Daily AI News Summary (02/23/2024)
Welcome to today's SMMRY. The key highlights:
-
Intel & Microsoft $15 billion chip deal 💰
-
Google's new org focused on AI safety 🤖
-
AnyGPT: A major step towards AGI 🧠
🚀 Headlines & Launches
Microsoft and Intel strike a custom chip deal that could be worth billions (2 minute read)
Intel will be producing custom chips, designed by Microsoft for Microsoft, as part of a deal that Intel says is worth more than $15 billion.
Nvidia's AI-friendly hardware sends its market value soaring higher than Australia's GDP (4 minute read)
Nvidia has posted a record revenue figure of $US22.1 billion for the December quarter --- up 265 per cent on the prior year. With a new market value of $US1.94 trillion --- about $US25 billion higher than Australia's GDP!
Google DeepMind forms a new org focused on AI safety (2 minute read)
Google DeepMind has created a new AI Safety and Alignment organization, which includes an AGI safety team and other units working to incorporate safeguards into Google's AI systems.
ChatGPT went temporarily "insane" with unexpected outputs (4 minute read)
On Tuesday, ChatGPT users began reporting unexpected outputs from OpenAI's AI assistant, with reports of the AI assistant having "gone insane". OpenAI acknowledged the problem and fixed it by Wednesday afternoon.
🧠 Research & Innovation
AnyGPT: A major step towards AGI (2 minute read)
Researchers in Shanghai have developed AnyGPT, a groundbreaking new AI model that can understand and generate data across virtually any modality like text, speech, images and music while still achieving strong zero-shot performance.
Real-time image generation with SDXL Lightning (Demo)
A demo application using fal.realtime and the lightning fast SDXL API provided by fal for realtime image generation.
25+ LoRAs for Mistral that outperform GPT-4 (5 minute read)
Predibase have released LoRA Land, a collection of 25 fine-tuned Mistral-7b models that consistently outperform base models by 70% and GPT-4 by 4-15%, depending on the task.
🛠️🔥 Top Tools
Lexica's Aperture v4 - Generate near 4k resolution images.
Phind-70B: Closing the code quality gap with GPT-4 Turbo while running 4x faster
Stability AI - Stable Diffusion 3
Match Group Teams Up with OpenAI
Groq® - Hyperfast LLM running on custom built GPUs
LongRoPE: Extending LLM Context Window Beyond 2M Tokens
These are the main links from today's SMMRY. Read the full edition or subscribe to the daily newsletter here.
