Recent AI News
- OpenAI launches the GPT retailer — OpenAI has publicly launched their much-anticipated GPT retailer. The GPT retailer mirrors the model of an app retailer and has some cool functions already. To entry the shop, login to your ChatGPT account and click on the GPT retailer.
Some cool ones we preferred are the next:
Cartoonize your self — Turn your selfie right into a Pixar-style animation.
Consensus — An AI analysis assistant that may search a library of 200 million pages.
Books — Personalized guide recommender.
Your ideas matter to us! We’d love to listen to about your favorites. Feel free to share them within the remark part beneath
2. Autogen studio 2.0 — Microsoft has launched a brand new model of Autogen Studio, their multi-agent framework. This newest model, with its revamped consumer interface, stands as the primary no-code resolution that actually impresses, permitting anybody to effortlessly craft a number of brokers, every with their very own distinct persona, instruments, and targets.
3. Google releases a surprising Medical Conversational Diagnostic Agent (AMIE) — AMIE is a cutting-edge medical diagnostic agent designed to have a dialog with a affected person, ask related questions, and make a analysis of their situation.
AMIE is a LLM-based system skilled utilizing self-play. Self-play is a coaching methodology the place an AI system basically ‘practices’ by interacting with itself.
AMIE was examined towards human medical doctors on actual examples and “considerably” beat the medical doctors in accuracy of analysis (see chart beneath) whereas having extra significant, empathetic conversations with the sufferers.
While AMIE continues to be in R&D, it is a promising step in making healthcare cheaper and extra accessible for all.
4. Text to Video era– 2024 marks a major leap in AI developments with the introduction of text-to-video era, exemplified by ByteDance’s newest innovation, the MagicVideo-V2 mannequin.
See an instance beneath. At its core, MagicVideo-V2 employs a posh pipeline that melds text-to-image era, video movement creation, and body interpolation, producing outcomes which might be each spectacular and seamless. We are eager to see how this subject evolves this yr.
5. OpenAI launches new fashions and lowers pricing — Among different OpenAI information, the corporate launched 2 new embedding fashions — text_embedding-3-small and text-embedding-3-large .
They additionally launched a brand new GPT4 mannequin which is best at code era and a few new instruments to handle and monitor API utilization. The firm talked about they’re planning to launch a brand new model of their flagship gpt-3.5-turbo mannequin within the coming weeks whereas additionally slashing its worth by 50%! Go OpenAI!
6. Meta releases Code Llama 70B mannequin — Meta launched a brand new Code LLama mannequin that’s 70B in dimension. This mannequin is the most important open supply code era mannequin and might generate code in Python, Java, PHP and C++. Its code era accuracy is similar to GPT4 mannequin.
