Google transforms shopping with virtual dressing rooms, automates advertising, enhances navigation, empowers health with skin condition search, and revolutionizes travel and product searches.
Google’s new generative AI lets you preview clothes on different models
Google’s having a new stab at virtual dress-up, introducing an AI-powered tool for Google Shopping. This tech wizardry predicts how clothes will look on an assortment of real models, mimicking how the garment would hang, stretch, or even crinkle. They’ve trained this AI by showing it countless pairs of pictures featuring people striking different poses in the same outfit. The aim? Making online clothes shopping less of a wild card.
Now folks in the U.S. can take a virtual spin in women’s tops from brands like Anthropologie and H&M. Just look for the “Try On” badge on Google Search. Men will have to wait a little longer though – their tops are slated for a later launch.
The folks at Google, in a statement from Senior Director of Consumer Shopping Product, Lilian Rincon, make it clear they believe shopping for clothes online should instill the same confidence as buying in store. Despite similar tech attempts by Amazon, Walmart, and a few start-ups, this confidence seems elusive, with surveys revealing a sizable chunk of online shoppers feel short-changed because their purchase doesn’t look as expected when it arrives.
Alongside the virtual try-on, Google’s adding AI-powered filters for clothing searches, helping users narrow down their hunt with inputs like color or style. Just like a personal shopping assistant, but without the small talk.
Google launches AI-powered advertiser features in push for automation
The tech giant just let slip about two shiny, new artificial intelligence (AI) features for advertisers. The big idea? Let the smart machines find the best spots for your ads across Google’s buffet of services. Google’s dialing it up a notch for the advertising folk. The new tech’s all about helping brands hit the bullseye with their ads.
One of these gizmos, “Demand Gen,” is poised to plunk down advertisers’ photo and video ads across a whole heap of Google’s services. Think Gmail, YouTube feed and Shorts – Google’s own answer to the TikTok frenzy.
The second doodad uses AI to rustle up the best ad spots aimed at boosting views of a brand’s video ads. The result? About 40% more eyeballs on those ads, as per Google’s own VP, Vidhya Srinivasan.
Google Maps gains glanceable directions and an expanded Immersive View
Google Maps has unveiled a spiffy set of features to make your trip more enjoyable – and maybe save you a few wrong turns along the way.
First up, “glanceable directions.” This gem, announced by Google back in February, is finally making its way to Android and iOS users. It’ll let you keep an eye on your travel progress, whether you’re hoofing it on foot, pedaling a bike, or driving, all from your lock screen or route overview. You’ll also get reminders about where to turn next, because, let’s face it, sometimes we all need a nudge in the right direction.
Next, the folks at Google have spiced up the Recents feature on your desktop. This handy shortcut now keeps a list of places you’ve looked up, ready for a quick peek on the side of your Maps screen. Soon, it’ll even remember places you’ve looked up after you’ve closed the Maps window. How’s that for a memory boost?
And remember that Immersive View thing Google has going on? The one that stitches together pictures to create a 3D-like image of the world? It’s getting an upgrade. Over 500 new landmarks, including the Prague Castle and the Sydney Harbour Bridge, are being added to its repertoire.
Christina Tong, a bigwig at Google Maps, says you can now peek inside of restaurants and cafes, and even check the weather and crowd levels with the time slider in Immersive View. No more guessing if your favorite brunch spot will be swamped on Sunday morning. That’s one more feather in Google’s cap.
Google Lens can now search for skin conditions
Google’s giving their Lens app a facelift, letting it help identify skin conditions. Just snap a photo of that weird mole or rash, and Lens will do its thing, showing you similar-looking conditions. It’s not the full-on skin-diagnosing AI tool Google launched in 2021 – the one that couldn’t crack the US because of FDA approval hurdles. But hey, it might help you decide if that rash warrants a doctor visit or just a trip to the pharmacy.
Lens is also getting buddy-buddy with Bard, Google’s chatty AI bot. Bard will use Lens to understand pictures you send. Say you send a photo of shoes and ask Bard what they are – Bard, now Lens-educated, will shoot back with an answer.
Google intros new AI-powered travel and product search features
Google jazzed up their Search Generative Experience (SGE), a fancy name for their trial search lab. This update is for folks hunting for travel info and bargain deals. Let’s say you want to know if a restaurant can handle your rowdy party of ten. SGE will pull together info from all over the web, including reviews, snaps, and details from the owner. Shopping for Bluetooth speakers? It’ll help you consider all the options, pulling up descriptions, reviews, ratings, prices, images, you name it.
Ng says, “With our SGE overhaul, we’re aiming to give you handy hints, so you can cut the planning and enjoy your time.” If you need to do deep-dive research for your summer getaway, SGE has got your back.
Google’s also cooking up a ‘Add to Sheets’ feature. This handy-dandy tool lets you stick a search result right into a spreadsheet and share with anyone, be it your mom, your buddy, or your work squad.
A leaked document of Amazon’s ideas for using ChatGPT and AI at work lists 67 ways to take advantage of the ChatGPT boom
Amazon’s been quietly hustling, and now we’ve got a leaked memo that shows they’re looking for 67 big and small ways to ride the ChatGPT wave. From sprucing up their products to fine-tuning the work dance, Amazon’s going all in on AI.
It seems the higher-ups asked the worker bees to brainstorm how this chatbot tech could give them an edge. And boy, did they deliver! Some ideas are cooking already – think security tool and an answer-any-question engineering app for their cloud products. Neat, huh?
And that’s not all. The paper shows Amazon’s got plans to use this chatbot to auto-craft code and marketing stuff. Sales reps could use it to comb through financial reports, a kind of speedy shortcut to the gold, if you catch my drift.
They’re even mulling over having ChatGPT speed up PRFAQ, a major decision-making process at Amazon. Amazon’s legal eagles have warned the staff to be careful when using ChatGPT, especially with sensitive info. Seems they’re walking a tightrope between innovation and security, which sounds about right for a tech giant.
OctoML launches OctoAI, a self-optimizing compute service for AI
OctoML has rolled out OctoAI, a fresh, self-tuning service to help businesses use and tweak open-source AI models. Unlike its original platform that catered to ML engineers with the aim of optimizing and packaging models, this one’s meant for businesses who want to leave the nitty-gritty of ML infrastructure to someone else. Imagine a valet service, but for your AI.
OctoAI picks out the best hardware based on users’ priorities (like cost or speed). If you want to skimp on expenses or make your AI model work faster than a cheetah, this service will have you covered. They also choose the best environment to run models, such as Nvidia GPUs or AWS’s Inferentia machines. Folks who like to take the reins can still control how their models run. However, CEO Luis Ceze reckons most people would rather let OctoAI do the heavy lifting.
On top of this, OctoML will offer souped-up versions of popular models like Dolly 2 and Stable Diffusion, making these models run up to three times faster while slashing the cost by fivefold. Although they’ll still support old customers who just want to optimize their models, OctoML is all in on this new compute platform.
Vectara lands $28.5M to supercharge enterprise search
A trio of ex-Googlers, who once called the Googleplex their home, have put their noodles together to whip up an AI tool they’re calling Vectara. Think of it like a supercharged search engine for businesses. You feed it text files, it finds and dishes out what you’re looking for.
The tool allows users to ask questions about their company’s data, and like a diligent librarian, returns a summary with all the juicy parts highlighted, citing the source data set. Plus, once it’s done indexing your documents, it tosses the original files, so there’s no risk of your data ending up where it shouldn’t.
Vectara plans to use its new wealth to expand its team and reach out to Europe, Middle East, and Asia Pacific markets. Their goal is to ensure they beat other players in the space like Forethought and Coveo to the enterprise search finish line. With the cognitive search market set to balloon to $15.28 billion by 2023, the race is on. They’ll need to bring their A-game, but if they play it right, they could grab a hefty chunk of that sweet pie.
92% of programmers are using AI tools, says GitHub developer survey
GitHub survey says a whopping 92% of US-based developers are using AI tools for coding, at work and off the clock. Partnering with Wakefield Research, they found out that most developers (70%) think AI tools are doing them a solid – helping them meet performance standards with better code quality, quicker output, and fewer mess-ups at the production level. This ain’t just for fun or open-source stuff, mind you. It’s serious business.
The survey shows developers want to learn more, design solutions, get user feedback, and work on their communication skills. They see AI as a tool, not the end game. They want their performance judged based on how they deal with bugs and issues, rather than just the volume of code they produce.
AI Regulation Is Here. Almost.
The EU parliament just voted to move forward with draft legislation known as the AI Act. This act is primed to be the West’s first hearty dish of AI rules. The draft act slaps a ban on things like real-time, remote biometric surveillance in public spaces. It tells companies to knock off the habit of scraping the internet to make facial-recognition databases. Heck, it even asks for a stop sign on predictive policing systems, the ones that act like fortune tellers for future crimes.
The legislation also wants to keep an eye on how companies train their AI. It could require these companies to raise their hands when they use AI to generate content. Companies might also need to tweak their AI so it doesn’t spit out illegal stuff and disclose the copyrighted data used for training. But, as always, there’s a catch: if companies don’t play by these rules, they could face fines worth up to 7% of their global revenue.