Meta’s Breakthrough AI Tool That Converts Text Into Rich, Detailed Soundscapes, Enabling Professional Musicians and Business Owners to Generate Unique Audio Content Effortlessly


AudioCraft: A Generative AI Tool For Audio and Music

Say hello to AudioCraft, a fancy new AI tool that can whip up tunes and sounds from just a bit of text. Think about a band member making a new tune without strumming a guitar, or a small business making catchy tunes for their Instagram ads. This gadget has three parts: MusicGen, AudioGen, and EnCodec.

MusicGen is like a jukebox, you put in a bit of text and out comes a melody, all made from music Meta has the rights to. AudioGen is your personal soundboard, creating everyday noises from text like a dog barking or car horns honking. EnCodec? It’s the magic sauce that makes the music sound even better. They’re also sharing all these tools for anyone to tinker with.

Meta’s giving this stuff away for free, so other folks can play around and make it even better. They’re hoping it’ll be a game-changer for the way we make and listen to music.

Up until now, AI has been great at making images, videos, and text, but sound was left in the dust. Making high-quality sound is a tough gig, especially music. But with AudioCraft, they’ve made a solid and simple tool that lets folks mess around with sounds and music easily. They see AudioCraft as a helper for musicians and sound designers to get creative and make some cool stuff. 


Google Speeds Up AI Summaries, Adds More Visuals To SGE

Google’s decided to spruce up its Search Generative Experiment (SGE), an AI-powered search feature that was launched at a shindig in May. What’s the new hotness, you ask? Now, SGE’s gonna show you pictures or videos that go with what you’re searching for. So if you’re curious about the smallest birds of prey or need help getting a stain out of your marble countertop, you’ll see pics of those tiny raptors or videos about stain removal right there in your search suggestions. Neat, huh?

They’ve also decided to put the publish dates next to links suggested by SGE. This way you’ll know just how fresh that info is. And the Googlers are working on making it easier for you to find web pages that back up the info in their AI summaries.

Apparently, they’ve also made some tweaks to speed up the AI-powered results, which is always good news. You can sign up to test these shiny new features through Search Labs. Once you’re in, you can access them through the Google app on your iOS or Android device or via Chrome on your computer.


Tinder tests AI photo selection feature to help users build profiles

Tinder’s cooking up a nifty tool that’s gonna pick your profile photos for you. Yep, you heard it right! Their new AI feature’s gonna dive into your photo album and fish out the top five pictures that show off the real you. This was dropped by Bernard Kim, the big cheese at Match Group (Tinder’s parent company), during an earnings call.

But that’s not all! Tinder’s also working on another AI feature to show the right content to the right folks to up their game in relevancy. They’re even hinting at AI helping to craft a snazzy bio for users because let’s face it, not all of us are wordsmiths.

To keep things kosher, Tinder’s got an AI verification feature that needs a video selfie from users to confirm they’re not bots or catfishers.


Uber Is Developing an AI-Powered Chatbot to Integrate Into App

Uber’s cooking up a chatbot powered by artificial intelligence for its app, aiming to make things like customer service smoother and more efficient. CEO Dara Khosrowshahi spilled the beans on the project recently, though he didn’t dish on the deets about what the bot can do. But, it’s not surprising considering Uber’s been using AI for a good while now to match you with your ride or delivery – it’s all about the science of day, distance, and more. 

Seems like they’re not alone in this tech rodeo, with DoorDash and Instacart also working on their own chatbots. Despite posting their first operating profit, Uber’s stock took a 6% hit when their revenue didn’t meet the mark and growth started to slow.


AMD considers making a specific A.I. chip for China to comply with export controls

AMD is considering creating an AI-specific chip for China, to stick to U.S. export rules. It’s a move following in the footsteps of rivals Nvidia and Intel. AMD’s CEO, Lisa Su, announced during an earnings call that the company views China as an “important” market and intends to follow U.S. export controls to the letter. The plan is to cater to Chinese customers seeking AI solutions while staying in line with these regulations.

“Accelerator” chips, which are used to train large quantities of data for AI apps, are in the spotlight here. AMD is ramping up production of its MI300 chip, hoping to compete with Nvidia’s graphics processing units, often used for AI training.

With few local alternatives, China is a major market for U.S. chip makers, especially in AI. AMD is betting big on its MI300 AI chip as it seeks to rival Nvidia and is banking on it to boost its data center business this year. AMD predicts about a 50% growth in this sector in the second half of the year, with the new AI chip playing a crucial role.


Google pulls its AI Test Kitchen app from Play Store and App Store

Google has yanked its AI Test Kitchen app from both the Play Store and the App Store to give full attention to its web platform. The app was launched last year to allow folks to play around with AI-powered projects like LaMDA 2. The company verified the change and said it’s easier to roll out updates on one platform only.

Google had announced a “Season 2” for the AI Test Kitchen with new experiments, but those never saw the light of day. At the moment, the Test Kitchen only features one text-to-music language model experiment named MusicLM, announced at Google IO.

During Google IO, Google launched a portal called Google Labs, where users can check out AI-based experiments. This platform includes the MusicLM experiment too. Now, users can tinker with AI on the AI Test Kitchen page with the lone MusicLM experiment, or the Google Labs page that features a variety of projects including Search Labs, the AI note-taking project NotebookLM, AI features for Workspace, and again, the MusicLM project.


CoreWeave came ‘out of nowhere.’ Now it’s poised to make billions off AI with its GPU cloud

Cloud startup CoreWeave is shooting for the stars, set to make big bucks off the AI boom. They’ve pivoted from crypto mining to focusing on accelerated GPU workloads – a fancy way of saying they make computers run faster using graphics cards. This caught the eye of Nvidia, who teamed up with them to build the world’s speediest AI supercomputer.

CoreWeave started out as three guys who turned their cryptocurrency mining side gig into a business in 2017. Their big break came when AI companies using their service started spreading the word, leading to a funding surge this year. They’ve raised over $400 million in total, with tech big shot Nvidia even pitching in.

Their impressive progress has led to major deals, like Microsoft planning to shell out potentially billions on CoreWeave’s cloud infrastructure. They’ve got a massive backlog of client demand and are building at 12 different data centers. They’re making big strides, especially considering most people hadn’t heard of them a few months ago.


Character-Creating Startup Inworld AI Raises More Cash At $500M Valuation 

Inworld AI, a whiz-kid startup that helps game developers cook up smart, learning characters, just secured a new round of funding. Once all is said and done, they’re looking at an extra $50 million in their piggy bank, pushing their valuation up to a cool half-billion dollars. Some big dogs like Lightspeed Venture Partners and Stanford University, among others, joined in this round.

Inworld AI ain’t your average startup, they’re all about mingling AI and gaming. Their “Character Engine” gives game developers the tools to create characters that can learn on the fly, remember stuff, get motivated, and even build relationships. Pretty snazzy, huh?

The big cheese of Inworld AI, Ilya Gelfenbeyn, says this extra cash will help them support developers now and stay on top of the ever-changing world of generative AI in the future.


AI chip firm Tenstorrent raises $100 mln from Hyundai, Samsung

Tenstorrent, a Canadian startup developing artificial intelligence (AI) chips and led by industry veteran Jim Keller, has raised $100 million in funding. This round includes investments from Hyundai Motor Group, Samsung’s Catalyst Fund, and several other investors. With this new funding, the total amount raised by Tenstorrent has reached $334.5 million, bringing the company’s valuation to $1 billion.

Tenstorrent is among a group of companies aiming to challenge Nvidia’s dominance in the market for AI chips. Jim Keller, who previously developed chips for Apple, Tesla, and Intel, joined Tenstorrent as CEO earlier this year.

The new funding includes $30 million from Hyundai, $20 million from Kia, and the remaining $50 million from Samsung’s Catalyst Fund and other investors. The funding round was structured as debt that will later convert to stock, so Tenstorrent will not officially have a new valuation until its next equity fundraising round, expected to take place next year.


VAST Data launches unified data platform for the age of AI

Tech company VAST Data is stepping up its game, launching a new platform called VAST Data Platform that’s aimed at fueling the AI age. This platform bundles storage, database, and virtualized computing services in a scalable system, making it easier for businesses to handle complex AI applications.

The goal is to go beyond just crunching numbers and creating reports. Instead, VAST Data wants to use deep learning applications to tackle some of humanity’s biggest challenges. This ambitious plan is built on the belief that as AI evolves, we’ll be able to use it to speed up processes that used to take years down to just days.

But how does the VAST Data Platform do all this? It works by simplifying the process of managing and learning from data into one unified stack. This “continuous computing engine” combines different components to handle data from the edge to the cloud. They’ve got their DataStore for scalable storage, their DataBase for quick and large-scale queries, and their DataEngine for understanding the data’s characteristics.


SettleMint’s AI assistant aims to help web3 developers write better smart contracts

SettleMint, a company that provides low-code blockchain programming tools for businesses, has introduced an AI assistant to its platform. The AI assistant aims to help developers create smart contracts, integrate data, and improve quality assurance testing. The goal is to aid developers in understanding the functionality of each line of code, rather than replacing human roles in coding.

The AI assistant from SettleMint has been trained primarily using OpenAI’s GPT-4 and is provided with the latest information on blockchain development. While the AI can suggest changes to prevent vulnerabilities, human audits are still required, particularly due to the complex nature of blockchain systems. Nonetheless, the introduction of this AI tool can help streamline the process by focusing human audit efforts on non-trivial parts of the code.


How Will Artificial Intelligence Change the News Business? Here are three theories of the case.

The news industry is in a bind with AI’s growing presence. Companies like OpenAI and Google are handing out powerful AI tools, like ChatGPT and Genesis, that can crank out articles or help journalists with the tough parts of their job. However, not everyone’s gung-ho about it – some big-shot publishers are suing, claiming AI companies are swiping their content without paying a dime.

Three theories are being bandied about:

One: AI might take over journalism, churning out stories like an assembly line. But it’s a risky play. If one site starts pumping out AI-generated content, others will follow suit, potentially driving down its worth. Plus, it’d mean going head-to-head with AI tools from tech bigwigs.

Two: AI could be the sidekick to journalists. Think of it like a super-powered assistant that takes care of drafting, transcription, and organizing story structure. It might give reporters more time for real journalism. But it’s not all roses and rainbows. If news bosses see this as a chance to cut staff or boost content production, it might backfire.

Three: AI might swallow journalism whole. As AI becomes commonplace in newsrooms, it could stir up labor spats, alter job roles, and shift power to bosses. Worse yet, there might be court fights over using news to train AI. If tech companies have to shell out big bucks, news orgs might end up just feeding AI models. That means news outlets become glorified wire services, filling in blanks for AI news machines. The fallout? Media as we know it could be obliterated, with tech companies ruling the roost.


One of Gaming’s Biggest YouTubers Wants to Replace Himself With AI

Famed gaming YouTuber, Kwebbelkop (real name Jordi Van Den Bussche), is planning to replace himself with AI. Exhausted from constant content creation, Van Den Bussche faced the common creator problem of not being able to step away without everything grinding to a halt. This inspired him to develop an AI tool capable of generating and uploading videos with little to no involvement from him.

Van Den Bussche’s AI platform consists of two versions. One version learns from the creator’s on-camera personality and content, producing new videos from that. The other simplifies the creation process, transforming basic prompts into full videos. While he’s keeping mum on the exact process behind these tools, they’re certainly launching at a key moment for AI in the creative space.

Although he’s not stepping away from creativity, Van Den Bussche is content to let AI take the reins. He’s convinced the system can reproduce his creativity and believes it’s a viable solution for creators wanting to continue their brand without the stress. However, critics argue that the appeal of creators lies in their humanity, and AI-generated influencers may not resonate as deeply with audiences.