Table of Contents

• This Week

• Business Use Cases

• Things to Pay Attention to

This Week

Bottom Line Up Front

• Moonshot's Kimi K2 Thinking, a Chinese AI model, has reportedly outsmarted OpenAI's GPT-5 and Anthropic's Claude Sonnet 4.5 on several benchmarks, including Humanity's Last Exam.

• The model was trained for under $5 million, significantly less than the cost of training GPT-5 or Claude Sonnet 4.5, and is open source, available for developers to experiment with.

• Nvidia CEO Jensen Huang has stated that China is likely to win the AI race, citing advantages such as lower energy costs and looser regulations, which could shift the balance of power in the AI industry.

Business Use Cases

Why is this important

• The AI revolution is facing a significant challenge due to a power shortage, with Microsoft CEO Satya Nadella stating that the biggest issue is not computing power, but rather access to electricity.

• Tech giants like Google, Microsoft, AWS, and Meta are spending billions to build data centers, but the construction of these centers is being hindered by the lack of available power, with some experts warning of a potential 45-gigawatt shortage by 2028.

• To address the power shortage, companies are exploring alternative energy sources, including nuclear power, solar power, and even putting chips in orbit in space, powered by solar energy, with Amazon championing Small Modular Reactors and Google planning to restart a reactor in Iowa in 2029.

Bottom Line Up Front

• Demand for AI skills is increasing, with mentions of AI skills in job postings rising 16% in three months, despite a 27% year-over-year decline in overall tech hiring.

• The fastest-growing AI jobs focus on data-related tasks, such as data labeling, data annotation, and data analysis, with businesses looking for employees who can interpret AI output and integrate machine insights into business decisions.

• Key skills required for AI jobs include math, statistics, programming languages like Python, as well as human skills like curiosity, ability to learn fast, and creativity, with CEOs like Cormac Whelan and Anthony Kappus emphasizing the importance of pairing hard skills with uniquely human skills.

Bottom Line Up Front

• A recent academic study by researchers at Zhejiang and Columbia universities found that generative AI can drive increased revenue and improve productivity in a large retail operation.

• The study showed that AI implementation can help increase revenue, with sales rising by 16.3% and conversion rates increasing by 21.7% in some cases.

• The benefits of AI are not limited to companies selling AI, but can also accrue to companies that implement AI into their operations, with US companies that have done so outperforming their peers, according to Mark, head of AI and Data Science at HSBC.

Why is this important

• Define specific, measurable objectives upfront using SMART criteria to avoid vague visions and ensure clear goals.

• Prioritize data quality over quantity by investing in cleaning, validation, and exploratory data analysis to prevent poor-quality data from poisoning AI models.

• Start simple by using straightforward algorithms and scaling to complex models only when necessary to avoid overcomplicating the model and ensure explainability..

Bottom Line Up Front

• Public sector leaders should carefully evaluate AI pilot offers from vendors, ensuring they are designed to validate value, not just technology, and are linked to mission outcomes.

• To test-drive AI technologies safely, leaders should select AI-native vendors with comprehensive platforms, assess business-outcome measurements, and apply a "Pilot-to-Value" protocol that combines governance, platform choice, and measurement.

• The risk of lock-in is not just about clouds or models, but also about operating model and data gravity, and can be mitigated by architecting for multi-model compatibility, owning prompts, data, and evaluations, and pricing for optionality.

Bottom Line Up Front

• Demand for AI agent developers is rising as marketing agencies and brands adopt AI, with companies like Lloyds Banking Group, Publicis Groupe, and WPP hiring for these roles.

• Salaries for AI agent developers can range from $70,000 to $180,000, with companies competing with each other and Silicon Valley to attract top talent.

• AI agent developers are expected to be problem solvers and polymaths, with skills in building and tweaking tech, coordinating practical and technical aspects, and experience in marketing and tech, according to experts like Shannon Moorman and Wesley ter Haar.

Things to Pay Attention to

Bottom Line Up Front

• The UT School of Law has launched a database called the AI Opportunity Inventory to analyze the benefits of AI for public policy.

• The database, developed under the UT AI Innovation and Law Program, will collect data on AI use from government institutions, civil society organizations, academics, and AI labs.

• Researchers, including those from UT and Brown University, will analyze the data to understand the public impact of AI, with the goal of ensuring AI aligns with the rule of law and is used for the best possible purpose, as stated by Kevin Frazier, AI Innovation and Law Fellow at UT Law.

Bottom Line Up Front

• Brands and influencers are not including intellectual property clauses addressing AI in their deals, leaving them exposed to potential risks.

• Influencers are relying on pre-existing agreements that include language for usage rights under name, image, and likeness, but these may not be sufficient to address AI-related issues.

• Experts, including Joseph Lawlor and Steven Stein, recommend that brands push for transparency and guidelines around the use of AI in content creation to protect themselves and creators from potential copyright and IP issues.

Bottom Line Up Front

• Stacey Rosen, the new volunteer president of the American Heart Association (AHA), aims to address the organization's blind spot regarding women's health in cardiology.

• Rosen believes that the key to addressing this issue lies in rethinking clinical trials, utilizing technology such as AI, and including more women in research studies.

• The AHA and the field of cardiology have evolved, but there is still a significant gap in understanding and addressing women's heart health, with women constituting only about 30% of research participants in cardiovascular studies.

Bottom Line Up Front

• The term "slop" refers to mass-produced, low-quality content, including that generated by artificial intelligence (AI), and has been popularized by individuals such as "deepfates" and Simon Willison.

• The creation of "slop" is not a new phenomenon, as it has occurred with the introduction of new tools and technologies throughout history, including the movable-type printing press and the rise of cinema.

• While much of this mass-produced content is of poor quality, it can also lead to the development of new art forms and provide opportunities for creators to emerge, with examples including authors such as William Shakespeare and Samuel Taylor Coleridge, and filmmakers like Francis Ford Coppola and Martin Scorsese.