Greeting 
Community 
,
May 30th 2024, Here are the top news in the AI world from last week (and a bit), as always If we’ve skipped over something important, give us a shout! And don’t forget to cast your vote for the news story that caught your eye the most.
GPT-4o is an advanced model designed for more natural human-computer interaction. It can process a combination of text, audio, images, and videos, generating text, audio, and images in response. With response times similar to humans, it improves upon GPT-4 Turbo’s performance in English, non-English languages, and code. GPT-4o excels in vision and audio understanding, offering enhanced capabilities at a lower cost.
Google has released updates for Gemini and Gemma AI models. Gemini 1.5 Pro now has enhanced capabilities for translation, coding, and reasoning, while the new 1.5 Flash model is optimized for high-speed tasks. Both models support multimodal inputs and long context windows. New features in the Gemini API include video frame extraction and parallel function calling. Additionally, the Gemma family now includes PaliGemma for vision-language tasks and the powerful Gemma 27B model. A Gemini API Developer Competition has also been launched​
Slack is facing significant backlash over its AI training policy, which involves using customer data for training its AI models by default. Users discovered that Slack’s privacy principles allowed the platform to analyze shared data, including messages and files, without explicit user consent. This policy raised concerns about privacy and data security among users, prompting Slack to revise its privacy principles.
The updated policy states that Slack will not use customer data to train third-party AI models or share data with LLM (large language model) providers. However, users are still critical of the default opt-in nature of the policy and the lack of a clear and user-friendly opt-out process. Critics argue that the process for opting out is cumbersome, requiring workspace owners to email Slack’s customer experience team without a specified timeline for processing these requests​
In a recent interview, Alex Kendall, co-founder and CEO of Wayve, discussed the future of autonomous vehicles and the company’s recent achievements. Wayve, a UK-based startup, raised $1.05 billion in a Series C funding round led by SoftBank, NVIDIA, and Microsoft. This funding marks the largest AI investment in the UK to date. Wayve plans to license its autonomous driving model to automotive manufacturers and makers of new autonomous robots.
Kendall highlighted Wayve’s unique approach to autonomous driving, which relies on video and language data for training rather than traditional Lidar and radar systems. This method allows their AI to handle a wide range of real-world driving scenarios. The company has partnered with major retailers like Tesco and Ocado to gather diverse data, which is crucial for training their models. Additionally, Wayve is collaborating with several top global automakers to integrate their technology into various vehicle platforms.