Welcome to Get The Gist, where every weekday we share an easy-to-read summary of the latest and greatest developments in AI—news, innovations, and trends—all delivered in under 5 minutes! ⏱

In today’s edition:

• US & UK Decline to Sign Paris AI Declaration on 'Inclusive' AI

• Perplexity’s Sonar Rivals GPT-4o and Claude 3.5

• Google’s Gemini Deep Research Now Available on iPhone

• Google Patents AI Chatbot Capable of Answering Phone Calls

• And more AI news….

Top Developments

Paris AI Summit

US & UK Decline to Sign Paris AI Declaration on 'Inclusive' AI

The Gist: The US and UK have refused to sign a Paris summit declaration on “inclusive and sustainable” AI, citing concerns over national security and regulatory clarity.

Key Details:

• The declaration, backed by 60 countries—including France, China, and India—calls for AI to be “open, inclusive, transparent, ethical, safe, secure, and trustworthy.”

• The UK government said the agreement lacked clear guidelines on AI governance and security risks.

• US Vice President JD Vance criticized Europe’s AI regulations, warning they could stifle innovation, and cautioned against cooperation with China.

• UK officials denied following the US lead but acknowledged strategic alignment with American AI policy.

Perplexity

Perplexity’s Sonar Rivals GPT-4o and Claude 3.5

The Gist: AI search startup Perplexity has introduced Sonar, an in-house model now available to all Pro users, boasting performance on par with leading AI models.

Key Details:

• Sonar is based on Meta’s Llama 3.3 70B and runs on Cerebras Inference, which claims to be the world’s fastest AI inference engine.

• Perplexity claims Sonar outperforms GPT-4o mini and Claude 3.5 Haiku, while matching the performance of GPT-4o and Claude 3.5 Sonnet.

• Sonar Pro, designed for complex tasks, offers deeper context retention and more citations, with competitive pricing.

• The Sonar API is available in two variants: Sonar and Sonar Pro, marketed as one of the most affordable AI APIs.

Google Gemini

Google’s Gemini Deep Research Now Available on iPhone

The Gist: Google has expanded its Gemini Deep Research feature to iPhone users, enabling advanced AI-powered research for Gemini Advanced subscribers.

Key Details:

• Users can ask detailed research questions, and the AI compiles organized reports with sources.

• The research process takes about 5-10 minutes, and users receive a notification when it's done.

• Reports can be exported to Google Docs, and there are daily limits on research requests.

• It is currently powered by Gemini 1.5 Pro, but Google is expected to upgrade it to Gemini 2.0 Pro in the future.

Quick Gist

Microsoft AI launched Chain-of-Retrieval Augmented Generation (CoRAG), a new framework that enhances traditional RAG models through iterative query refinement. Read More

Google has patented a personalized AI chatbot capable of answering phone calls on users' behalf, utilizing on-device processing for quick responses and tailored interactions. Read More

Accenture and Google Cloud are partnering to enhance the adoption of cloud solutions and generative AI in Saudi Arabia. Read More

xAI plans to significantly increase its workforce by hiring thousands of "AI tutors" this year to enhance training for its chatbot Grok. Read More

Google DeepMind researchers introduced an improved distributed training approach called Streaming DiLoCo that significantly reduces bandwidth requirements while maintaining model quality. Read More

Apple is partnering with Alibaba to introduce AI features to the iPhone in China, aiming to boost sales amid declining demand and fierce local competition. Read More

Meta is in discussions to acquire South Korean AI chip startup FuriosaAI to bolster its custom chip initiatives amidst an ongoing NVIDIA GPU shortage. Read More

Positron secured funding from investors to scale production of its energy-efficient AI chips, which reportedly achieve greater power efficiency than Nvidia’s H100 GPUs. Read More

THAT’S IT FOR TODAY!