OpenAI's App-Marketplace

PLUS: Amazon's Leaked Document

Hey there, this is Beeps! Here are our main topics of the week:

  • ⚡️ OpenAI plans a App-Marketplace for AI Models

  • 💻 Dropbox Unveils Dash and Dropbox AI

  • 🖥️ Google warns employees of its own AI chatbot Bard

  • 🤖 Voicebox by Meta Unleashed

  • 👨‍⚖️ SDXL 0.9: Stability AI's Text-to-Image Powerhouse

  • 🚫 DeepMind's RoboCat: The Swiss Army Knife of Robotics

Read time: 5 minutes

So, our buddies at OpenAI are planning to run an AI marketplace. Here, developers can peddle their AI models, all built on OpenAI tech. Oh, and this isn't rumor-mongering; reliable insiders spilled the beans to The Information. Imagine, tailored AIs from spotting financial fraud to answering market-specific queries, all up for grabs! This move might give Salesforce and Microsoft a run for their money, as OpenAI widens its customer circle. Two potential sellers include Aquant and Khan Academy, who fancy flogging their ChatGPT-powered AIs. Since ChatGPT hit the scene, businesses have been falling over themselves to adopt it. Fasten your seatbelts, AI's biggest yard sale is about to kick off!

Dropbox is here with their shiny new tools - Dropbox Dash and Dropbox AI. Apparently, we live in a world where Googling the entire human history is easier than finding our own content, so they've decided to save us from the 'work about work' nightmare. Dash, your AI-powered universal search tool, ropes in all your scattered content from various apps into one search bar. Stacks, Start Page, and soon, generative AI that answers your questions - it's all happening. Then, there's Dropbox AI. Need a quick summary of a colossal document or an answer without sifting through hefty files? Say hello to AI-powered file previews. Did I mention Dropbox Ventures? A $50M fund for propelling AI startups. Time to grab some popcorn and see how this show unfolds!

Even Google seems a tad wary of its own baby, the AI chatbot Bard. The company is warning employees to keep their secrets safe from chatty pals like OpenAI’s ChatGPT and indeed, their own creation. Can't blame them, given the potential for human reviewers peeking in or chatbots inadvertently spilling beans while learning. Even Amazon's lawyers are playing cautious, advising against sharing confidential stuff with ChatGPT. Apple has joined this 'Keep-it-secret' club too, banning employees from using ChatGPT and GitHub Copilot. Oh, and they've also thrown some serious cash at AI startups. Meanwhile, Google's Bard, built on their LaMDA engine, faces a delayed release in the EU over privacy concerns.

Ladies and gents, meet Voicebox, the latest brainchild in generative AI for speech. This whizz-kid is multilingual, mastering six languages, and can edit your audio clips like a pro. Pesky dog barks or car horns in your audio? Voicebox will erase them without messing with the content or style. What's really cool is this AI can mimic your voice based on a tiny two-second sample! And here's the kicker - it can transfer your voice style across different languages. Speaking in French with your own voice while knowing zilch French? Oui, c'est possible! With Voicebox, the future of AI-driven virtual assistants and audio editing for creators looks pretty darn promising. Say hello to the era of seamless cross-lingual communication!

Roll out the red carpet for Stability AI's SDXL 0.9, the latest and greatest in the text-to-image world. This brainy model puts its predecessor to shame with massively improved image and composition detail. But wait, there's more! It isn't just about text prompting; we're talking image-to-image variations, filling in image gaps, and even image extensions! Powered by a whopping 3.5 billion parameters, it's your go-to for hyper-realistic visuals in films, TV, music, and more. Think of it as your own personal art director, but without the tantrums and existential crises. And the best part? You don't need a supercomputer to run it - just a decent GPU. Let's just say, the future of AI imagery is here and it's looking pretty picture-perfect.

Meet RoboCat, DeepMind's newest AI model that's acing the robotics scene. No ordinary kitty, RoboCat can juggle multiple tasks across different robotic arms. It's like a Swiss Army knife, but for robots! Drawing inspiration from Gato, another DeepMind genius, RoboCat was trained on a buffet of robotics data from both simulations and real-life scenarios. It then honed its skills by practicing tasks about 10,000 times! Talk about perseverance. Sure, its success rate varies, but don't hold that against it - it's still learning. The goal? Teaching RoboCat new tricks with fewer than 10 demos. In short, RoboCat is the future, making "herding cats" look easy in comparison.

  • Cisco Unveils AI Supercomputer-Friendly Networking Chips (link)

  • DeepMind's Latest AI Agent Takes Gaming to Another Level (link)

  • ChatGPT's Function Calling Marks a Significant Milestone (link)

  • Generative AI: The Future of Google Search Engine (link)

  • Data Wars: In the Battle for World Control, Data Reigns Supreme (link)

  • Opera's New Browser One with Free ChatGPT Integration (link)

  • ByteDance Invests $1B in Nvidia GPUs to Fuel AI Progress (link)

  • Self-Healing Code: The Future of Software Development (link)

  • Is it Time to Jump on the Generative AI Company Bandwagon? (link)

  • Generative AI and Robotics Forge Synthetic Data Generator (link)

  • ChatGPT Plugin Store Hits Milestone with 500 Available Plugins (link)

  • The Underground Trade of NVIDIA Computer Chips in China (link)

  1. Fini AI - Turn your knowledge base into AI chat... in 2 minutes.

  2. Zuddl - Stunning Webinars that Drive Better Results

  3. Puzzicle - Turn video lessons into interactive games

  4. Spell AI - Delegate your tasks to autonomous AI agents in your browser

  5. WorkGPT - WorkGPT is an agent framework in a similar fashion to AutoGPT or LangChain

Meme