AI Weekly Roundup - Cracking Open the Biggest Brainwaves in Artificial Intelligence!

Ladies and Gents, grab your soda and popcorn, because we've got a news roundup that'll leave you feeling more shocked than if you found out your microwave was secretly a Decepticon!

Get this – Snap has rocked the boat with SnapFusion. It's a techno-marvel that can turn text into an image on your mobile in less time than it takes to sneeze.

StabilityAI is here to uncrop your life with its new feature in Clipdrop. Basically, this ingenious tool paints you a background for any image as if it were the Van Gogh of AI. It's like instant room extension, but for photos. Best part, you can try this wizardry for free!

Meanwhile, Google's Bard is beefing up. It now has a talent for running math calculations in the background, making it 30% smarter at number-crunching. It also has a new party trick – it can fling any table it creates straight into Google Sheets.

Microsoft comes into the ring with Orca, a goliath of a model with a whopping 13 billion parameters. It's a heavyweight that can hold its own against the likes of ChatGPT and even dance a bit with GPT-4.

Google's also raising the bar with Visual Captions. It's a system that brings your words to life by adding images to video chats. We're living in the future, people!

AI history is made with AlphaDev, a brainchild of Google DeepMind, discovering sorting algorithms that beat the pants off human benchmarks. These AI-masterpieces have now found a home in the LLVM standard C++ sort library. The smarty pants AI has also crafted a new hashing algorithm!

Adobe's rolling out the red carpet for enterprise customers with its Firefly generative AI model. Picture getting a bespoke suit tailored, but it's a model customized with your branded assets.

In other news, HuggingChat (the humble cousin of ChatGPT) developed by HuggingFace now has a fancy web search feature.

Tafi – the folks who own Daz 3D – just dropped a text-to-3D character engine. It's like being able to conjure a virtual genie with a mere text prompt!

Runway's Gen-2 for text-to-video is finally out of the oven. Free trials are on offer - hurry while stocks last!

Europe is stepping up its game, planning to label AI-generated content. It's like a warning label, but for disinformation.

Google presents the amazing SQuId, a model with 600 million parameters that's all about evaluating how natural synthesized speech sounds in various languages.

Together launched its v1 versions of RedPajama-INCITE models, which outperform the Falcon-7B. It's like the tortoise beating the hare in a marathon!

Wordpress is offering its new Jetpack AI Assistant to generate blog posts, structured lists, and tables. It's like having a little elf in your editor doing your work for you.

Lastly, Google Research is flaunting StyleDrop. Imagine being able to summon any image you want with just a text prompt and a reference image. It's like having your own personal artist at the tip of your fingers.

So, folks, that's all the tech hoopla for now. Tune in next time for more!