📰 Dev.to · Kevin

My new Symfony 8 + FrankenPHP + AI template ships with model failover, circuit breakers, and Claude Code guidelines. Zero to production in one command.

How I built a self-hosted RSS aggregator with AI categorization and smart alerts — and why rule-based fallback made it actually reliable.

OpenAI acquires Astral (makers of uv, ruff, and ty) to power Codex while shipping GPT-5.4 with 1M context. The AI coding wars are about owning the entire develo

Your Monday roundup: Nvidia GTC wrap-up, Mistral Small 4 unifies three models into one, Cursor's Composer 2 Kimi controversy, world model funding wave, and more

Remember when picking an LLM was a whole thing? GPT-4 versus Claude versus Gemini — benchmarks...

Two papers this week—Mamba-3 from Together AI/CMU/Princeton and AttnRes from MoonshotAI—signal a fundamental shift in how AI researchers are thinking about mode

Nobody wants to say it out loud, but the math stopped working. For two years, the dominant...

It happened quietly. Last week, a Meta engineer was debugging a technical question on an internal...

Document processing has been stuck in a binary choice for years: use traditional OCR for speed and...

Mamba-3 claims to beat Transformers on language modeling. Nvidia drops a technique to shrink LLM memory 20x. The foundational architecture of AI is genuinely in

The demos are gorgeous. An AI agent receives a vague task, spawns sub-agents, calls a dozen APIs,...

Jensen Huang Just Said $1 Trillion Like It Was Nothing — GTC 2026 Breakdown It's GTC week,...

Nvidia dropped a 120B hybrid architecture model, Qdrant raised $50M, swarm-native coding agents launched, and the open source world shipped K8s-style agent sche

There's a move happening in engineering teams right now that nobody's calling out loudly enough....

Atlassian laid off 1,600 people for AI. xAI scrapped its coding tool and hired from Cursor. Gumloop raised $50M for non-engineer automation. A lawyer warned abo

Last quarter, three different companies I know personally—not "a source told me," literally people...

What a week. If you blinked, you missed Meta snapping up a viral AI social network, Amazon quietly...

Atlassian cut 10% of its workforce this week to 'self-fund' AI investment. When the company that makes Jira starts feeding its employees into the AI machine, ev

When deploying large language models, selecting the right inference engine can save time and money....

Everyone's Building AI Agents. Almost None of Them Work. I spent last weekend doing...

Karpathy's autoresearch gives an AI agent a real LLM training setup and lets it run 100 experiments overnight on a single GPU. Nearly 20k GitHub stars in 4 days

The demo was perfect. A reasoning model, given a gnarly algorithmic problem, spent forty-five seconds...

From Karpathy's nine-nines math to the Claude Marketplace, this week's AI news all pointed to the same uncomfortable truth: reliability engineering is the hard

A weekly roundup of the biggest AI stories: Anthropic vs the Pentagon, GPT-5.4, Karpathy on why 90% reliability is nowhere near enough, and a KV cache breakthro