xAI has released the weights and architecture for Grok-1. Grok-1 is a Mixture-of-Experts model with 314 billion parameters. The released model is a raw base model checkpoint from the Grok-1 pretraining phase, which means that it is not fine-tuned for any specific application. It was trained from scratch by xAI using a custom training stack.
Monday, March 18, 2024Sam Altman and Jony Ive, former Apple design chief, have officially joined forces to develop an AI-powered personal device that won't resemble a smartphone. They are currently seeking funding of up to $1 billion.
Saturday, March 9, 2024Researchers have used a generative AI tool called RFdiffusion to make completely new antibodies for the first time. The proof-of-principle work paves the way for the design of antibody drugs at the touch of a button. The technology is still a long way from reaching the clinic.
There have been a number of news stories about AI developers recently. While some developers are concerned that the technology may soon take over their jobs, these claims are likely just attempts at grabbing attention. AI may not remove the need for developers - it will likely just create the need for AI developers. The future of software development will likely still involve humans - the exaggerated claims made by startups today are just marketing.
The government of Saudi Arabia plans to invest around $40 billion into artificial intelligence. It is in talks to create a tech fund through a partnership with Andreessen Horowitz and other financiers. The fund would make Saudi Arabia the world's largest investor in artificial intelligence. It will back an array of tech startups tied to artificial intelligence, including chip makers, data centers, and AI companies.
OpenAI is expected to release a major AI model, possibly GPT-5, sometime in mid-2024, likely during the summer. The new model will likely be a multimodel large language model with similar capabilities to GPT-4, but better. OpenAI is still reportedly training the model, after which it will go through internal safety testing to identify any issues before public release. Other issues, besides those that may come up from testing, may delay the launch.
OpenAI's Voice Engine can create synthetic voices based on a 15-second clip of someone's voice. The text-to-voice generation platform is currently in limited access. Developers testing the platform are required to get explicit and informed consent from speakers, not build ways for individual users to create their own voices, and disclose to listeners that the voices are AI-generated. Example clips generated by Voice Engine are available in the article.
Microsoft is planning to build a $100 billion supercomputer called Stargate to power OpenAI's next generation of artificial intelligence systems. Stargate could take up several hundred acres of land and require up to 5 gigawatts of power. The facility could open as soon as 2028. Its creation is largely dependent on whether OpenAI can deliver its next major AI upgrade, which is reportedly due in early 2025.
Sam Altman and Jony Ive, former Apple design chief, have officially joined forces to develop an AI-powered personal device that won't resemble a smartphone. They are currently seeking funding of up to $1 billion.
Saturday, March 9, 2024Sam Altman and Jony Ive, former Apple design chief, have officially joined forces to develop an AI-powered personal device that won't resemble a smartphone. They are currently seeking funding of up to $1 billion.
Sam Altman and Jony Ive, former Apple design chief, have officially joined forces to develop an AI-powered personal device that won't resemble a smartphone. They are currently seeking funding of up to $1 billion.
Sam Altman and Jony Ive, former Apple design chief, have officially joined forces to develop an AI-powered personal device that won't resemble a smartphone. They are currently seeking funding of up to $1 billion.
Sam Altman and Jony Ive, former Apple design chief, have officially joined forces to develop an AI-powered personal device that won't resemble a smartphone. They are currently seeking funding of up to $1 billion.
Andrej Karpathy has put together a GPT-2 implementation in just 1,000 lines of code in a single file.
Sam Altman and Jony Ive, former Apple design chief, have officially joined forces to develop an AI-powered personal device that won't resemble a smartphone. They are currently seeking funding of up to $1 billion.
Andrej Karpathy has put together a GPT-2 implementation in just 1,000 lines of code in a single file.
Sam Altman and Jony Ive, former Apple design chief, have officially joined forces to develop an AI-powered personal device that won't resemble a smartphone. They are currently seeking funding of up to $1 billion.
Andrej Karpathy has put together a GPT-2 implementation in just 1,000 lines of code in a single file.
Sam Altman and Jony Ive, former Apple design chief, have officially joined forces to develop an AI-powered personal device that won't resemble a smartphone. They are currently seeking funding of up to $1 billion.
Andrej Karpathy has put together a GPT-2 implementation in just 1,000 lines of code in a single file.
Sam Altman and Jony Ive, former Apple design chief, have officially joined forces to develop an AI-powered personal device that won't resemble a smartphone. They are currently seeking funding of up to $1 billion.
Anthropic has trained three new models in the Claude 3 family, the strongest of which matches GPT4’s reported benchmark results. It is also a multimodal model and performs well on vision tasks. Importantly, Claude's coding ability is substantially improved with this release.
Tuesday, March 5, 2024OpenAI introduced a Read Aloud feature for ChatGPT, allowing it to verbally respond to users in 37 languages with five voice options and enhancing its multimodal capabilities and usability for on-the-go interactions across web and mobile platforms.
- Claude 3 successfully summarizes a video into a blog post, showcasing its long context capabilities.
Andrej Karpathy issued a long context challenge to make a blog post from a recent video of his. Claude 3 was able to perform this task with some data pre-processing help. The resulting blog post is high quality and interesting.
The influence of the printing press on the dissemination of information and the formation of nation-states mirrors the impact of the Internet, particularly through Aggregators like Google and Meta. However, the emergence of generative AI challenges the Aggregator model by providing singular answers that reflect embedded worldviews, potentially undermining the universal appeal that underpins Aggregator economics and suggesting a shift towards personalized AI to maintain their dominance.
Your personal AI pitch deck coach.
Adobe's Project Music GenAI Control is a platform designed to generate and customize music from text descriptions or reference melodies.
Cognition announced a new system called Devin that scores 14% on SWE-Bench, a challenging benchmark that tests AI’s abilities to write code. GPT-4 scored 1.7%. This model shows strong in-context learning abilities.
- Anthropic releases a prompt library for Claude 3, offering effective user prompts for various tasks.
The release of Claude 3 has been quite popular, but the prompting style for these models is slightly different. Anthropic has collected a set of user prompts that work well for a wide variety of tasks and topics.
Apple researchers have developed methods for training large language models on both text and images, leading to state-of-the-art performance in multimodal AI tasks.