In a business as rapidly developing as AI, trying to keep pace can be a challenge. In the meantime, here is a summary of recent news in the world of machine learning and noteworthy research and experiments that may not have been discussed independently. This week, the leadership controversy surrounding the AI startup OpenAI was the talk of the town. The company’s Board ousted Sam Altman, the CEO and co-founder, supposedly over what they perceived as his commercialization of AI at the expense of safety. Following this, he was reinstated as CEO, mostly due to the efforts of Microsoft, a major investor in OpenAI, leading to the replacement of most of the original board members. This incident underscores the challenges faced by AI companies, as the pressure to tap into monetization-oriented sources of funding continues to grow. The high costs of training and creating AI models make it nearly impossible to avoid this fate. According to CNBC, training a large language model such as GPT-3, the predecessor to OpenAI’s flagship text-generating AI model, GPT-4, could cost over $4 million. This estimate does not include the expenses of hiring data scientists, AI experts, and software engineers, all of whom command high salaries. Many large AI labs have strategic agreements with public cloud providers, as compute has become more valuable than gold. OpenAI’s chief rival, Anthropic, has taken investments from Google and Amazon, while Cohere and are exclusively backed by Google Cloud. However, these investments carry a risk, as tech giants have their own agendas and significant influence over startups. OpenAI attempted to maintain some independence with a “capped-profit” structure but was vulnerable to the power of companies like Microsoft. It is unlikely that this situation will change without increased investments in public supercomputing resources or AI grant programs. AI startups, like most startups, are forced to give up control over their destinies if they want to grow. Other noteworthy AI stories from the past few days include OpenAI ensuring that it won’t destroy humanity, California’s Privacy Protection Agency preparing to place guardrails on AI, Google’s Bard AI chatbot answering questions about YouTube videos, X’s chatbot Grok set to launch for Premium+ subscribers, the release of a video generator by Stability AI, Anthropic’s release of Claude 2.1, questions about the forces that control the AI revolution, AI21 Labs raising cash, the machine learning models that help people, and research advancements involving neural networks.

By admin