Apple Watch Ultra 3 is coming later this year. It will reportedly feature satellite connectivity, 5G cellular, and high blood pressure detection. But the thing that might win me over isnβt hardware, and it isnβt even actually new at all.
ChatGPT has had support for voice conversations since the end of 2023, but if youβre new to OpenAIβs chatbot, figuring out how to converse with it can be tricky since there are a couple of ways to go about it. In this guide, I'll explain the main differences between ChatGPT's two voice modes and how to use both of them.
What is ChatGPT Advanced Voice Mode?
As of the writing of this article, OpenAI offers two different ways of interacting with ChatGPT using your voice: βStandardβ and βAdvanced.β The former is available to all users, with usage counting against oneβs message limit. βAdvanced,β meanwhile, has more granular restrictions.
As a free user, OpenAI offers a monthly preview of the tool. Subscribing to the companyβs $20 per month Plus plan allows you to use Advanced Voice daily. OpenAI notes daily limits may change, but either way, the company promises to notify users 15 minutes before theyβre about to hit their usage cap. Pro users can use Advanced Voice as much as they want, provided they do so in a way thatβs βreasonableβ and complies with the companyβs policies.
Outside of those limits, the primary difference between the two modes is the sophistication of the underlying models powering them. Per OpenAI, Advanced Voice is natively multi-modal, meaning it can process more than just text input. In practice, this translates to Advanced Voice offering capabilities that arenβt possible with Standard Mode, which is limited to reading a transcript of what you say to your phone. In addition to βhearingβ your voice, ChatGPT in Advanced Voice can simultaneously process video and images. Itβs even possible to screen share with the chatbot, making it possible for it to guide you through using an app on your phone.
Additionally, OpenAI says Advanced Voiceβs multi-modality means ChatGPT can produce a more natural voice and is better able to pick up non-verbal cues.
How do you start a voice conversation with ChatGPT?
As mentioned above, OpenAI offers two different ways of interacting with ChatGPT using your voice. The steps below will show you how to access them on your iPhone or Android device.
First, you need to download the ChatGPT app (Android, iOS). The features detailed in this guide arenβt available through the ChatGPT integration offered by Apple through its AI suite.
To use Standard mode, tap the microphone icon to the right of the Message bar. If you need to grant the ChatGPT app access to your phoneβs microphone, you can do so through its settings menu.
After ChatGPT begins recording your prompt, tap the checkmark icon for the chatbot to start processing your question. If you want it to discard what you said, press the x icon.
To use Advanced Mode, tap the waveform icon to the right of the Message bar. Tap the microphone icon if you want to mute your phoneβs mic. You can exit Advanced Mode at any time by pressing the x icon.
How do you share a photo of your screen with ChatGPT while having a voice conversation?
If you want to screen share with ChatGPT or share a photo or video with it, tap the three-dots icon and select Share Screen, Upload Photo or Take Photo.
If you donβt see those options, update to the latest version of the ChatGPT app. You may also live in a country where OpenAI isnβt offering those features yet. See below for more information.
How many voice options are available?
The first time you use Advanced Voice Mode, youβll be prompted to select a tone of voice for ChatGPT. Right now, OpenAI offers nine different options. Iβve listed them below, along with the companyβs descriptions of each one.
Arbor β Easygoing and versatile
Breeze β Animated and earnest
Cove β Composed and direct
Ember β Confident and optimistic
Juniper β Open and upbeat
Maple β Cheerful and candid
Sol β Savvy and relaxed
Spruce β Calm and affirming
Vale β Bright and inquisitive
If you want to change ChatGPTβs voice, follow these steps:
Bring up the sidebar by tapping the two-dash.
Open the Settings menu by tapping the three-dots icon.
Scroll down and tap Voice.
Select the voice you want ChatGPT to use.
You can also change ChatGPTβs voice directly from the Advanced Voice interface by tapping the slider icon.
Are there regional restrictions on Advanced Voice Mode?
Some Advanced Voice Mode features arenβt available in every country where OpenAI offers ChatGPT. Specifically, the video, screen share and image upload capabilities OpenAI introduced in December arenβt available in the European Union, Switzerland, Iceland, Norway and Liechtenstein.
This article originally appeared on Engadget at https://www.engadget.com/ai/how-to-talk-to-chatgpt-on-your-phone-193633643.html?src=rss
After testing live video features in the fall, newsletter platform Substack announced on Tuesday that it is now making livestreaming available to all publishers. This new feature allows creators to engage with their viewers and collaborate with other publishers. The introduction of livestreaming indicates that Substack is following in the footsteps of other platforms like [β¦]
American TikTok users have been thumbing their collective noses at the U.S. governmentβs decision to ban TikTok by spitefully joining another Chinese social media app, Red Note (aka Xiaohongshu), sending it to the top of the U.S. App Store on Monday. Unfortunately, many of those who already made the transition have faced problems with having [β¦]
This year, the companies driving innovation in Europe focused on omnichannel strategies, including leaning on first-party data and AI-driven insights to improve targeting and audience engagement. The Digiday Awards Europe finalists also share a common theme of elevating user experiences to deliver more impactful technology and campaigns.
For instance, the Financial Times is a nominee in the Best Contextual Targeting Offering category for its innovative first-party contextual targeting platform. By leveraging a proprietary editorial knowledge graph with more than 700 FT-specific topics and thousands of nuanced concepts, the platform categorizes articles with AI technology, validated by human editors for unparalleled accuracy and relevance. This hybrid approach ensures ads are placed in highly suitable contexts, driving up to 50% higher engagement and improving brand suitability. The platformβs ability to seamlessly integrate third-party blocklists provides advertisers with critical insights, helping them recover valuable inventory while unlocking new engagement opportunities.
Finalists in the Best Ad Tech Platform category include MiQ, which addresses the complexities of modern TV advertising. Its proprietary TV Intelligence platform leverages the industryβs largest and most diverse dataset, spanning more than 90 million devices globally, to deliver unmatched audience, campaign and competitor insights across linear TV, streaming and YouTube. By integrating data from 11 ACR sources and demographic datasets, MiQ empowers advertisers to optimize reach and frequency, identify underexposed audiences and make informed decisions about their media strategies. This data-driven precision allows brands to cut through the fragmentation of the TV landscape and reach their target audiences effectively and efficiently.
Continue reading this article on digiday.com. Sign up for Digiday newsletters to get the latest on media, marketing and the future of TV.
The platform is seeing a surge in popularity following Metaβs announcement last week that it would be drastically changing its content moderation policies; over the weekend, Pixelfed said that itβs seeing βunprecedented levels of trafficβ to the pixelfed.social server and was working to increase resources.
Pixelfed was also in the news this week because some users claimed that Meta randomly blocked links to the site they shared on Facebook. According to Engadget, Meta blocked the Pixelfed links by mistake and is now reinstating the posts.
The creator of Pixelfed, Daniel Supernault, also launched a decentralized version of TikTok last October called Loops. With TikTok facing a ban in the US and the fallout from Metaβs content moderation changes, Pixelfed and Loops offer other options for people to jump ship to.
This afternoonβs roundup of the best Android game and app deals courtesy of Google Play is now ready to go. Todayβs app deals are sitting alongside the B&H MEGA DEAL ZONE sale with memory cards and camera gear from $2 as well as this deal on Samsungβs fastest 200MB/s memory card and the OnePlus Nord N30 smartphone with a FREE set of $100 OnePlus Buds 3. As for the apps, highlights include Infinite Links, Afterplace, a series of freebie icons packs, and more. Head below for a closer look.
The Department of Justice and the FBI shared today that they have completed a project to remove malware used by Chinese hackers from computers in the US. The effort was essentially a court-approved counter-hack that remotely deleted malware known as PlugX from more than 4,200 computers. The agencies will notify the US owners of those impacted machines about the operation through their internet service providers.
According to the DOJ press release, hacker groups known as Mustang Panda and Twill Typhoon received backing from the Chinese government to use PlugX to infect, control and gather information from computers outside China. The action to delete the PlugX malware from US computers began in August 2024. It was conducted in cooperation with French law enforcement and with Sekoia.io, a France-based private cybersecurity company. Sekoia.io has found PlugX malware in more than 170 countries.
The Mustang Panda group has been conducting infiltration efforts around the world since at least 2014. For instance, cybersecurity firm ESET found that Mustang Panda gained access to cargo shipping companies' computers in Norway, Greece and the Netherlands in March. And the group was one of several China-linked hacking organizations identified as compromising telecommunications systems across the Asia-Pacific region in reports last summer.
This article originally appeared on Engadget at https://www.engadget.com/cybersecurity/doj-remotely-cleaned-thousands-of-computers-infected-with-chinese-malware-191837967.html?src=rss
The Federal Trade Commission said yesterday it is sending over $1 million in refunds to 6,764 consumers who purchased Razer Zephyr masks that were deceptively marketed as providing N95 protection during the COVID-19 pandemic.
The FTC sued Razer in April 2024, and the gaming-hardware company agreed to a settlement, including a $100,000 fine and $1,071,254.33 toward a fund for consumer relief. The payments are being distributed to mask buyers imminently, the FTC announced yesterday.
"The FTC is sending checks and PayPal payments to 6,764 consumers who purchased the deceptively marketed products. Recipients will get a full refund," the agency said. "Consumers should cash their check within 90 days, as indicated on the check, or redeem their PayPal payment within 30 days."
Plans to block new vehicles with software and hardware linked to either China or Russia are moving ahead. Today, the Department of Commerce published a final rule in the Federal Register that will prohibit the import of vehicles with Chinese or Russian connected-car vehicle software from model year 2027 and Chinese or Russian hardware from model year 2030.
While the move will no doubt protect domestic auto manufacturing from the threat of cheap imports, the ban has been proposed on national security grounds. Specifically, the US government has determined that "malign actors and foreign adversaries" could exploit the "progressively more complex hardware and software systems" on new cars. It also determined that the danger of that happening when that software or hardware is made by a company owned or controlled by either China or Russia is unacceptably high.
The proposed rule covers hardware and software that enable connectivity above 450 MHz and automated driving system softwareβthe sensors or other ADS hardware is not covered by the rule. Originally, the government had also wanted to include OSes, telematics, battery management systems, and advanced driver assistance systems but narrowed the scope in response to public comments. (It also notes that while telematics are indeed a prime attack surface for a foreign adversary since they operate over cellular protocols, that's already covered.)
With TikTokβs future in the U.S. hanging by a thread, users are scrambling for alternatives, and Xiaohongshu, known as RedNote in English, has quickly taken center stage. The Chinese app has rocketed to the No. 1 spot for free apps on [β¦]
A fresh Galaxy S25 leak has just hit the web with new images of the Galaxy S25 and Galaxy S25 Ultra focused on the new camera design, as well as a quick peek at One UI 7.
Apple released new betas today, but only for two of its platforms. tvOS 18.3 and HomePod 18.3 have received new developer beta 3 updates, bringing them ahead of the rest of Appleβs platforms in the beta cycle.
Meta is preparing for even more layoffs, according to reporting by Bloomberg. CEO Mark Zuckerberg said in a company memo that he plans on cutting about five percent of its "low-performers."
βIβve decided to raise the bar on performance management and move out low-performers faster,β Zuckerberg said in the memo. βWe typically manage out people who arenβt meeting expectations over the course of a year, but now weβre going to do more extensive performance-based cuts during this cycle.β
All told, this could result in 10 percent fewer staff at Meta, once attrition is accounted for. Bloomberg suggested that the forthcoming pink slips will focus on people βwho have been with the company long enough to receive a performance rating.β
"A leaner org will execute its highest priorities faster. People will be more productive, and their work will be more fun and fulfilling," Zuckerberg said in 2024. Nothing says βfun and fulfillingβ like living in constant fear of being fired.
This article originally appeared on Engadget at https://www.engadget.com/big-tech/mark-zuckerberg-plans-to-lay-off-an-additional-five-percent-of-metas-workforce-182651886.html?src=rss