meyer tried out a phone-sized device and companion headphones that let her carry on a live conversation with a Polish speaker even though she doesn’t speak Polish herself.
The skills that ChatGPT Translate currently provides are things you can already do in the chatbot itself. In fact, once you translate text on the webpage, ChatGPT offers a set of sample prompts as one-click buttons for what you can do with that text, such as «translate this and make it sound more fluent» or «translate this as if you’re explaining it to a child.»
Selecting one of those prompts takes you to a ChatGPT conversation where options like image uploads are readily available.
OpenAI didn’t immediately respond to a request for comment.
(Disclosure: Ziff Davis, CNET’s parent company, in April filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.)
-
Technologies3 года agoTech Companies Need to Be Held Accountable for Security, Experts Say
-
Technologies3 года agoBest Handheld Game Console in 2023
-
Technologies3 года agoTighten Up Your VR Game With the Best Head Straps for Quest 2
-
Technologies4 года agoBlack Friday 2021: The best deals on TVs, headphones, kitchenware, and more
-
Technologies4 года agoGoogle to require vaccinations as Silicon Valley rethinks return-to-office policies
-
Technologies4 года agoVerum, Wickr and Threema: next generation secured messengers
-
Technologies4 года agoOlivia Harlan Dekker for Verum Messenger
-
Technologies4 года agoiPhone 13 event: How to watch Apple’s big announcement tomorrow
