What are AI agents? How they plan, use tools and take actions
AI agents go beyond chat: they plan, call tools, browse and act toward a goal. How the agent loop works, the main design patterns, real uses and the risks.
AI Desk3 min read
Page 3 of 9
Articles & Analysis — Page 3
Deep dives, analysis, features and opinion on artificial intelligence.
Knowledge distillation explained: how small AI models learn from big ones
Distillation trains a compact "student" model to imitate a large "teacher". How it works, why it powers today's small models, and its legal and quality limits.
AI Desk3 min read
LLM-as-a-judge explained: using AI models to grade AI output
Language models can score other models' answers at scale. How LLM-as-a-judge works, its known biases, and how to design rubrics and checks that make it reliable.
AI Desk3 min read
10 open-source AI projects every developer should know
From Hugging Face Transformers to llama.cpp, vLLM, Ollama, LangChain and ComfyUI: the open-source projects that form the backbone of today's AI development.
AI Desk3 min read
How to evaluate an open-source AI project before you depend on it
Stars aren't enough. Check maintenance, community, license, security, documentation and governance before building on an open-source AI library or model.
AI Desk3 min read
GPU vs TPU vs NPU: the AI chips explained
Why AI runs on GPUs, what Google's TPUs and the NPUs in laptops and phones do differently, and how memory, bandwidth and interconnects shape AI performance.
AI Desk3 min read
Why NVIDIA dominates AI: chips, CUDA and the software moat
NVIDIA's lead in AI rests on more than fast GPUs: the CUDA software ecosystem, networking, full systems and developer habits. How the moat formed and its limits.
AI Desk3 min read
How much energy does AI use? What we actually know
Data centers, training runs and everyday queries: what the International Energy Agency and researchers say about AI's electricity use, and why estimates vary so much.
AI Desk3 min read
Liquid cooling explained: why AI data centers are moving beyond air
Dense AI racks produce far more heat than traditional servers. How direct-to-chip and immersion cooling work, their benefits for energy and water, and the trade-offs.
AI Desk3 min read
What is an NPU? AI PCs and on-device AI explained
Neural processing units let laptops and phones run AI locally. What NPUs do, what "40 TOPS" means, which features use them and whether you need an AI PC.
AI Desk3 min read
Humanoid robots explained: why companies are betting on human-shaped machines
Why humanoid robots are attracting huge investment, what they can and cannot do today, the role of AI foundation models, and the hurdles before they reach homes.
AI Desk3 min read
How foundation models are changing robotics: vision-language-action models
Robots that follow instructions like "pick up the extinct animal": how vision-language-action models such as RT-2 connect web knowledge to physical actions.
AI Desk3 min read
The SAE levels of driving automation, explained (0 to 5)
What Level 2, Level 3 and Level 4 really mean, who is responsible for driving at each level, and why "self-driving" marketing often confuses drivers.
AI Desk3 min read
How self-driving cars see the world: cameras, lidar and radar
The sensors behind autonomous vehicles, what each does best, how AI fuses them into a picture of the road, and the debate between camera-only and multi-sensor designs.
AI Desk3 min read
The EU AI Act explained: risk levels, obligations and deadlines
The world's first comprehensive AI law in plain English: banned practices, high-risk systems, general-purpose model rules, transparency duties, fines and timelines.
AI Desk3 min read
How the world regulates AI: the EU, US, China, UK and beyond
How the EU, US, China, the UK and global bodies regulate AI — from comprehensive laws to sector rules and voluntary frameworks — and what it means for companies.
AI Desk3 min read
What is AI alignment? The problem of making AI do what we intend
Why it is hard to make AI systems pursue the goals we actually want, the main techniques used today — from RLHF to constitutional AI — and the open research problems.
AI Desk3 min read
Frontier AI safety frameworks explained: how labs decide a model is safe to release
Responsible scaling policies, preparedness frameworks and dangerous-capability evaluations: how leading AI labs test models for catastrophic risks before release.
AI Desk3 min read