June 7, 2026
Headlines

Aaron Di Blasi

An example from the ARC Dataset, problems designed to test an AI system's ability to adapt to novel tasks. o3 scored 75.7% on the Semi-Private Evaluation set under the competition's $10k compute budget (around $20 per task) and 87.5% at high-compute configurations ($2000-$3000 per task).

Why o3 Will Not Take Your Job in 2025

Aaron Di Blasi25 mins

This article aims to set the record straight. By dissecting o3’s true strengths and weaknesses, we will demonstrate why this model, despite its impressive performance, does not pose an imminent threat to the workforce. Instead of succumbing to alarmist predictions, we’ll explore the nuances of AI’s role in enhancing, rather than replacing, human productivity.

Read More
A book cover titled "Top 20 Prompting Techniques" for the year 2024. The design features a dark background with diagonal lines and gradient colors ranging from pink to blue, adding a dynamic, modern aesthetic. The title text is prominently displayed in a mix of red and light blue fonts, creating a striking contrast against the dark background. The year "2024" is positioned at the bottom right corner in white, further standing out against the deep hues. The overall design is sleek and contemporary, suggesting a focus on modern techniques and innovative approaches.

Top 20 Prompting Techniques In Use Today: A Real LLM Prompting Guide For Professional Results Using an Interface or API

Aaron Di Blasi100 mins

Here, Aaron Di Blasi, Publisher for the AI-Weekly newsletter, offers a comprehensive guide on effective prompting techniques to help Large Language Models (LLMs) achieve professional results, whether you use an interface or an API. Below, Aaron outlines and describes in detail the Top 20 prompting techniques being used today to enhance LLM performance.

Read More
An image shows a network diagram with interconnected nodes on a dark blue gradient background, centered around a prominent white rectangle labeled 'LLM' with 'Large Language Model' written below it. Surrounding the central node are various icons connected by lines, representing different elements and applications related to LLMs. These include icons for a microphone, speech bubbles, documents, graphs, a toolbox, a globe, a neural network, a lightbulb, a book with a graduation cap, and a flowchart. Each icon is enclosed in a circular node, emphasizing the diverse functionalities and fields influenced by large language models.

Winners & Losers in the Age of Augmented LLMs

Aaron Di Blasi8 mins

We are beginning to witness a pivotal transformation in the AI landscape, where technology trends, market demands, and generative AI capabilities are reshaping the industry. This article delves into the winners and losers emerging from these advancements, highlighting the importance of addressing the core issue of hallucinations in LLMs through innovative augmentation techniques like Retrieval Augmented Generation (RAG) and Memory-Augmented Models (MoME). By categorizing use cases into information retrievers, automators, copilots, and coworkers, I outline a strategic framework for leveraging augmented AI to stay ahead in a competitive market, emphasizing the success of companies like Apple and Microsoft, while cautioning against the pitfalls of ignoring AI augmentation.

Read More