In the video, several important developments in the field of artificial intelligence were highlighted. Firstly, Apple’s involvement with OpenAI was discussed, with the news that Apple would be getting a board observer seat on OpenAI’s board of directors. This move was seen as significant given Apple’s previous distance from OpenAI, despite their use of OpenAI’s API for certain functions. The appointment of Phil Schiller, Apple’s former marketing chief, to the role was noted as a strategic move by Apple.
Next, Salesforce CEO Mark Benioff introduced Salesforce Einstein Tiny Giant, a 1 billion parameter model that excels in micro-model function calling. This development was seen as a step towards on-device AI processing, with a focus on privacy, security, and efficiency. The shift towards smaller, task-specific models orchestrated by a generalist model was highlighted as a potential future direction for AI computing.
Another noteworthy development was the release of Moshi by Open Science, a real-time multimodal model capable of listening and speaking with high accuracy. This model was compared to OpenAI’s delayed GPT-4 voice functionality, showcasing how Open Science had beaten OpenAI to market in this aspect. The features of Moshi, including its ability to understand emotions and speak with accents, were highlighted in the video.
The video also touched on the changing landscape of media production, with the introduction of 11 Labs’ famous voices feature that allows users to have historic personalities say custom messages. This innovation was seen as a glimpse into the future of media creation, where AI-generated content tailored to individual preferences could become more prevalent. Additionally, 11 Labs’ voice isolator model, capable of extracting clear voices from audio samples with background noise, was noted for its potential applications in various real-world scenarios.
Lastly, the video discussed security issues related to AI, particularly focusing on OpenAI’s recent security breaches. The concerns raised by a former OpenAI employee regarding internal vulnerabilities and the potential risks posed by foreign adversaries were highlighted. The importance of maintaining high security standards in AI development, especially in the context of potential AI secrets being compromised, was emphasized as a critical area for ongoing vigilance and improvement in the field.
