Custom GPTs by Google, Midjourney Dive Into Hardware, and Nvidia & Apple Join OpenAI’s Funding
PLUS HOT AI Tools & Tutorials
Hello and welcome to our weekly roundup!
Have you tried Google Gems? I've tested a few premade versions and haven't decided on my attitude yet. It seems we need to prepare a post on this topic.
Well, let's talk a little about Gems and then discuss the rest of the news. We've had more than enough of them over the past week. Let’s get started.
Want to launch a project but have been putting it off? Perhaps you just need a good boilerplate:
This Creators’ AI Edition:
Featured Materials 🎟️
News of the week 🌍
Useful tools ⚒️
Weekly Guides 📕
AI Meme of the Week 🤡
AI Tweet of the Week 🐦
(Bonus) Materials 🎁
Featured Material 🎟️
Google Releases Custom Gems
As promised at Google I/O, Google released its platform for creating custom chatbots. It’s called Gems. Gemini Advanced (and Business & Enterprise) users can now create their own AIs. And yes, it works roughly the same way as with Custom GPT. But with a limitation: for now, you can't share or sell your Gems to other users.
With Gems, you can create a team of experts to help you think through a challenging project, brainstorm ideas for an upcoming event, or write the perfect caption for a social media post.
You can also create a custom chatbot with different qualities and character traits by describing a set of instructions, and it will have access to your YouTube, Gmail, or Drive account. But if you're too lazy to do that, Google suggests using one of five premade options. The current list of Gems looks like this:
Learning coach for breaking down complex topics
Brainstormer for inspiration and creating new ideas
Career guide for career consultations with detailed plans to refine your skills
Writing editor that helps you to write better
Coding partner that levels up your coding skills
Gems look good, but honestly, sometimes Google raises questions. In terms of AI, the company is clearly in the role of a laggard and the release of a Custom GPT competitor without one of the main features (the ability to share Gems themselves with other users) is perplexing.
Fine, let's hope that at least the implementation is in order.
Keep your mailbox updated with key knowledge & news from the AI industry
3 New Google Gemini AI Models
In addition to Gems, Google also released three experimental Gemini models for developers. The list of updated models looks like this:
Gemini 1.5 Flash-8B: A new compact model with more features
Gemini 1.5 Pro model (Exp 0827): An improved version for coding and handling complex cues
Gemini 1.5 Flash (Exp 0827): Gets a significant performance boost that promises better performance across the board
You can try out all these models at Google’s AI Studio.
News Of The Week 🌍
Bloomberg: Nvidia and Apple Could Join OpenAI’s Next Funding Round
OpenAI's list of key investors is about to be joined by a couple of more well-known companies. Bloomberg reports that Nvidia and Apple may join the list of investors in the next round of funding. Microsoft, which already owns 49% of OpenAI, may also participate. The ChatGPT developer's valuation is expected to rise to $100B in this funding round.
Apple and Nvidia's support for OpenAI is an expected move. As part of the partnership, the startup trains its gen AI models on Nvidia GPUs. It is also responsible for integrating ChatGPT into iOS for Apple Intelligence.
ChatGPT Usage has Doubled Since Last Year
OpenAI has disclosed data on the demand for its chatbot. According to the developer, the ChatGPT audience has doubled year-on-year, with more than 200 million users using it weekly. In addition, 92% of Fortune 500 companies now use ChatGPT regularly, and usage of its automated API has doubled since the release of GPT-4o mini in July.
Yes, everyone needs AI. I think this news complements the previous one very nicely and answers the question of why Apple and Nvidia are willing to invest in OpenAI (if people asking such a question even exist).
Midjourney Is Getting Into Hardware
The most famous image generation platform developer is starting to develop physical devices. The other day, Midjourney's Twitter account announced that the company is officially getting into hardware. The new team responsible for the project will be based in San Francisco.
We don't know what exactly this mysterious project is, but we have a couple of hints. In February, Midjourney hired Ahmad Abbas. He is an ex-Neuralink staffer who helped develop the Apple Vision Pro, Apple's mixed-reality headset. In addition, Midjourney CEO David Holz is no stranger to hardware: he co-founded Leap Motion, which built motion-tracking peripherals.
It's interesting to watch Midjourney go from a pretty geeky startup with a product available via Discord to a full-fledged IT company.
Let's hope their hardwired project doesn't fail.
Sharing is caring! Refer someone who recently started a learning journey in AI. Make them more productive and earn rewards!
Google Meet Now Supports AI Note-Taking Feature
Google has started rolling out new AI features to Meet for Google Workspace customers. Starting this week, users can transcribe their calls and summarize that discussion. If you've used Fireflies or Otter AI, you know exactly what we're talking about. The service also creates files in Google Docs and then attaches them to events in Calendar. At the moment, the feature only supports English.
On the surface, this story doesn't seem very important, but it's yet another reminder of how vulnerable some AI startups are. I'm afraid this announcement is bad news for many projects that are developing AI assistants for Meet and Zoom.
Anthropic Publishes System Prompts for Its AI Models
Anthropic has published system hints for its Claude family of models and promised to continue to do so. In doing so, the company promises to set a new standard of transparency for a rapidly evolving industry. System hints are rules or instructions that define how a model should respond to requests, describing exactly what it can and cannot respond to and the sentiment embodied in the output.
Anthropic claims this solution will help developers better understand how these often-mystified models actually work in practice.
A rather unconventional move on Anthropic's part. The most intriguing thing now is whether OpenAI, Google, Meta and other companies will adopt this practice.
Useful Tools ⚒️
Elisi – Boost Performance with AI coach
CompanyGPT 2.0 – Search engine for companies and people
Develyn – Open source AI agent for developer relations
Ultimaps - Create color-coded interactive maps in seconds
GPT Engineer – Chat with AI to build for the web
GPT Engineer is a no-code platform designed to help users build interactive web applications quickly by translating plain language into code. It allows for rapid prototyping and development, enabling users to create authentic web apps in seconds. The platform features collaborative tools, version control integration, and instant preview capabilities, making it user-friendly for both technical and non-technical individuals.
Share this post with friends, especially those interested in AI stories!
Weekly Guides 📕
Dreamina AI : Create Exactly What You Imagine
How To Make Viral MONETIZABLE Videos Using AI | Shorts & Reels
How to Start a Blog with AI and Automation in 2024
Soundraw: The Go-To AI Music Software for Musicians & Content Creators
AI Meme Of The Week 🤡
AI Tweet Of The Week
The full thread is here.
(Bonus) Materials 🏆
AI Enterprise Market Map: Cutting through the hype
Generative AI Transformed English Homework. Math Is Next
How AI Analytics Spurs Innovation at Newly Public Firms
China’s AI Engineers Are Secretly Accessing Banned Nvidia Chips
OpenAI expert Scott Aaronson on consciousness, quantum physics and AI safety
AI is changing the world, and Nvidia’s billionaire CEO Jensen Huang is banking on it
Share this edition with your friends!