Welcome back to Daily Zaps, your regularly-scheduled dose of AI news ⚡️ 

Here’s what we got for ya today:

• 👯‍♀️ 1 Worker + AI = 2 Workers

• 💸 UAE pledges $1.4 trillion investment in US

• 💿 KBLaM: Knowledge Base augmented Language Model

• 🪖 US DoD spending less on AI in 2025 budget

Let’s get right into it!

Sign Up | Advertise | Tools

RESEARCH

1 Worker + AI = 2 Workers

A study at Procter & Gamble involving one-day workshops with 776 professionals found that AI can significantly enhance performance in product development. Teams with access to GPT-4 or GPT-4o outperformed individuals without AI by 40%, while individuals using AI improved their results by 37%, matching the performance of non-AI teams. AI-supported teams delivered the highest quality overall and were three times more likely to produce top-tier solutions.

The technology also helped bridge expertise gaps between technical and commercial staff, especially benefiting less experienced employees. Participants using AI reported more positive emotions, and most AI-generated content was retained. However, the study's short timeframe and reliance on chat-based AI may limit its relevance to real-world, long-term corporate settings, where questions remain about AI’s impact on skill development, process integration, and sustained effectiveness.

GLOBAL NEWS

UAE pledges $1.4 trillion investment in US

The United Arab Emirates has pledged $1.4 trillion over the next decade to boost its investments in U.S. sectors such as AI infrastructure, semiconductors, energy, and manufacturing, building on its existing $1 trillion already invested. The announcement follows UAE National Security Advisor Sheikh Tahnoon bin Zayed’s meetings with U.S. leaders and tech executives, including Trump, Satya Nadella, Jensen Huang, Jeff Bezos, and Larry Fink. This massive commitment includes plans like expanding U.S. aluminum production and mirrors similar investment expectations Trump has expressed toward Saudi Arabia.

BIG TECH

KBLaM: Knowledge Base augmented Language Model

Large language models (LLMs) excel in language tasks but face challenges in integrating external knowledge efficiently. Traditional methods like fine-tuning and Retrieval-Augmented Generation (RAG) have drawbacks—fine-tuning is costly, and RAG adds complexity with external retrieval components. In-context learning, while simple, becomes computationally inefficient with large knowledge bases. To overcome these limitations, Microsoft introduces KBLaM (Knowledge Base-Augmented Language Model), a novel approach that integrates structured knowledge directly into LLMs using a scalable, efficient mechanism called rectangular attention.

This method encodes structured triples (entity, property, value) into key-value vectors and embeds them within the model’s attention layers, enabling dynamic retrieval during inference with linear scaling in memory and computation. KBLaM supports over 10,000 knowledge triples on a single GPU, updates knowledge without retraining, and reduces hallucinations by refusing to answer when information is missing. It offers improved interpretability, lower latency, and better memory efficiency than RAG or traditional in-context learning.

GOVERNMENT

US DoD spending less on AI in 2025 budget

In fiscal year 2025, the U.S. government is operating under a full-year Continuing Resolution (CR), maintaining funding at prior-year levels due to the absence of new appropriations legislation. This CR includes adjustments such as a $6 billion increase in defense spending compared to fiscal year 2024 enacted levels. Concurrently, the National Defense Authorization Act (NDAA) for FY 2025 authorizes $895.2 billion for the Department of Defense, reflecting a modest 1% increase over the previous year.

Despite these overall increases, the NDAA allocates approximately $141 billion for Research, Development, Test, and Evaluation (RDT&E) accounts, which is $2 billion less than the administration's request and $7 billion less than the prior year's allocation. This reduction in RDT&E funding suggests a potential decrease in resources available for artificial intelligence (AI) initiatives within the defense sector, as AI projects often rely heavily on RDT&E investments. However, the NDAA continues to prioritize AI, emphasizing its role in strengthening defense and cybersecurity operations.