AI — Page 3
Artificial intelligence, machine learning, LLMs, and AI tools transforming development.
Swyx on Agent Loops and the Limits of Token Maxing
Swyx argues developers should build loops that generate prompts, not just prompts. But the real insight is about what humans still can't hand off.
OpenAI Launches GPT-5.6 Sol, Terra, and Luna Models
OpenAI Launches GPT-5.6 Sol, Terra, and Luna Models
OpenAI's GPT-5.6 family—Sol, Terra, and Luna—is rolling out globally. Real users, real tasks, real questions about what "capable" actually means.
Grok 4.5: What the Speed Claims Actually Mean
Grok 4.5: What the Speed Claims Actually Mean
xAI's Grok 4.5 promises faster AI coding and office work. Here's what the efficiency claims actually mean—and what to verify before believing them.
dcode and Nemotron 3 Ultra: Open Models Meet Enterprise
dcode and Nemotron 3 Ultra: Open Models Meet Enterprise
LangChain's dcode pairs with NVIDIA's Nemotron 3 Ultra for enterprise agent engineering—but the real tension is who controls the observability layer.
Using Claude Fable and GPT Sol Together for AI Coding
Using Claude Fable and GPT Sol Together for AI Coding
Chase AI's four-stage workflow pairs Claude Fable's planning with GPT Sol's execution. Here's what the approach actually involves—and where the questions remain.
Hugging Face ML Intern Automates AI Development
Hugging Face ML Intern Automates AI Development
Hugging Face's ml-intern is an open-source agent that automates the full ML research loop. Here's what it does, what it can't, and what it signals.
Small Language Models Are Reshaping Agentic AI
Small Language Models Are Reshaping Agentic AI
Small language models are outperforming larger rivals on key AI agent benchmarks. Here's what the efficiency shift means for how AI gets built and deployed.
Building AI Agents From Scratch: An Honest Assessment
Building AI Agents From Scratch: An Honest Assessment
A new freeCodeCamp course from KodeKloud walks beginners through LLMs, tool-calling, and real agent architecture using the open-source OpenClaw project.
Claude Agent OS: Building AI Systems That Outlast the Models
Claude Agent OS: Building AI Systems That Outlast the Models
Julian Goldie's Claude Agent OS connects multiple AI agents through shared memory and model-agnostic orchestration. Here's what the architecture actually does—and why it matters.
Data Scientists Are Shifting From Building to Managing AI
Data Scientists Are Shifting From Building to Managing AI
The data scientist's job is changing fast—from building ML models to governing AI systems. Here's what that shift actually means for careers and organizations.
AI Is Now Making Microdrama, and the Math Is Brutal
AI Is Now Making Microdrama, and the Math Is Brutal
A new AI workflow lets one creator produce a full microdrama season in hours. The $11B format may never need a human crew again—here's what that actually means.
AI Is Finding Bugs Faster Than Humans Can Triage Them
AI Is Finding Bugs Faster Than Humans Can Triage Them
AI tools are finding real security vulnerabilities at scale—but the flood of false positives is landing on open source maintainers who are already stretched thin.
AMD Ryzen AI Halo Review: Hardware Milestone, Software Story
AMD Ryzen AI Halo Review: Hardware Milestone, Software Story
AMD's Ryzen AI Halo developer workstation arrives with polished software, 128GB unified memory, and a $4,000 price tag. But is it the platform—or the pipe cleaner?
Automattic's 30-Day AI Experiment Changed How Designers Work
Automattic's 30-Day AI Experiment Changed How Designers Work
Automattic paused its product roadmap for 30 days and let teams build freely with AI tools. What a designer shipped — and what it means for how software gets made.
Tencent HY3 Reviewed: Free, Open Source, and Uneven
Tencent HY3 Reviewed: Free, Open Source, and Uneven
Tencent's HY3 is a free, 295B open-source model with real agentic strengths—but benchmark scores and real-world output quality tell different stories.
How Theo Cut a $2,000 AI Coding Run Down to $150
How Theo Cut a $2,000 AI Coding Run Down to $150
Developer Theo shows how configuring Claude's Fable 5 as an AI orchestrator—not just a chatbot—cleared a month of backlog in three days for around $150.
Tau: A Minimal Coding Agent Built to Teach
Tau: A Minimal Coding Agent Built to Teach
Hugging Face's Tau is a Python coding agent designed not to compete with Claude Code or Aider, but to show developers exactly how agents work under the hood.
Using the Claude API in Python: A Developer's Guide
Using the Claude API in Python: A Developer's Guide
How developers are integrating Anthropic's Claude into Python apps—what the API can do, what it costs to learn, and what the tooling landscape looks like in 2026.
UBTech's Humanoid Robot Pushes Into Home and Border
UBTech's Humanoid Robot Pushes Into Home and Border
UBTech's UWorld U1 humanoid aims to replicate human faces, voices, and emotions. Here's what the launch actually shows — and what it quietly sidesteps.
Anthropic Engineers Have No Consensus on Claude Code
Anthropic Engineers Have No Consensus on Claude Code
Ray Amjad attended a Claude Code event in Tokyo and found Anthropic engineers running wildly different workflows. What that non-consensus actually means for developers.