GPT-4o - OpenAI combines audio, text, and vision in one model
More topics: Google I/O 2024 with many AI updates, and Humanoid Robot for 16,000 dollars?
Hi AI Enthusiasts,
Welcome to this week’s Magic AI news, where we bring you exciting updates from the world of Artificial Intelligence (AI) and Technology. OpenAI and Google have presented a lot of new things this week. Stay curious! 😎
This week’s Magic AI tool is an AI-powered scheduling tool. That AI tool can optimize your schedules for better productivity, and work-life balance. A must-know tool for anyone with a full schedule.
Let’s explore this week’s AI news together. 👇🏽
Top AI news of the week
💬 GPT-4o: OpenAI combines audio, text, and vision in one model
OpenAI presented the new model GPT-4o (omni) for ChatGPT. It combines audio, text, and vision capabilities in one model. Here are the new features:
- GPT-4o is available for all users in ChatGPT (free and paid)
- Lower latency and more cost-efficient because of the end-to-end approach (all-in-one model)
- It is possible to interrupt ChatGPT. More realistic conversations are possible.
- Different voices (sarcastic, serious, anxious) and languages possible
- Available via API
- New Desktop App (initially only for Mac)
Have you seen the presentation of GPT-4o on Monday? If not, you should watch the presentation in the following video.
Our thoughts
The demos are very impressive. It looks like ChatGPT is a real conversation partner. You can talk to ChatGPT in real time and interrupt it, just like in a normal conversation.
In addition, you can use ChatGPT as a learning partner to solve math or programming problems. The progress is impressive when you consider where we were a year ago. In our opinion, we’ll see this kind of models in many areas for example in call centers and education.
However, one thing also irritated us. Microsoft invests billions in OpenAI, and OpenAI uses an iPhone and a MacBook Pro for the presentation. Why no Microsoft products? 🤔
✍🏽 What is your opinion on GPT-4o? Impressive or not?
More information
- Hello GPT-4o - OpenAI website
🤖 Google I/O 2024 with many AI updates
Google presented enhancements across its Gemini and Gemma model family and a new video generation model.
Some updates in a brief overview:
- Quality improvements for Gemini Pro 1.5 with 1 million token context window
- Introducing Gemini 1.5 Flash with a 1 million token context window for high-frequency tasks
- New Gemini Pro 1.5 model with a 2 million token context window (Currently only with a waitlist)
- Launch of the open-source model Gemma 2 with 27B parameters in June
- New open-source vision-language model called PaliGemma
- New video generation model Veo, a competitor to OpenAI’s Sora
Want to know more about the updates? Then, you can watch the keynote in the following video:
Our thoughts
Google has presented a bunch of AI updates. Most of them are not available in Europe. Unfortunately!
We are excited to see how Google will integrate these updates into Google products. And above all, when these updates will be available, especially in Europe? We are fans of open-source models. For this reason, we also welcome the release of Gemma 2 in June.
More information
- Gemini 1.5 Pro updates, 1.5 Flash debut and 2 new Gemma models - Google Blog
- New generative media models and tools, built with and for creators - Google Blog
🦾 Humanoid robot for 16,000 dollars?
The Chinese robotics company Unitree Robotics has presented the Unitree G1, the first affordable humanoid all-purpose robot. The robot is equipped with state-of-the-art sensors and AI technology and can perform complex tasks in the home, care, and even in industrial environments - starting at USD 16,000.
The robot has a height of 1.27 m and a weight of about 35 kg. The following video shows the impressive capabilities of the robot:
Our thoughts
Wow, that’s an impressive price. We are excited to see how such robots will be used in industry and in our everyday lives in the future.
✍🏽 Would you use a humanoid all-purpose robot like Unitree G1?
More information
- Unitree G1 Humanoid agent AI avatar - Unitree website
Magic AI tool of the week
Do you ever have a full calendar and struggle to organize it? Yes, then an intelligent calendar scheduling tool can help you. With Reclaim.ai***, you can easily manage your schedule with AI.
This tool can automatically schedule meetings at the best time across your team. So, your team can stay focused on their most important work. In addition, Reclaim.ai also offers you many integrations for your favorite work tools like Slack, Zoom, or Jira. This AI tool makes time scheduling simple.
Articles of the week
- How to Deploy a Web App With Docker on Render for Free?
- Time Shifting in Pandas using the time series data from Tesla stock
- Responsible Development of an LLM Application + Best Practices
💡 Do you enjoy our content and want to read super-detailed articles about AI? If so, subscribe to our blog and get our popular data science cheat sheets for FREE.
Thanks for reading, and see you next time.
- Tinz Twins
P.S. Have a nice weekend! 😉😉
Leave a comment