🚙 After 15 years, Uber finally turned a profit.
🏢 Adam Neumann wants to buy WeWork out of bankruptcy.
👓 Pokémon Go creators + ex-Apple execs announced new AI glasses.
👾 Google’s most capable LLM is here + RIP "Bard." It's just Gemini now.
🖼️ DALL-E 3 will automatically watermark AI images, according to OpenAI.
Apple teamed up with researchers at the University of California, Santa Barbara (Go gauchos!) to create a new AI model called MGIE (MLLM-Guided Image Editing).
The model can edit images — think of the stuff you do in Photoshop, but with plain English prompts. It can crop, resize, rotate, brighten, sharpen, remove objects, and so on, plus edit specific parts of your image, like changing the shape, size, and texture of items.
Perhaps the most interesting or “revolutionary” thing about the model is that it derives expressive instruction from your prompt.
Imagine using a phrase to ask AI to edit your photo for you. It might look something like this: “Make the sky more blue.” But MGIE can produce instructions like this: “Increase the saturation of the sky region by 20%,” explains Michael Nuñez for VentureBeat.
The model also works by creating a representation of the edit that the end user wouldn’t see, which it uses as a guide for how to manipulate pixels — “an end-to-end training scheme that jointly optimizes the instruction derivation, visual imagination, and image editing modules.”
That’s a lot for me to process too, but the key point is that the model is being recognized as a breakthrough for using multimodal large language models (MLLMs) for interpreting user input and making edits at the pixel-level. People are also loving that it's open-source. The project is on GitHub and a demo is on Hugging Face.
Tired of explaining the same thing over and over again to your colleagues? It’s time to delegate that work to AI. guidde is a GPT-powered tool that helps you explain the most complex tasks in seconds with AI generated documentation.
Simply click capture on our browser extension and the app will automatically generate step-by-step video guides complete with visuals, voiceover and call to actions.
The best part? Our extension is 100% free.
🤗 AI for Good: Sensay is a platform that uses AI to help dementia patients and their families preserve cherished memories, likenesses, stories, and personalities while also assisting patients by digitally replicating people closest to them.
🔨 Midjourney phone: Okay, maybe not a phone, but it does look like Midjourney could be cooking up something in the hardware department. The company recently hired an engineer from the Apple Vision Pro to be its “Head of Hardware.”
PRODUCTIVITY
- Keycheck is a library of keyboard shortcuts for 100+ different apps.
- Linga is a reading app, translator, and vocabulary builder all rolled into one.
DEV TOOLS
- Friendly Fire connects GitHub to Slack so you can smartly assign pull requests and notify code reviewers.
- Context lets you test LLM prompts & models side-by-side against many inputs.
DESIGN
- toddle is a visual development tool for building products.
- MyDevPage lets you create your portfolio in minutes without code.
FUN
