As it promised last week, Samsung has launched Food, a "personalized, AI-powered food and recipe" app in eight languages and 104 countries around the world. It draws on the food database of Whisk, an app Samsung acquired a few years back — and resembles a version of Whisk the company revealed last year. Given Samsung's large presence in kitchens with its smart fridges and other appliances, the release of a food and recipe app seems a logical step for the company.

The app allows users to search for recipes around the world, save them and make weekly eating plans. The company prepared over 160,000 recipes for launch, with that number set to increase down the road. Samsung Food can also be run on mobile phones and Samsung Family Hub smart appliances like refrigerators, while allowing users to manage ingredients, shopping, etc.

Samsung

Users can save recipes anytime, and the app analyzes them, standardizes the format and organizes them to create shopping lists based on the ingredients. It can also provide recipe recommendations based on available food items as managed by the user. It even has a "personalize recipe" function that uses the AI to alter recipes recipes and create vegan or vegetarian versions, for instance. "Users will even be able to create fusion recipes, such as Korean versions of Italian dishes, and adjust cook time or skill level of recipes," Samsung adds.

It uses AI to create recommendations for individualized daily meal plans based on dietary preferences and favorite cuisine types. Nutritional ingredient breakdowns can be viewed at any time, and users can add items to shopping lists and then send them straight to a retailer's e-commerce checkout. With connected cooking, it lets users preheat ovens, set timers and transfer cook settings to supported appliances via a step-by-step guided cooking mode. 

Last week, Samsung said it hoped to secure a million users for the app around the world. While there are numerous recipe apps out there (Mealtime, Paprika, Yummly etc.), Samsung may have an edge with the millions of its smart appliances sold — making it a known quantity to consumers. 

Samsung plans to add new features, like integration with Samsung Health to sync parameters like BMI and calorie consumption, while offering suggestions for diet management. The app will incorporate AI vision tech by 2024, allowing Samsung Food to recognize food items through the camera and provide details like nutrition information. Samsung Food is now available for download on Android and iOS. 

Google took a step towards transparency in AI-generated images today. Google DeepMind announced SynthID, a watermarking / identification tool for generative art. The company says the technology embeds a digital watermark, invisible to the human eye, directly onto an image’s pixels. SynthID is rolling out first to “a limited number” of customers using Imagen, Google’s art generator available on its suite of cloud-based AI tools.

One of the many issues with generative art — apart from the ethical implications of training on artists’ work — is the potential for creating deepfakes. For example, the pope’s hot new hip-hop attire (an AI image created with MidJourney) going viral on social media was an early example of what could become more commonplace as generative tools evolve. It doesn’t take much imagination to see how something like political ads using AI-generated art could do much more damage than a funny image circulating on Twitter. “Watermarking audio and visual content to help make it clear that content is AI-generated” was one of the voluntary commitments that seven AI companies agreed to develop after a July meeting at the White House. Google is the first of the companies to launch such a system.

Google doesn’t go too far into the weeds about SynthID’s technical implementation (likely to prevent workarounds), but it says the watermark can’t be easily removed through simple editing techniques. “Finding the right balance between imperceptibility and robustness to image manipulations is difficult,” the company wrote in a DeepMind blog post published today. “We designed SynthID so it doesn’t compromise image quality, and allows the watermark to remain detectable, even after modifications like adding filters, changing colours, and saving with various lossy compression schemes — most commonly used for JPEGs,” DeepMind’s SynthID project leaders Sven Gowal and Pushmeet Kohli wrote.

Google DeepMind

The identification part of SynthID rates the image based on three digital watermark confidence levels: detected, not detected and possibly detected. Since the tool is embedded into the image’s pixels, Google says its system can work alongside metadata-based approaches, like the one Adobe uses with its Photoshop generative features, currently available in an open beta.

SynthID includes a pair of deep learning models: one for watermarking and the other for identifying. Google says the two trained on diverse images, culminating in a combined ML model. “The combined model is optimised on a range of objectives, including correctly identifying watermarked content and improving imperceptibility by visually aligning the watermark to the original content,” Gowal and Kohli wrote.

Google acknowledged that it isn’t a perfect solution, adding that it “isn’t foolproof against extreme image manipulations.” But it describes the watermark as “a promising technical approach for empowering people and organisations to work with AI-generated content responsibly.” The company says the tool could expand to other AI models, including those tasked with generating text (like ChatGPT), video and audio. 

Although watermarks could help with deepfakes, it’s easy to imagine digital watermarking turning into an arms race with hackers, with services that adopt SynthID requiring continual updating. In addition, the open-source nature of Stable Diffusion, one of the leading generative tools, could make industry-wide adoption of SynthID or any similar solution a tall order: It already has countless custom builds that can run on local PCs out in the wild. Regardless, Google hopes to make SynthID available to third parties “in the near future” to at least improve AI transparency industry-wide. 

Back at I/O, we got a sense of some of Google's big plans for generative artificial intelligence, including the integration of the Duet AI virtual assistant into Workspace productivity apps. The company is now starting to roll out Duet AI to those who pay for versions of Docs, Gmail, Sheets and so on that are geared toward professional use. The virtual assistant is also making its way to Google Meet and Google Chat.

There's a free trial of Duet AI for Google Workspace. However, to use the virtual assistant in the cloud-based apps after that, large organizations will need to pay $30 per month per user, according to The Verge. Microsoft charges the same for 365 AI Copilot in Office apps. Duet AI pricing for smaller organizations hasn't been finalized yet.

Duet AI has a lot of tools in its belt that are designed to lighten your workload. You can use it to generate an entire email draft in Gmail. It can pore over emails and a swathe of documents and spreadsheets to create a summary of recent work activities (such as your company's quarterly performance). The assistant can then build a presentation for you based on its findings — but definitely review the slides to make sure all the facts and figures are accurate before showing them to your bosses. And, of course, Duet AI can generate images as well.

The virtual assistant will be able to help you out on other fronts too, including in Google Meet and Chat. It has somewhat different functions in Meet, since it can help you "look and sound your best with studio look, studio lighting and studio sound," Google Workspace general manager and vice president Aparna Pappu wrote in a blog post. Duet AI can use face detection and dynamic tiles to give everyone in a meeting room their own video tile, perhaps with the idea of making all the participants feel more engaged and connected.

Meet will be able to offer translated captions in 18 languages. Google says the app can tell when someone is speaking in another language and display translated text in real time. (For comparison, Zoom has offered live transcription to free users for nearly two years.)

Duet AI's note-taking functions should prove useful too. A "take notes for me" feature will prompt the assistant to jot down action items and capture video snippets, then send a summary to attendees after the fact. If you're unable to join a meeting right at the beginning (or you slept in, we're not judging), Duet AI can catch you up on what you missed with an in-progress summary.

Google

If you can't make it to a meeting at all, you can send Duet AI on your behalf. The assistant can get your points across to other attendees and still make sure you get a summary afterwards. However, if all participants use Duet AI as their proxy, Meet will seemingly be able to detect that and end the meeting.

On top of all of that, the company is stuffing Duet AI into Google Chat too. Google is trying to make the app more of a competitor to Slack and Microsoft Teams, which could make a lot of sense for organizations that are already paying for Workspace for other productivity apps.

You'll be able to ask Duet AI questions about topics of discussion. The assistant will be able to get you up to speed on conversations you missed and summarize documents that others have shared. Google has refreshed Chat's user interface and added more shortcuts and an enhanced search function that includes suggested queries and autocomplete. Autocorrect could help you avoid typos in your messages too. Chat communities will also eventually be able to include up to 500,000 users, while those who need to talk something out using their voices will be able to hop into a Meet-powered huddle.

As for privacy, Google says that users' interactions with Duet AI will be private to them. "No other user will see your data and Google does not use your data to train our models without your permission," Pappu wrote. "Building on these core commitments, we recently announced new capabilities to help prevent cyber threats, provide safer work with built-in zero trust controls, and better support our customers’ digital sovereignty and compliance needs."

Generative AI tools have had a largely fraught relationship with privacy so far, with some companies reportedly restricting workers' ChatGPT use over concerns that they may accidentally leak proprietary data. Italy also briefly banned the chatbot due to privacy worries. Just this week, ChatGPT owner OpenAI debuted an Enterprise offering. The company says it won't train its AI models on Enterprise subscribers' usage of the service — a similar claim to the one Google just made with Duet AI.

Google will surely see Duet AI as a competitor to ChatGPT and Microsoft Copilot. It's letting third-party companies use the underlying tech for other apps, such as an upcoming GE Appliances app that can generate recipes based on the food in a fridge (it's hard to think of an AI-powered app that would improve my life more than this). Meanwhile, as The Wall Street Journal notes, Google will make its PaLM 2 language learning model more broadly available after only granting select customers access until now.

Google (and Alphabet) CEO Sundar Pichai told the publication that the company will offer Gemini — its direct answer to ChatGPT — to cloud customers when it's ready fro primetime. Google is also scaling up its infrastructure to handle the demands of AI functionality, but Pichai admitted that the company doesn't have a crystal ball in that regard. “It’s tough to fully predict the demand,” he told the Journal. “I think we will be OK, but does it keep me up at night? Yes.”

Google has been letting Android users set up locked folders in the Photos app for quite some time, giving them the option to keep sensitive images and screenshots safe from prying eyes. Now, the tech giant has started rolling out an update for the app that allows users to back up their locked folders, so they can access files saved in them from any of their devices. In addition, users can now access and set up locked folders — a feature that started as a Pixel exclusive — from the Photos app on iOS and the web. That means users don't even need an Android device to be able to enjoy the feature anymore. 

As a refresher, users can set up a locked folder by going into the Utilities section of the Photos app. They'll then see "cards" for potentially more than one feature, where they can directly create a locked folder for any image they want to keep safe. After they're done, they can choose whether to back up the folder or not. They can also easily add more images from their gallery by long pressing on one and then ticking the checkboxes that appear on their upper left corner. All images sent to locked folders will still be inaccessible anywhere else, and users can still only access them after confirming their identity using their device's password or biometric authenticaiton.

Google

Google has also started rolling out a simplified settings page for the Photos app. When users click on their account icon and check the settings page, they'll now see a new layout that makes it easier for them to find what they're looking for by navigating the new privacy, backup, sharing and notifications sections. 

Snapchat has added a new generative AI feature to its app. Called “Dreams,” it’s in some ways similar to the company’s signature AR effects, known as lenses. But instead of real-time camera-based effects, the feature uses generative AI to remix users’ selfies into “fantastical images that transform their persona into new identities.”

The feature, which can be found in the app’s “memories” section, begins by asking users to take selfies showing their face at different angles. The app will then creates a series of eight images based on themes like “time travel” or “alternate universes.” Eventually, Snap says, users will be able to create Dreams that include their friends’ likenesses as well.

Dreams is the latest generative AI experiment from the company, which launched its MyAI chatbot earlier this year using OpenAI’s models. (Dreams uses open source tools and internal data, though the company hasn’t provided details about specific partners.)

The feature also highlights how the company is using interest in the technology as a source of revenue. MyAI was initially limited to Snapchat+, the app’s premium subscription tier, before it was released to all the app’s users this spring. The company has since added specialized features for subscribers, including the ability for MyAI to reply to photo Snaps with its own AI-generated images.

Likewise, Dreams will have both a free and paid component. Snap is allowing non-Snapchat+ subscribers to access just one — so use it wisely — “pack” of eight selfies, while subscribers will get access to one pack a month (the company says it plans to update Dreams with new themes and styles regularly). All users will be able to buy additional packs for a $0.99 in-app purchase.

Snap

In practice, the images appear to have some of the same limitations as other AI-based image generators. A promotional image shared by Snap showed what appeared to be the tips of partial fingers strangely placed over the subject's chest. When I tried Dreams to create my own AI selfies, some of the resulting images also had strange-looking hands, though it at least showed the correct number of fingers placed in an anatomically correct position.

Still, I can see how the feature could keep Snapchat users — who have collectively sent more than 10 billion messages to MyAI — coming back. And as tools like Midjourney have moved behind paywalls, Snap’s offerings might just seem like a better deal for those looking to try out generative AI.

During its I/O annual developer conference earlier this year, Google said it was giving users the ability to transition their first-generation Nest Cam Indoor and Nest Cam Outdoor devices from the old Nest app to its new Home application. Well, the time has come — for the indoor camera, at least. Google has just started rolling out the ability to manage the camera model through the public preview version of its Home app, and it will take a few weeks before it's done making its way to all users. 

Nest cameras used to have their own application before the company was acquired by Google. Over time, Google added support for all its smart home products to its Home app, but users of the the oldest Nest cameras had to stick to the original application to be able to manage their devices. With this update, and as long as the user is on public preview, people will be able to access their first-gen Nest Cam Indoor device through the Home app and enjoy its updated camera history experience. They'll also be able to add the camera to their Favorites tab, so that its feed shows up as soon as they open the app. 

When Google first announced the ability to add the old Nest cam models to the Home app, it said the transfer would take users around 10 minutes. After that, they'll no longer need the old application to see the camera's data and footage, enabling a single app experience for a lot of users. 

The long-awaited web version of Threads has finally arrived. After announcing that its Twitter competitor would be accessible from browsers earlier this week, Meta’s rollout is finally underway. Instagram head Adam Mosseri confirmed Thursday the update is now live for all users.

That means if you’ve been impatiently refreshing threads.net in waiting for it to be available, you should check again. Up to now, many Threads users have been frustrated by the lack of a web version. Though it was possible to manually navigate to a specific user’s profile, people have been unable to browse their feeds, post, reply or even log in to their accounts from browsers. 

As the app has seen a significant drop-off in engagement following its initial launch, many have viewed the lack of a web version as a significant barrier to the app. Both Mark Zuckerberg and Mosseri have been promising a web version was high on their agenda for Threads, though Mosseri recently said the team was working out some bugs with the experience. 

Now that the web version is finally live, it will be interesting to see which of the many “missing” features Threads will get next. Users have also complained about the lack of content search — right now people can only search for users, not for specific posts — and hashtags, both of which help people discover new content on Instagram and Twitter.

One feature, however, that seems unlikely is the addition of DMs. Mosseri has said he’s reluctant to add another inbox to user’s lives, though he recently raised the possibility of incorporating Instagram DMs into the Threads experience. But he did confirm that post-editing features are in the works, so at least Threads users won’t have to wait more than a decade for an edit button.

If you've been wondering who's making the most money from the AI boom, NVIDIA may have the answer in it's latest earnings report. The company announced revenue of $13.51 billion in the second quarter, more than doubling the $6.7 billion it made last year and crushing market expectations. On top of that, it earned $6.18 billion in GAAP net income, nine times the $656 million it made in Q2 2022. 

NVIDIA's gaming segment did pretty well too, thanks to $2.49 billion in Q2 revenue, up 22 percent from last year. During the quarter, it started shipping the budget-oriented GeForce RTX 4060 GPU, announced the Avatar Cloud Engine (ACE) for games and saw the addition of 35 DLSS games including Diablo IV. (Earlier this week, it unveiled DLSS 3.5 designed to use AI to make ray-traced games look better.)

But it was very much the AI and data center segments that pushed NVIDIA to new heights. It saw a record $10.32 billion in revenue in that sector alone, up 141 percent from Q1 2023 and 171 percent from a year ago. 

Earlier this year, CEO Jensen Huang said that back in 2018, NVIDIA had a "bet the company" moment when it started using AI to power DLSS, "and while we were reinventing CG with AI, we were reinventing the GPU for AI." He later added that "the future is a large language model (LLM) at the front of just about everything," from VFX to heavy industry. 

NVIDIA's prescience is now paying off with the company's flagship H100 Tensor Core GPU. It's also been building more complex systems like the HGX box, which puts eight H100 GPUs into a single computer. All of that helped it create immense cashflows with top customers spending heavily on NVIDIA GPU tech to build complex AI models — like Microsoft with its Azure segment. 

In addition, the company's use of custom software and apps makes it difficult for customers to switch rivals like AMD. "Our Data Center products include a significant amount of software and complexity which is also helping for gross margins," said NVIDIA finance chief Colette Kress in an analyst call.

All that led to a perfect storm of profit. "During the quarter, major cloud service providers announced massive NVIDIA H100 AI infrastructures. Leading enterprise IT system and software providers announced partnerships to bring NVIDIA AI to every industry. The race is on to adopt generative AI," Huang said in a statement. "Companies worldwide are transitioning from general-purpose to accelerated computing and generative AI." The company expects more to come, forecasting around $16 billion in revenue for Q3. 

You now have an easier way to control your Tesla from your iPhone. Not a Tesla Appnotes the latest version of Tesla's iOS software now supports Shortcuts actions, making some common tasks available through Siri in addition to on-screen taps and widgets. You can use Apple's voice assistant for simple tasks like opening the trunks or windows through to special modes. If you've ever wanted to activate Bioweapon Defense Mode by talking to your phone, it's now an option.

Shortcuts also enable automations, so you can string together multiple tasks or perform those tasks on set schedules. You can create a panic mode that flashes lights, honks the horn and closes windows all at the same time, for instance, or automatically warm the cabin before your morning commute.

The available commands aren't comprehensive, and include features you probably aren't going to use – you don't really need a shortcut for an emissions test. This could still be significantly more convenient than wading through the Tesla app, however. It's also notable that there isn't any official integration with Alexa or Google Assistant as of this writing, so Android users will have to put in more work to go hands-free.

This app update doesn't give Tesla cars a built-in voice control system. You'll need to turn to cars from the likes of Mercedes or Volvo if you want to talk directly to your vehicle. It does bring voice into the equation, though, and may be appreciated if you'd rather not wade through the official app to handle common functions.

Threads by Instagram will get a web version as soon as this week, people familiar with the matter told The Wall Street Journal. Earlier this month, Meta CEO Mark Zuckerberg promised a web version with better search functionality, and Instagram head Adam Mosseri recently said that one is in testing. Currently, a full version of Threads is only available on iOS and Android, with limited read-only functionality on browsers.

A web version is near the top of the list of most-desired features for Threads, but the company is exercising caution with the release. "It’s a little bit buggy right now, you don’t want it just yet," Mosseri said Friday on Instagram. "As soon as it is ready we will share it with everybody else."

Threads recently added new features to Threads like the ability to set notifications and view posts in chronological order. The company also started labeling state-controlled media outlets after some were seen posting propaganda. Another new update is the "repost" tab makes it easier to see all reposted content. (X, previously called Twitter, recently renamed "retweets" to the more generic "reposts," ironically following Threads' lead.) 

A web version would be coming at a good time for Threads. After a torrid launch with over 100 million users signing on in the first week, the number of daily active users (DAUs) dropped down to 80 percent by mid-August. Still, Threads is by far the most successful alternative to X, which counted around 238 million DAUs in August 2023 and 364 million monthly active users, X reported last year. 

In any case, the launch of a web version will be particularly useful for social media power users, just when Twitter has put one of its key tools for those folks, Tweetdeck, permanently behind a paywall.