Rumors of MetaGPT, FTC Investigates OpenAI, and Stability AI Equity Disputes
There's never a dull moment in the current AI hype cycle!
Created on July 14|Last edited on July 14
Comment
Meta Rumored to Launch Commercial LLM
Meta is reportedly preparing to launch a commercial version of LLaMA, their open-source large language model (LLM). Previously introduced in February as a foundational model for researchers, the upcoming commercial edition offers developers and businesses the chance to use and refine this AI technology.
The initial LLaMA models released by Meta were smaller, with parameter counts ranging from 7 to 65 billion. In comparison, GPT-3.5 by OpenAI features 175 billion parameters, while the rumored GPT-4 might even surpass 1 trillion. Though no definitive information has been shared, it's conceivable that Meta could have trained a larger model for commercial use.
Currently, this is pure speculation. It's feasible that Meta may have employed a 'mixture of experts' model, similar to the rumored architecture of GPT-4, or perhaps developed an entirely unique innovation. Given Meta's renowned dedication and talent in the field, the result could offer surprising performance.
FTC Investigates OpenAI
OpenAI, the organization behind the well-known ChatGPT AI assistant, is under investigation by the US Federal Trade Commission (FTC). The FTC is assessing allegations that OpenAI violated consumer protection laws, which may have exposed personal data and harmed reputations. The investigation will focus on OpenAI's risk management strategies surrounding its AI models, specifically its handling of potential false or misleading statements. The FTC's probe comes after recent instances of defamation involving ChatGPT, with the AI assistant making false statements about individuals. This regulatory scrutiny follows an earlier issue in Italy where OpenAI faced claims of breaching the EU's GDPR privacy regulations.
Stability AI Equity Disputes
Stability AI, the startup associated with the viral text-to-image generation system Stable Diffusion, is facing legal allegations of fraud and embezzlement. A lawsuit filed by Cyrus Hodes, who claims to be a co-founder, accuses CEO Emad Mostaque of misleading him into selling his 15% stake in the business for a mere $100. The lawsuit alleges that Mostaque withheld information about fundraising and a planned pivot into AI text-to-image generation, which led to a $100 million funding round and a $1 billion valuation soon after Hodes sold his stake. The suit also accuses Mostaque of destroying evidence, mishandling company funds, and focusing on other projects while promising to use AI for addressing the Covid-19 pandemic.
Add a comment
Tags: ML News
Iterate on AI agents and models faster. Try Weights & Biases today.