- unwind ai
- Posts
- Last Week in AI - A Weekly Unwind
Last Week in AI - A Weekly Unwind
From 31-March-2024 to 6-April-2024
Gargi Gupta & Shubham Saboo
April 07, 2024
It was yet another thrilling week in the AI field with advancements that further extend the limits of what can be achieved with AI.
Here are 10 AI breakthroughs that you can’t afford to miss 🧵👇
Hume AI has introduced a new conversational AI EVI, designed with emotional intelligence, with whom you can speak just like you’d to your human confidant. It goes beyond words to understand the emotions. Our voice carries a wealth of emotional information through nuances in tone, rhythm, and timbre, and EVI can recognize these subtleties in real-time. It speaks, pauses, stutters, and laughs just like a human!
OpenAI has announced Voice Engine, an AI model for voice cloning that uses a 15-second audio sample and text input to almost perfectly clone the voice. Voice Engine was first developed in late 2022. It is currently being tested for applications like non-readers and children, content translation, and improving essential service delivery in remote settings.
Microsoft and OpenAI are partnering to build a $100 billion supercomputer, called Stargate, to launch by 2028. This will involve erecting a series of supercomputer installations in the US over the next six years, to push the boundaries of what can be done with AI. This supercomputer will probably be 100x more expensive than the largest data centers currently in operation.
Pinecone has released the first-ever hallucination-free LLM, Luna, with 122B parameters, that admits when it does not know the answer to a question, essentially avoiding making up responses. However, it comes with significant limitations, that is its reduced functionality in tasks beyond simple QA, often responding with “I don’t know.”
DALL.E-3 now lets you edit generated images with simple text prompts within ChatGPT. It’s like the inpainting feature in Adobe’s Firefly where you can select an area of the image and put a prompt in natural language to make the desired edits. You can add, remove, or modify the objects in the selected area.
Apple has unveiled a new method ReALM for improving how AI understands references made during conversations and to items displayed on a screen or operating in the background. While ReALM models are significantly smaller in size, they outperform GPT-4 significantly.
This is significantly going to improve the user experience and functionality of assistants like Siri for on-device assistance.
Assembly AI has released a highly accurate, multilingual speech-to-text AI model Universal-1, trained on over 12.5 million hours of multilingual audio data. It is trained on four major languages: English, Spanish, French, and German, and shows a high speech-to-text accuracy in almost all conditions, including heavy background noise, accented speech, natural conversations, and changes in language. It beats OpenAI Whisper Large and NVIDIA Canary on several benchmarks.
Researchers at Anthropic have released a paper on how “many-shot jailbreaking,” that is, feeding the LLM a sequence of crafted dialogues, can bypass the AI’s built-in safety checks. The simplicity of this method, its effectiveness, and its ability to scale to longer context windows make LLMs by companies like Anthropic itself, Google DeepMind, and OpenAI more vulnerable.
Cohere has released a new model Command R+ specially built for real-world enterprise use cases. With 128k tokens context window, Command R+ is optimized for advanced RAG to improve response accuracy and provides in-line citations that mitigate hallucinations. The model is competitive with more expensive models like Mistral Large and GPT-4 Turbo.
OpenAI has rolled out an array of new features for its fine-tuning API, which was launched for GPT-3.5 back in August 2023, alongside expanding its Custom Models program. These updates are designed to provide developers with even more tools to tailor AI behavior, improve accuracy, and optimize costs.
Which of the above AI development you are most excited about and why?
Tell us in the comments below ⬇️
That’s all for today 👋
Stay tuned for another week of innovation and discovery as AI continues to evolve at a staggering pace. Don’t miss out on the developments – join us next week for more insights into the AI revolution!
Click on the subscribe button and be part of the future, today!Subscribe
📣 Spread the Word: Think your friends and colleagues should be in the know? Click the ‘Share’ button and let them join this exciting adventure into the world of AI. Sharing knowledge is the first step towards innovation!
Reply