Insider's Edit: Anthropic and Musk Take Aim at OpenAI

Here are this week's top stories on AI Business:

Anthropic Unveils Business Friendly Claude 3 AI Models

OpenAI rival Anthropic unveiled its Claude 3 family of models, which is the startup’s first multimodal versions and geared towards addressing companies’ biggest generative AI concerns: cost, performance and hallucinations.

The startup, which boasts multi-billion dollar investments from Amazon and Google to take on the Microsoft-OpenAI juggernaut, unveiled three new models in its Claude 3 family: Haiku, Sonnet and Opus. These can accept and generate text and images.

The models show ascending levels of capability – Haiku, then Sonnet and Opus – as well as pricing. Notably, Anthropic’s technical paper on Claude 3 shows all three models beating OpenAI’s GPT-3.5 and Gemini 1.0 Pro in knowledge, reasoning, math, problem-solving, coding and multilingual math.

Opus beats even GPT-4 and Gemini Ultra – OpenAI’s and Google’s most advanced models, respectively - according to Anthropic. Opus exhibits “near-human levels of comprehension and fluency on complex tasks, leading the frontier of general intelligence,” Anthropic researchers wrote in a blog post.

Read more

Elon Musk Sues OpenAI to Save Humanity

Elon Musk is suing OpenAI for alleged breach of contract, claiming the ChatGPT creator has violated its founding pledge to be a nonprofit, which the tech billionaire said he funded and nurtured.

In a bombshell 46-page complaint, Musk lists OpenAI CEO Sam Altman and President Greg Brockman as defendants, contending that he was essentially defrauded by their decision to create a for-profit arm in which Microsoft has a 49% stake in exchange for at least $10 billion.

It is a peculiar lawsuit in that the beneficiary is “humanity” instead of the sole plaintiff, Musk, although he had to show he suffered financial harm for the complaint to make it to court. The lawsuit also exemplifies the rarefied group in which elite tech leaders belong; they can afford to litigate theoretical issues with purported planetary significance.

In the lawsuit filed with a California state court, Musk wants OpenAI to make its AI models open to the public and stop using its technology to benefit Microsoft and others. It also wants the court to determine that GPT-4 and Q* (Q star) constitute AGI - not OpenAI's new board - and thus would be outside the clutches of Microsoft. Musk also seeks monetary damages that he will then give to a charity.

Read more

Human-in-the-Loop: Mission Critical for AI Usage and Evaluation

Udo Sglavo, the vice president of advanced analytics at SAS, penned an opinion piece on Human in the Loop (HITL) evaluations.

"The HITL model, in its essence, recognizes and capitalizes on the distinctive strengths inherent in both machine intelligence and human intuition. It serves as a testament to the belief that the synergy between artificial and human intelligence not only elevates the quality of outcomes but also nurtures a profound sense of trust in the capabilities of AI systems," he writes.

Read the thought leadership piece

In other news

• Researchers created a new piece of malware to show how easily generative AI models like Gemini Pro and GPT-4 can be circumvented.
• Hugging Face unveiled the latest version of its code generation model StarCoder – enlisting the help of Nvidia to bring it to life.
• Google co-founder Sergey Brin said issues with images generated by its Gemini AI model were likely due to inadequate testing.
• And robotics startup Figure has signed a partnership with OpenAI to develop next-generation AI models for robots.

Stay updated on AI news and insights. Subscribe to the AI Business email newsletter!