Google employees have been testing several potential ChatGPT competitors as part of the tech giant’s bid to launch a response to OpenAI’s technology, according to CNBC. A previous New York Times report said that Google CEO Sundar Pichai declared “code red” and accelerated AI development to be able to unveil at least 20 AI-powered products this year. Now, CNBC has detailed several of the products the company is working on, including a chatbot called “Apprentice Bard” that uses Google’s LaMDA conversation technology.
Apparently, Google management asked the LaMDA team to prioritize working on a ChatGPT competitor, telling them that it takes precedence over any other project and even warning them not to attend unrelated meetings. Apprentice Bard reportedly looks and functions like ChatGPT in that a user can type a question or a prompt in a text box and then get a written response.
CNBC says it saw samples proving the bot’s answers include information from recent events — something ChatGPT isn’t capable of, because it only has limited knowledge about anything that happened after 2021. In one example, Apprentice Bard was able to answer whether there will be another round of layoffs at Google. (Unlikely for this year, it said, seeing as the company is doing well financially.) If you’ll recall, Google’s LaMDA tech had gotten a former company engineer fired after he claimed that it had gained sentience.
Google is also reportedly testing a new search page that uses a question and answer format. The experimental home page features five potential question prompts replacing “I’m feeling lucky” under the search bar. After a user types in their query, the page generates human-like responses in gray bubbles. Underneath those responses are suggested follow-up questions, followed by the typical search results with links and headlines. In addition, Alphabet is working on a project called “Atlas” under its cloud unit. While CNBC didn’t have details on what it is, it’s still reportedly part of Google’s “code red” efforts.
It’s unclear at this point which of Google’s projects are going to be released to the public. Google AI chief Jeff Dean told employees during an all-hands meeting to discuss the company’s response to ChatGPT that it’s moving “more conservatively than a small startup.” Providing people wrong information will have a much bigger impact for a known company like Google, after all. Indeed, The Times previously said that the tech giant is prioritizing safety, accuracy and blocking out misinformation when it comes to the development of its search chatbot. If reports that Microsoft is incorporating ChatGPT’s technology into Bing as soon as this March are true, however, we’ll also likely see Google’s search chatbot in the near future.
After a few years of staying mostly under the radar, Instagram co-founders Kevin Systrom and Mike Krieger are back with a new project. It’s an app called Artifact, a name Systrom told Platformer’s Casey Newton is designed to evoke the project’s three t…
If you’ve been thinking about adding Microsoft Word, Excel and Powerpoint to your home office, this may be a good day to sign up. Today only, Amazon is offering a 12-month subscription to Microsoft 365 Family bundled with a $50 Amazon gift card for $100. The price of the yearly subscription is usually $100, so you’re essentially getting a free gift card for signing up. The deal is only good through the end of the day and the subscription will auto-renew at the end of the year, so set a reminder if you want to cancel before that kicks in.
The year-long subscription will come as a digital download and runs on PCs or Macs as well as smartphones and tablets running Apple’s iOS or Android OS. It covers up to six people and can run on five devices at the same time, with offline access included. Microsoft 365 is the new branding for the well-known Office software and the Family subscription includes access to Word, Excel, Outlook email and PowerPoint. It also comes with the Microsoft Teams video calling app, which updated last year to include a Live Share feature enabling easier real-time collaboration. You also get Clipchamp video editor, which Microsoft acquired in 2021. Included security add-ons like Microsoft Defender and ransomeware protection will help protect your data and devices.
Each person gets 1TB of OneDrive personal cloud storage, which not only lets you store a large amount of files in the cloud, it lets you share photos and files with other OneDrive users, across pretty much any device.
The Microsoft 365 family subscription is activated as a digital download, but the $50 gift card will arrive as a physical card in the mail. And as we said, the deal ends today, so take advantage if you want to add Microsoft’s productivity apps to your setup.
Ever found yourself turning down the radio so you can focus on finding a parking spot? Music didn’t stop you seeing, but it was taking up some tangible mental resources. But what if you had a way to immediately make the music more calming? Or to change…
Companies continue to find new ways to impress with true wireless earbuds. There’s no doubt the popularity of Apple’s AirPods helped make them a mainstay, but plenty of others offer reliable connectivity, great sound and active noise cancellation (ANC)…
Companies continue to find new ways to impress with true wireless earbuds. There’s no doubt the popularity of Apple’s AirPods helped make them a mainstay, but plenty of others offer reliable connectivity, great sound and active noise cancellation (ANC)…
When Apple debuted the HomePod in 2018, it was already late to the smart speaker game. Sure, the company has never been worried about tardiness, choosing instead to focus on being the best. Unfortunately, that wasn’t the case with its first attempt. The HomePod sounded good, but lacked a lot of the basic functionality and voice assistant smarts the likes of Amazon and Google offered. Apple has spent the last five years improving both HomePod and HomePod mini, adding features like multi-room audio, multi-user support and an intercom tool.
In 2021, Apple discontinued the original model in favor of the $99 HomePod mini. But now, the larger version is back with a familiar look, but lots of changes on the inside. Like the first, the new HomePod ($299) is best suited for those who’ve committed to Apple’s ecosystem. So if that’s not you, I understand if you don’t want to go any further. If it is, just know that the company has given its speaker a lot more tools than it had at launch five years ago, including more capable assistance from Siri and more smart home abilities. Plus, the second-generation HomePod is $50 cheaper than the original was at launch.
A familiar design updated on the inside
If you were hoping for a wholesale redesign with the new HomePod, Apple undoubtedly disappointed you. However, there are some noticeable changes upon close inspection. First, the speaker’s touch panel is now slightly recessed like the HomePod mini. On the original version, that panel sits flush with the top rim. When you trigger Siri, lighting for that panel now goes all the way to the edge too. Next, the power cord is now detachable. This means if you have an issue with that very necessary component, it should be easier to get a replacement (via Apple Care). If you have excellent vision, you might also be able to discern the 2023 model is 0.2-inches shorter at 6.6-inches tall (vs. 6.8 inches) if they’re sitting side by side.
On the inside though, Apple made a host of changes. Both versions have an upward facing woofer with a group of tweeters around the bottom. For the second-generation, Apple reduced the number of tweeters from seven to five, angling them slightly upward where previously they were almost perfectly side-firing. The company also cut the number of voice microphones from six down to four. And perhaps the most important change internally, Apple swapped the iPhone 6’s A8 chip for the much more modern S7 – the silicon that powers the Apple Watch series 7.
Software and setup
Despite the lack of changes on the outside, the second-gen HomePod is a better speaker because Apple has been improving it and the HomePod mini over time. The company added stereo pairing and multi-room audio alongside AirPlay 2 a few months after launch of the first generation model. Multi-user support and audio handoffs were added in 2019 and the intercom feature arrived ahead of the HomePod mini in 2020. And in 2021, HomePod gained the ability to play any TV audio via an Apple TV 4K over eARC and both spatial audio and Apple Music lossless streaming. So many of the key features Apple is chatting up on the new HomePod are things it has been slowly adding since 2018. And many of which, I’d argue, should’ve been there from there from the start.
Everything you need to set up and control the HomePod is found in Apple’s Home app. First, you’ll need to add a new speaker like you would any other smart home device with this software, including assigning it a room or location in your home. The app allows you to set up Automations and Scenes along with configuring how you’d like to interact with Siri.
Billy Steele/Engadget
You have the ability to disable voice cues or the long press on the HomePod’s touch panel in order to activate the assistant (both are enabled by default). You can also have the speaker light up and play a sound when using Siri – or do one or neither. The Home app will let you use Siri for Personal Requests too, where the speaker can recognize your voice for things like messages, calls and reminders so long as your iPhone is close by. The software allows you to disable the HomePod’s intercom feature as well, should you prefer not to use that tool.
Like the first HomePod, the 2023 model is equipped with room calibration. This detects reflections off of walls and other surfaces to configure the sound for the best performance. The HomePod does this the first time you play music, but it isn’t constant monitoring. Instead, like it did on the previous generation, the setup leverages an accelerometer to determine when it has been moved. If it has, the speaker will then reinitiate the room sensing process when music is played.
A smart home with Siri
Billy Steele/Engadget
When we reviewed the original HomePod in 2018, one of our biggest gripes was with Siri’s limited abilities. Sure the speaker sounded good, but the lack of polish with the voice assistant made it seem like a work in progress. Apple has done a lot to improve Siri over the last five years, so a lot of those issues with the original have been fixed.
First, the HomePod, like Siri on your iPhone, is capable of recognizing multiple users. Personal Requests can allow it to peek at your Calendar, Notes, Reminders, Messages, Find My and more when you ask. Plus, HomePod can give each member of your family (up to six people) their unique responses from certain iPhone apps. What’s more, Siri can create recurring home automations without you having to pick up your phone and swipe over to the appropriate app.
Even with fewer microphones to pick up your voice, the new HomePod doesn’t suffer any performance setbacks. It’s just as capable as ever at picking out your voice even in a noisy room. And, while playing music with voice commands used to be limited to Apple Music, services like Pandora, Deezer and several others can now be controlled via Siri.
HomePod is also a smart home hub, easily pairing with HomeKit and Matter accessories. The new model is equipped with temperature and humidity sensing, and that info is displayed prominently in the Home app. Of course, you can use that data to create automations with other smart home devices. I don’t have a compatible smart thermostat, but I was able to successfully trigger a smart plug once the HomePod detected a certain temperature in my living room.
Sound Recognition is arguably the most notable new feature that Apple is debuting with the new HomePod, but it’s not ready yet. It can listen for smoke or carbon monoxide alarms and send a notification to an iPhone, iPad or Apple Watch. What’s more, you’ll be able to “check in” on what’s happening via an audio feed or camera. Sound Recognition won’t arrive until later this spring, so we’ll have to wait to test it.
Sound quality
Billy Steele/Engadget
Audio quality wasn’t an issue with the first-gen HomePod and it’s still great here. But like the first version, Apple’s choice for tuning won’t appease everyone. There’s a continued emphasis on voice, so things like vocals in music or dialog on a TV show or movie take center stage. At times it’s fine though, with some genres and content, it leads to a rather subpar experience.
Don’t get me wrong, for the most part HomePod sounds outstanding, especially when you put it up against other smart speakers. The fact is many of those don’t sound very good at all, so Apple continues its track record for making a device that has serious audio chops and smart features. RTJ4 has ample bass for its bombastic hip-hop beats while Sylvan Esso’s No Rules Sandy gets enough low-end for its synth-heavy rhythms. Thanks to the HomePod’s excellent clarity, detailed styles like bluegrass and jazz shine. Sometimes the bass is a little too subdued for metal (Underoath’s Voyuerist) or full-band country (Zach Bryan’s 2022 live album), and the vocals too forward, but overall, it sets the standard for smart speaker sound. And the HomePod only shines brighter as a stereo pair.
When you add the second HomePod in the Home app, the software asks you if you want to use them together. Once you tell it which side the additional unit is on (left or right), the app completes the setup for you, assigning the appropriate channel to each speaker. Now everywhere you previously saw a single HomePod icon, you’ll see two, constantly reminding you of the stereo arrangement. The double speaker option is great for music. On Béla Fleck’s My Bluegrass Heart, the two-speaker arrangement adds a lot of dimensionality to each track. There was already a spatial element to the tunes, but dual HomePods heightens the sensation, making it seem like you’re in the middle of the musicians while they record each song.
Billy Steele/Engadget
When you select a HomePod for use with the Apple TV you still need to deselect your TV speakers. Those aren’t disabled just because you’ve asked the streaming box to also send audio to a HomePod or two. TV audio with stereo pair is fine, but I can’t see using a single unit for the same purpose. In a multi-room setup you might want to send the sound from a live event like the Super Bowl to a solo speaker, but having just one as your lone living room audio from a TV isn’t a great experience. While the HomePod beams sound in all directions, with a single speaker it’s clear the sound is coming from a fixed location – something that’s a lot less noticeable with a pair or with a soundbar.
As is the case with vocals, dialog takes prominence with a HomePod and Apple TV. At times it can seem slightly muffled when watching things like live sports. Otherwise, there’s good clarity, nice bass and great dimensional audio when streaming Formula 1: Drive to Survive or Slow Horses. But, honestly, if you’re looking to improve your living room audio, a soundbar and sub is a better option. The driver arrangement in those speakers does a better job of filling a room completely and evenly. Plus, most companies give you the ability to use adjustments or presets to dial in how much you want those speakers to focus on dialog so you can fine-tune things to your liking.
The competition
Billy Steele/Engadget
Because the HomePod is best suited to people who have married themselves to Apple’s ecosystem, the best alternative to the second-gen HomePod is the cheaper HomePod mini. If all you’re after is some Siri assistance with your smart home and a speaker that’s good enough for casual listening and podcasts, the $99 option will work well for you. Plus, Apple just unlocked the smaller speaker’s inactive temperature and humidity sensor and it’s due to get Sound Recognition.
Wrap-up
Apple has been preparing for a new HomePod for five years, constantly improving both the original version and the HomePod mini. And the fact that the company has made Siri a more capable companion certainly helps. Plus, there’s more smart home abilities than before. Apple hasn’t strayed from its emphasis on the spoken (or sung) word for HomePod’s sound profile, but that’s okay. It’s clear that the company is focused on expanding the toolbox for its smart speakers after their debut, so I’d expect that much like the original HomePod, this is just the beginning for the second generation.
The CEO of TikTok, Shou Zi Chew, will testify before the House Energy and Commerce Committee on March 23rd. Chew will discuss the app’s privacy and data security measures, its impact on kids and ties to China (where parent company ByteDance is headquartered). This is Chew’s first appearance in front of a congressional panel. TikTok’s security and relationship with Chinese authorities have drawn the attention of US officials over the last few years. However, as CNBC reported, discussions between the US and TikTok appear to have stalled.
The relationship has been a precarious one for several years. The company has tried to assuage concerns by storing US user data on domestic Oracle servers and deleting such data from its own servers in the US and Singapore. However, other stories, like TikTok having to fire four employees (two each in China and the US) who accessed the data of several journalists, aren’t helping. They were said to be looking for the sources of leaks to reporters.
Also in December, a mammoth spending bill passed by lawmakers included a ban of TikTok from federal government-owned devices. However, the relationship has thawed since its nadir during the Trump administration, when the president pushed for the social network to shed its US assets. That, of course, never happened.
Samsung has revealed a sharp decline in profit for 2022, mainly due to the weak demand for its chips and smartphones, the company’s main moneymakers. The Korean tech giant has posted KRW 302.23 trillion (US$245.4 billion) in annual revenue, which is a new record high for the company, in its latest earnings report. Its operating profit, however, was down KRW 8.5 trillion (US$6.9 billion) from the year before. Sales for Samsung’s more affordable phones went down, and while flagship sales held up to market expectations, they’re still lower than in previous quarters. The company expects demand for mass-market smartphones to weaken even further in 2023 “due to persistent macroeconomic conditions.” That’s not great news when the company is holding its first Unpacked event of 2023 tomorrow.
Carl Pei says the upcoming flagship will be “more premium” than the Phone 1.
Engadget
Nothing’s Carl Pei has confirmed the company’s second phone will launch in the US later this year. He described the 2023 flagship as “more premium” than the Nothing Phone 1, which probably means it’ll be more expensive, too. In an interview with Inverse, Pei said the Phone 2’s US launch would be Nothing’s top priority this year. Pei added that American carrier demands (“red tape”) were the main reason Nothing didn’t launch the Phone 1 in the US.
Baidu’s service will allow users to get conversation-like search results.
NurPhoto via Getty Images
Chinese search giant Baidu aims to introduce a ChatGPT-like AI service that gives users conversational results, according to a Bloomberg report. Open AI’s ChatGPT has taken the tech world by storm, thanks to its ability to answer fact-based questions, write in a human-like way and even create code. Microsoft invested $1 billion in Open AI back in 2019, and reportedly plans to incorporate aspects of ChatGPT into its Bing search engine. It’s the big thing – even Google reportedly sees the technology as a threat to its search business and plans to accelerate the development of its own conversational AI technology.
It’s a good time to pick up one of Apple’s latest M1-equipped iPad Air models, as they’ve dropped back down to all-time low prices. The 64GB WiFi model is now on sale for just $500, or $99 (17 percent) off, while the 256GB is available for $650, also $…
A few days ago, speech AI startup ElevenLabs launched a beta version of its platform that gives users the power to create entirely new synthetic voices for text-to-speech audio or to clone somebody’s voice. Well, it only took the internet a few days to start using the latter for vile purposes. The company has revealed on Twitter that it’s seeing an “increasing number of voice cloning misuse cases” and that it’s thinking of a way to address the problem by “implementing additional safeguards.”
While ElevenLabs didn’t elaborate on what it meant by “misuse cases,” Motherboard found 4chan posts with clips featuring generated voices that sound like celebrities reading or saying something questionable. One clip, for instance, reportedly featured a voice that sounded like Emma Watson reading a part of Mein Kampf. Users also posted voice clips that feature homophobic, transphobic, violent and racist sentiments. It’s not entirely clear if all the clips used ElevenLab’s technology, but a post with a wide collection of the voice files on 4chan included a link to the startup’s platform.
Perhaps this emergence of “deepfake” audio clips shouldn’t come as a surprise, seeing as a few years ago, we’d seen a similar phenomenon take place. Advances in AI and machine learning had led to a rise in deepfake videos, specifically deepfake pornography, wherein existing pornographic materials are altered to use the faces of celebrities. And, yes, people used Emma Watson’s face for some of those videos.
ElevenLabs is now gathering feedback on how to prevent users from abusing its technology. At the moment, its current ideas include adding more layers to its account verification to enable voice cloning, such as requiring users to enter payment info or an ID. It’s also considering having users verify copyright ownership of the voice they want to clone, such as getting them to submit a sample with prompted text. Finally, the company is thinking of dropping its Voice Lab tool altogether and having users submit voice cloning requests that it has to manually verify.
Crazy weekend – thank you to everyone for trying out our Beta platform. While we see our tech being overwhelmingly applied to positive use, we also see an increasing number of voice cloning misuse cases. We want to reach out to Twitter community for thoughts and feedback!
