- Daily Zaps
- Posts
- 🚗 ChatGPT For Your Daily Commute: GM's Plan to Integrate AI into Your Car
🚗 ChatGPT For Your Daily Commute: GM's Plan to Integrate AI into Your Car
CarGPT, Chopped AI, What is Visual ChatGPT? Uber For Package Delivery
DZ Team
March 14, 2023 • Estimated Reading Time: 8 minutes
Have you guys been keeping up with the Silicon Valley Bank saga? It's been a rollercoaster ride! But hey, good news - Uncle Sam's got everyone’s back and will make sure all deposits are safe and sound. 🙌
Now that we can breathe easy, let's get back to the fun stuff: nerding out over AI!
Here's what we got for you today:
🚗 GM Wants To Bring ChatGPT to Cars
💬 What is Visual ChatGPT?
🤖 GPT-4 Coming This Week?
📦 Drone Package Delivery Network
🚀 Cool A.I. Tools, Demos, and News
🖼️ A.I. Generated Image of the Day
⚡️ A.I. Meme of the Day
GM Wants To Bring ChatGPT to Cars
Picture this: you're cruising down the road and your car starts chatting with you - but it's not your backseat driver, it's ChatGPT! Yup, you heard that right. General Motors, the makers of Chevy, Cadillac, Buick, and GMC vehicles, are cooking up a virtual personal assistant that uses AI models similar to ChatGPT.
So how does it work? Well, the assistant is voice-activated, meaning you can chat with it hands-free. And the whole thing is powered by Microsoft's Azure cloud service, which has exclusive rights to some of the coolest AI tech out there - including ChatGPT, DALL·E (an image creator), and Microsoft's Bing chatbot.
We might not be at the point of letting AI drive us to work just yet, but hey, at least we'll have some friendly company on the ride there! More details here.
What is Visual ChatGPT?
When it comes to AI art generators, most of them rely on Visual Foundation Models (VFM) like Stable Diffusion to create images. While they can get the job done, they have their limitations.
However, Microsoft is taking things to the next level with their latest creation, Visual ChatGPT. This awesome tool uses Open AI's GPT Large Language Model (LLM) as well as Microsoft's own Prometheus model to generate images. And here's the cool part: to create Visual ChatGPT, Microsoft combined a bunch of different VFMs together and bolted them onto the flexible GPT model.
They did this by creating a "Prompt Manager," which essentially helps bridge the gap between ChatGPT and these VFMs. This allows ChatGPT to leverage the VFMs and receive their feedback in an iterative way until the resulting image meets the user's requirements or reaches the ending condition. It's a pretty neat way to bring together different AI models to create something truly unique and powerful! Demo here.
GPT-4 Coming This Week?
ChatGPT, along with numerous other AI text-based tools, are currently powered by OpenAI's GPT-3.5. However, many users of these tools have been eagerly anticipating the release of GPT-4, which is expected rumored to be a significant improvement and possess multi-modal capabilities, including the ability to process images and videos. More details here.
“We will introduce GPT-4 next week, there we will have multimodal models that will offer completely different possibilities – for example videos," Braun said. The CTO called LLM a "game changer" because they teach machines to understand natural language, which then understand in a statistical way what was previously only readable and understandable by humans.
Drone Package Delivery Network
A Wing Delivery Network that's like Uber but for boxes! The idea is to transport small items like packages and meals from one spot to another using smaller, more efficient vehicles that match the size of the goods being transported.
It just makes sense, right? Why use big, clunky vehicles to transport a few little things when you can use smaller, more nimble ones? That's the key to making last mile delivery sustainable, cost-effective, safe, and reliable. That's why you've got to check out this super cool video below and learn more about this innovative company here.
🚀 Cool A.I. Tools, Demos, and News
The world of AI and machine learning is moving incredibly fast, and it's tough to keep up with all the cool new developments! Here are some of the most exciting AI-powered tools and news.
🖼️ 100% A.I. Generated Image of the Day
Midjourney Prompt - editorial photo | GENRE: Drama | EMOTION: Melancholic | SCENE: A panoramic view of the New York City skyline from the top of the rock at dawn, with a blueish hue in the sky at eye level | TAGS: 8k, shot from movie, cinema composition shot, professional color grading, epic volumetric lighting, sharp focus, film grain, atmosphere, melancholic mood, urban landscape, morning blue hour, cityscape, panoramic shot, majestic view, manhattan --ar 16:9
🤭 A Meme of the Day
Regarding SVB, the FDIC had everything under control! Just like many other bank closures…FDIC closed SVB on Friday and opened replacement bank back up on Monday under new management like nothing ever happened. Easy peasy for customers with no loss in deposits.
The real losers in this fiasco were those VCs who spread the panic on social media, blogs, podcasts, you name it. They're getting roasted on Twitter and other platforms right now. Ouch!
🔬 A.I. / Machine Learning Research
🌴 PaLM-E - What happens when we train the largest vision-language model and add in robot experiences? The result is PaLM-E , a 562-billion parameter, general-purpose, embodied visual-language generalist - across robotics, vision, and language.
🧐 Visual ChatGPT - One of the most significant breakthroughs is ChatGPT, which is built upon InstructGPT. Although powerful, ChatGPT is limited in its ability to process visual information since it is trained with a single language modality, while Visual Foundation Models (VFMs) have shown tremendous potential in computer vision, with their ability to understand and generate complex images.
🧮 MathPrompter -Large Language Models (LLMs) have limited performance when solving arithmetic reasoning tasks and often provide incorrect answers. Unlike natural language understanding, math problems typically have a single correct answer, making the task of generating accurate solutions more challenging for LLMs.
👣 GigaGAN - GigaGAN offers three major advantages. First, it is orders of magnitude faster at inference time, taking only 0.13 seconds to synthesize a 512px image. Second, it can synthesize high-resolution images, for example, 16-megapixel pixels in 3.66 seconds.
♻️ Refer Daily Zaps
Can you do me a solid and refer a buddy, colleague, or family member to Daily Zaps? As a thank you, I'll send you an Airtable database / list of over 170 AI-focused startups complete with links to their sites, founder names, funding amounts, investors, and all that good stuff. Sound good? Let's do this!
Peace out,
Daily Zap Team
How much did you enjoy this email? |