Today:

OpenAI announcement LEAK! Cost Cutting 20x plus massive API updates. run for cover...

OpenAI, the brains behind ChatGPT, wants to reel in more developers with cheaper and faster tools. Next month, they're gonna roll out some cool updates that might make building AI apps up to 20 times cheaper.

They're also giving devs new gadgets like the ability to understand and chat about pictures, which is super handy for everything from movies to doctor stuff. They're gonna show off these new tools at their very first developer meet-up in San Francisco come November 6.

WATCH THE VIDEO ON YOUTUBE

Google’s AI-powered search experience can now generate images

Google's upped its game with a new feature that lets you whip up images using just a text prompt through their Search Generative Experience (SGE). Similar to something Microsoft's been doing with Bing Chat and OpenAI's model, Google's version lets you type in a request and get a few AI-generated image options to pick from.

They're trying to keep it on the up-and-up by avoiding generating any iffy content and marking all images as AI-created. Only folks 18 and up can use it, and they're playing it safe by not allowing creation of realistic faces or images of famous people.

On top of that, you can also generate written drafts with different tones or lengths right from the search bar, which you can pop into Google Docs or Gmail if you dig what you see.

READ THE ARTICLE ON THE VERGE

Get Ready for a Deluge of AI-Written Emails, Thanks to Microsoft Outlook

Outlook’s about to make your inbox even lazier. Microsoft's cooking up some new artificial intelligence (AI) features that'll read your emails, get a feel for how you tap the keys, and start cranking out messages just like you. Imagine: it'll shoot off apologies for being late and might even pen a sweet nothing to your honey. This isn’t just finishing your sentences; it's drafting whole notes with your vibe.

And heads up: this robot ghostwriter comes with no disclaimer. Recipients won't know it’s AI chirping at them, not you. The feature's being wrapped into the $30 a month per user package for businesses, and it's got a few eyebrows raising about authenticity in our inbox connections.

READ THE ARTICLE ON BLOOMBERG

Adobe previews AI upscaling to make old, fuzzy videos and GIFs look fresh

Adobe's cookin' up something sweet for old, grainy videos and GIFs! They've got a new tool in the works, dubbed “Project Res-Up”, that uses AI to drastically upgrade the quality of low-res images and video. Think making your blurry, vintage clips sharp and clear, kinda like giving grandma's old recipe a modern twist.

It ain’t available for a test drive just yet, but early peeks show it works wonders, like turning a fuzzy old movie clip into something way crisper and adding new details that weren’t there before. Although it's still in the experimental phase, this tech might sneak into some of Adobe’s popular editing apps down the road.

READ THE ARTICLE ON THE VERGE

New Federal Bill Would Ban AI-Generated Deepfake Vocals

U.S. senators from both parties are making moves to put the kibosh on AI-created fake celeb vocals, like the “Fake Drake” tune that dropped earlier this year. Their draft bill, dubbed the NO FAKES Act, aims to shield artists and celebs from unasked-for digital doubles of their image or voice. It wants to lay down federal law that lets folks sue if someone spins up a digital copy of them without the thumbs up.

While there are some state-level protections, this would make it a nationwide deal and even let your heirs keep control for 70 years after you kick the bucket. The senators are trying to walk the line, making sure to respect free speech while also keeping artists' rights in the crosshairs. Music industry bigwigs are cheering for a final version that shields against misuse, but they're also hoping to keep the innovative spark of AI alive and well in the biz.

READ THE ARTICLE ON BILLBOARD

Hook wants to help you create a legal remix of your favorite track for TikTok

Gaurav Sharma, a seasoned player in the music streaming game, kicked off Hook, aiming to make remixing tunes for TikTok and other social media a smooth and lawful endeavor. Creators can tweak a short bit of their jam of choice (under 60 seconds) using Hook’s slick AI-powered tools.

And the real game-changer? It plans to ensure artists get their fair shake by compensating them for these fresh, fan-made remixes. A test version of the app is on deck for next month, with a full-blown launch penciled in for 2024. With a snug $3 million already banked from investors like Steve Cohen’s Point72 Ventures, it's all systems go for Hook to potentially become a powerhouse in both the creator and music worlds, while also having artists and labels singing its praises.

READ THE ARTICLE ON TECHCRUNCH

A New Tool Helps Artists Thwart AI—With a Middle Finger

Artists are cheesed off because AI is snatching their work from the web without asking and then using it to create new stuff - no "please", no "thank you", and definitely no payday. Some techy folks rolled out a new tool named Kudurru, a kind of virtual watchdog that sniffs out these AI bandits when they try to pilfer images and throws them a curveball.

Artists can choose to either shoo them away or send them a sassy image (like a middle finger) instead of their artwork. While it's a cheeky “take that!” to the tech giants, it's not solving the bigger issue. Some are saying the real fix needs to come from the bigwigs in government, making laws that say “You want to use it? You gotta ask first.” But until those rules are laid down, artists are having a little fun sticking it to the image-swiping AIs with a one-finger salute.

READ THE ARTICLE ON WIRED

Energy-efficient transistor could allow smartwatches to use AI

Scientists whipped up a new kind of transistor that's way easier on batteries and could let smartwatches and other wearables use artificial intelligence without conking out too quickly. These transistors, made of molybdenum disulphide and carbon nanotubes, can handle lots of steps in smart processes all at once, unlike the usual silicon ones that can only do one thing at a time.

Mark Hersam and his team from Northwestern University in Illinois showed that this new transistor could help AI accurately sort through heart data from electrocardiogram tests. The hope is that, in the future, this energy-saving transistor can be used in wearables, like fitness trackers or blood pressure monitors, without them needing to chat with distant computers (cloud-based AIs) or run out of juice too fast. Plus, keeping data local means better privacy.

READ THE ARTICLE ON NEW SCIENTIST

AI reads text from ancient Herculaneum scroll for the first time

Nebraska college kid, Luke Farritor, cracked the code to read an ancient scroll from Herculaneum, a city wiped out by a volcano in 79 AD. Using fancy computer algorithms, he detected Greek letters on the charred, rolled-up papyrus, giving historians and scholars a wild ride into the past.

This brainy 21-year-old scored a cool $40,000 for spotting more than 10 characters in this puzzle from the past, and the discovery could unlock a treasure trove of ancient texts from the only library from that time period to survive intact. Now, experts are buzzing with excitement to see what other ancient secrets can be unveiled with a little help from modern tech.

READ THE ARTICLE ON NATURE

New proteins, better batteries: Scientists are using AI to speed up discoveries

Scientists are using AI to speed up protein designs, aiming to make quicker medical solutions, from drugs to vaccines. Susana Vazquez-Torres from the University of Washington says AI has made it so they can find potential treatments in just months. Imagine proteins as keys.

Before AI, scientists had to test every possible key to find the right one. Now, they can design the perfect key on the first go. While this AI magic works wonders with proteins, because there's lots of past data to learn from, it's trickier for things like new solar panel materials where data isn't as abundant. But for many, including Vazquez-Torres, the rise of AI in science isn't a job-killer. It's a game-changer.

READ THE ARTICLE ON NPR

Biden to push for government standards on AI

Joe Biden's cookin' up some new rules for AI, trying to reel in the wild west that’s technology these days. He's crafting an executive order to get cloud computing companies to keep a sharp eye on users cooking up powerful AI systems, using the big bucks the government spends on tech to shape up standards.

There’s talk of getting the National Institute of Standards and Technology to jazz up industry guidelines and make AI systems safer and more reliable. The order's also supposed to make it easier to bring in top-notch AI pros from other countries and amp up AI education here at home. While the specifics are still under wraps, and the White House is keeping mum, this move is a part of a bigger plan to get a grip on AI risks while still keeping the U.S. in the global tech race.

READ THE ARTICLE ON POLITICO