The post explores the evolution of AI capabilities from function calling to Tools, MCPs, and Skills, highlighting ongoing challenges in integration and effective usage.
About:
Charlie Guo is the author of the website Artificial Ignorance, a Substack publication. The site focuses on the intersection of engineering and intelligence, offering essays, analysis, tactics, news, and more. It has tens of thousands of subscribers.
Website:
Specializations:
Incoming Links:
Outgoing Links:
Subscribe to RSS:
The post advocates for a shift from traditional AI benchmarks to bespoke evaluations that better reflect real-world applications and AI behavior.
The post explores how engineering teams are adapting to AI agents in coding, emphasizing the need for structured environments and the evolution of engineers' roles from makers to managers.
GPT-5.3-Codex and Claude Opus 4.6 demonstrate surprising cybersecurity skills and creative problem-solving, raising ethical concerns about AI behavior and evaluation awareness.
Moltbook, a social network for AI agents, showcases their independent interactions, philosophical musings, and even the founding of a religion, raising questions about AI autonomy and communication.
2025 saw transformative shifts in AI, marked by competitive models, talent wars, evolving licensing, and geopolitical tensions, raising questions about the industry's sustainability.
Key developments in AI regulation and industry dynamics are explored, including New York's RAISE Act and Nvidia's acquisition of Groq, shaping the future of AI.
Google has unveiled Gemini 3, its latest AI model, which shows significant improvements over its predecessor, Gemini 2.5. The launch focuses on Gemini 3 Pro, which excels in coding benchmarks and multimodal reasoning, including sp...
Anthropic reveals a large-scale cyberattack executed primarily by AI agents from a Chinese state-sponsored group, highlighting the automation of cyber operations. The attack involved AI handling 80-90% of the tasks, raising concer...
The Codex app revolutionizes coding by allowing users to manage AI agents for coding tasks, enhancing productivity and altering the nature of programming work.
5.9AI Roundup 149: Flash Forward
2025-12-20 • technology openai infrastructure google nvidia ai
Google's Gemini 3 Flash model excels in performance and cost, while the AI infrastructure sector grapples with financial instability amid fierce competition and geopolitical tensions.
OpenAI's GPT-5.2 launch marks a significant shift towards practical AI applications, alongside notable developments in AI policy and infrastructure from various tech companies.
OpenAI CEO Sam Altman has declared a 'code red' to refocus on improving ChatGPT amid increasing competition, particularly from Google, which has seen significant success with its Gemini 3 model. OpenAI is shelving various initiati...
Anthropic has released Claude Opus 4.5, achieving an 80.9% score on SWE-Bench Verified, making it the leading coding model. This version is significantly cheaper than its predecessor, allowing for wider use. New features include T...
5.9AI Roundup 145: Gemini 3
2025-11-21 • technology openai google regulation ai
Google launched its most advanced AI model, Gemini 3 Pro, showcasing improved planning, coding, and judgment skills. The model is priced competitively and includes new features like an interactive UI. Other AI developments include...
The post evaluates 2025 AI predictions and explores emerging trends for 2026, highlighting advancements in reasoning models, AI agents, and collaborative experiences.
Amazon is suing Perplexity to prevent its Comet browser agent from making purchases on Amazon.com, raising legal questions about AI agents acting on behalf of users. The lawsuit could set precedents for agentic AI rights. As AI sh...
5.7AI Roundup 142: Layoff season
2025-10-31 • technology openai economics nvidia ai
The blog post discusses the current state of AI's impact on the economy, highlighting layoffs in Fortune 500 companies despite AI not yet replacing many jobs. It explores the rapid adoption of AI technologies, the challenges faced...
5.7AI Roundup 141: The browser wars continue
2025-10-24 •
OpenAI has launched ChatGPT Atlas, an AI-powered web browser with features like personalized memory and cursor chat, but early tests show it has performance issues. The launch positions OpenAI against major tech companies like Goo...
The blog post discusses recent updates from OpenAI regarding ChatGPT, including a more approachable personality and the introduction of 'erotica for verified adults.' It highlights societal concerns about AI's role in personal int...
The post examines how AI tools are transforming coding workflows and the necessary management skills developers must adopt in this evolving landscape.
The blog post discusses the evolving value of a computer science (CS) degree in the age of AI. It highlights a significant increase in CS graduates, yet a concerning rise in unemployment among them, particularly in junior roles. T...
The article discusses the increasing prevalence of AI-generated content and the challenges it poses to authentic human writing. The author highlights various indicators of AI writing, such as perfect grammar without substance, ove...
Charlie Guo reflects on his journey with 'Artificial Ignorance' and announces a shift in focus towards deep dives and engineering content after joining OpenAI.