Claude 3.5 Haiku AI Launches, New York Leads in Responsible AI, Australian SMBs Embrace AI, Microsoft Unveils Phi-4 AI Model, and more...
11:36 Share

Good morning. It's December 13th, and this is your daily brief in AI. Here's everything you need to know. On December 12th, 2024, Anthropic launched its latest AI model, Claude 3.5 Haiku, making it available globally through the Claude chatbot on both web and mobile applications.

This new model follows the introduction of CLAWD 3.5 Sonnet in October 2024 and boasts significant improvements, including lower latency and enhanced instruction following capabilities. CLAWD 3.5 Haiku has achieved a 40.6% score on the software engineering benchmark, surpassing both the earlier 3.5 Sonnet model and OpenAI's GPT-40.

It is now the fastest model in Anthropic's lineup, outperforming its predecessor, Claude 3 Opus, in various performance benchmarks. The model is accessible to all users, including those on free tier subscriptions, and has become the default option for this group without any specific announcements from Anthropic. Users can access Claude 3.5 Haiku for free, but their usage is capped at approximately 10 exchanges per day, depending on server demand.

For users requiring higher usage, the Claude Pro plan is available at $20 per month, offering up to five times the free tiers usage and priority access during peak hours. Claude 3.5 Haiku is particularly effective for enterprises, excelling in user-centric applications and creating tailored experiences from large datasets.

Potential applications for the model include data extraction, content moderation, personalization, and scalable chatbot development. Recent optimizations include support for Amazon Web Services' Tranium 2 AI chipset, enhancing its latency-optimized inference capabilities.

With a 200,000 token context window, CLAWD 3.5 Haiku can manage extensive inputs better than OpenAI's GPT-4, which has a limit of 128,000 tokens. However, it does not support image analysis or web browsing features, which limits its capabilities compared to some competing models.

Governor Kathy Hochul has made a strong statement regarding New York's role in the development of artificial intelligence. At the AI Summit New York, she emphasized the importance of AI for public good, economic growth, and social equity. Hochul's administration aims to position New York as a leader in AI technology, being the first governor to fully embrace its potential.

During her address, Hochul announced the introduction of two supercomputers, which are the most powerful public-use machines in the United States, located at the University at Buffalo and the University at Albany. This move is part of a broader investment strategy in technology infrastructure, including a historic $6 billion federal approval for Micron's semiconductor manufacturing expansion in upstate New York.

To further advance AI initiatives, Hochul launched the Empire AI Initiative, a $400 million 10-year partnership that focuses on leveraging AI for public good. This initiative will involve collaboration between the state, private sector, and academic institutions, guided by New York's first Emerging Tech Advisory Board.

The board, co-chaired by IBM CEO Arvind Krishna and Girls Who Code CEO Tarika Barrett, will develop recommendations to ensure equitable AI growth. Hochul also highlighted the need for ethical AI use and increased diversity in technology leadership, advocating for the inclusion of underrepresented groups in the AI sector.

Hochul noted that other states, such as California, are taking notice of New York's AI initiatives and encouraged them to adopt similar projects. She raised important questions about AI's potential to address critical issues like food insecurity, cancer therapies, and healthcare modernization. With confidence in New York's approach, she declared the state's intention to lead in AI innovation, positioning it as a model for others to follow.

In conclusion, Hochul urged industry leaders to prioritize ethical practices and workforce diversity, asserting that innovation must be responsible and inclusive. A recent Salesforce report highlights a remarkable trend among Australian small and medium businesses, or SMBS, where 88% of those utilizing artificial intelligence have reported an increase in revenue.

This surge in AI adoption is noteworthy, as 85% of Australian SMBS are actively experimenting with AI, surpassing the global average of 75%. The data clearly indicates that for these businesses, adopting AI is not merely about staying competitive. It is essential for thriving in a rapidly evolving market. However, challenges persist.

A significant 47% of SMBS express concerns about falling behind in AI adoption, while 79% believe that enhancing data quality could further boost their revenue. Despite their enthusiasm for AI, 73% of Australian SMBS face difficulties in keeping up with technological advancements, and 45% report lacking the time to master the necessary tools.

Security remains a critical issue, with 79% of Australian business leaders willing to pay a premium for trusted technology vendors. Adrian Tauzy from Salesforce ANZ emphasizes that companies embracing AI can gain a competitive edge through improved operating margins and enhanced customer engagement.

Globally, 78% of growing SMBS plan to increase their AI investments over the next year, highlighting a widening gap between those who adopt AI and those who do not. The key benefits of AI for Australian SMBS include enhanced revenue, productivity, and customer experience, with popular applications like content generation, automated chatbots, and personalized recommendations.

The report, based on feedback from 3,350 SMB leaders worldwide, including 150 from Australia, underscores the significant impact of AI on business operations. A notable success story is UrbanRest, an Australian flexible accommodation provider which achieved a staggering 1,400% revenue increase and a 40% reduction in field service visits by leveraging Salesforce's AI solutions.

Microsoft has unveiled FI4, a new AI language model set to enhance user experiences and transform industries within its Azure AI ecosystem. This model is designed for seamless integration with the Azure AI framework, allowing businesses to deploy it effectively for natural language processing and custom AI applications. Engineered for efficiency and speed, FI4 delivers precise outputs while minimizing the need for extensive computational resources.

Its optimized transformer architecture strikes a balance between scalability and modularity, ensuring effective performance even on constrained hardware. The introduction of smaller AI models like PHY4 signals a trend towards reduced environmental impact and cost-effectiveness, aligning with corporate sustainability initiatives.

As a lightweight alternative to larger models such as GPT-4, PHY-4 is particularly focused on completion tasks, including form autofill and text predictions. The future success of PHY-4 will depend on its adaptability to user needs and evolving market demands in the dynamic AI landscape. With the global AI market projected to reach $1.5 trillion by 2030, businesses are increasingly integrating AI solutions into their operations.

FI4's specialization enhances its effectiveness in customer service by providing domain-specific accuracy and managing nuanced queries. Microsoft emphasizes responsible AI development with FI4, implementing safeguards against biases and ethical concerns, though challenges in transparency and accountability remain.

The model's compact design supports sustainability goals by reducing energy consumption, making it particularly suitable for industries like healthcare and finance that require rapid insights and compliance. Looking ahead, Microsoft plans to expand FIFOR's capabilities with updates that will include multilingual support and extended APIS, enhancing its adaptability across various sectors.

Amazon is positioning itself as a frontrunner in the realm of artificial intelligence, with projections suggesting it could reach a market cap of $3 trillion by the year 2027, driven largely by its AI initiatives. The company has recently unveiled Rufus, a generative AI-powered shopping assistant aimed at improving product recommendations and boosting sales. This innovation aligns with Amazon's strategy to enhance customer experiences through advanced technology.

With a forward price-to-sales ratio of approximately 3.4, Amazon stands out as an appealing investment opportunity, especially when compared to tech giants like NVIDIA and Apple, as AI adoption accelerates across industries. Amazon's development of custom AI chips, Tranium and Inferentia offers cost-effective AI solutions tailored for budget-conscious cloud users, expanding its market reach significantly.

To further support third-party sellers, Amazon is creating generative AI tools for advertising, which is vital as it accounts for over 60% of the company's sales, thereby improving overall performance. The recent advancements in AI are transforming the tech landscape, paving the way for the rise of new trillion-dollar companies in the near future.

Over the past decade, Amazon has seen a notable 442% increase in annual revenue, indicating that current growth expectations may be conservative. Additionally, the company has successfully reduced fulfillment processing times by up to 25% using AI-powered robotics, enhancing efficiency and lowering operational costs, which positively impacts profitability.

In the last 18 months, Amazon Web Services, or AWS, has launched nearly double the AI features compared to its major cloud competitors, highlighting its dedication to AI innovation. As the largest cloud service provider globally, AWS offers extensive AI capabilities, allowing customers to develop AI models thanks to its vast data access.

Furthermore, Amazon Bedrock provides a comprehensive repository of AI models from leading companies, assisting customers in training and deploying their AI models.

This has been your daily Brief in AI. To read more about these stories, follow the links in the episode bio. You can also subscribe to these updates via email at www.brief.news. For more daily podcasts about the topics you love, visit brief.news forward slash podcasts. We'll be back Monday with everything you need to know.