🇬🇧 English Article: en.thedavestack.com/postgresql-18/
Check it out and let me know, what's your favorite new feature?
#PostgreSQL #Postgres #Database #DevOps #Performance #Security #Postgres18
🛡️ Enterprise-Grade Security: Native OAuth 2.0 support (great for SSO/MFA) and the official deprecation of MD5 auth, pushing everyone towards SCRAM.
I wrote a full deep-dive on these features for The Dave Stack, including a complete Docker lab so you can test them yourself.
And for Devs, it's packed with gems: 💻
• Native uuidv7() for faster, time-ordered primary keys.
• Virtual columns are now the default, saving disk space.
• More flexible RETURNING OLD/NEW clauses for easier auditing.
For anyone in Ops, the biggest win might be the Pain-Free Upgrades. 😌
Planner stats are now preserved across major upgrades. This means no more dreaded performance dips after an update. A massive operational improvement!
The PostgreSQL team officially launched v18 on Sep 25th, and it's a major update.
I put its headlining feature, Asynchronous I/O (AIO), to the test. The result on my own Docker lab: a 34.87% performance improvement. 🚀
Thread with more key features below. 👇
http://backlog.md, a Git-native "Project as Code" tool—a great alternative to Jira for keeping the dev flow.
Inspired by it, I built mcp-backlog-md: an MCP wrapper to connect it to an AI.
I wrote about the tool and my build process in my latest post.
👉https://www.thedavestack.com/backlog-md/
I tested Claude Code, Gemini CLI, and OpenCode to find the best AI agent for the terminal.
My verdict is in: one of them is in a league of its own.
Find out which one and why in my latest blog post.
Save yourself hours of testing 👇
https://en.thedavestack.com/coding-agents-my-experience/
AI tool fatigue is real. I'm not leaving VS Code, just want to enhance it.
Found the perfect extension: Cline. It adds powerful, controllable AI without replacing my setup. It's AI that truly works with me.
My deep-dive review:
en.thedavestack.com/cline-bot/
Of course, @ProtonMail has a great tool for this too. This got me thinking: what other email services have similar one-click unsubscribe features? Let me know! 👇
x.com/protonmail/s...
Are you a digital hoarder of newsletters? 😅 Gmail's new "Manage Subscriptions" feature is a game-changer. It lists all your subscriptions and lets you unsubscribe with a single click. A super easy way to declutter!
blog.google/products/gma...
The battle that's destroying engineering productivity: #Monorepo vs. #MultiRepo?
The harsh truth: most teams choose by habit and end up on the losing side, trapped in either dependency hell or complexity paralysis.
en.thedavestack.com/git-reposito...
#TeamMonorepo #TeamMultirepo
🤯 @Microsoft is rewriting the playbook for code editors with open-source AI in @code!
👉 https://en.thedavestack.com/microsoft-opens-up-ai-in-vs-code/
💬 Will you stick with your current editor, or does this VS Code move make you rethink your tools?
🚀 Google introduces Gemma 3n, and it wants to live (and see and hear) on your phone! 🤖 I've written an analysis of this new on-device AI and what it means for developers. The end of total API dependency for AI? 🤔 Full post on: https://www.thedavestack.com/gemma-3n/
#Gemma3n #AI #Google
Ever wished interacting with Kubernetes was more intuitive? Do you know how `kubectl-ai` leverages AI to understand natural language commands for K8s management? 🤖 Could this change how we work with clusters? 🤔
Find out more: https://en.thedavestack.com/kubectl-ai-simplify-kubernetes/
#Kubernete
Anthropic drops Claude Opus 4 & Sonnet 4! 🚀
Opus 4: "world's best coding model."
Sonnet 4: big leap in coding/reasoning.
New: extended thinking w/ tools, better memory, Claude Code GA, new API features.
The AI race is on! Who's next?
Details: https://www.anthropic.com/news/claude-4
#AI #Claude4
GitHub's Copilot: now an "AI coding agent"—beyond autocomplete, a true dev partner! Story: https://github.blog/news-insights/product-news/github-copilot-meet-the-new-coding-agent/
Your AI coding assistant experiences? How will this shape IT roles? Discuss! 💬
#GitHubCopilot #AICoding
Incredible! @GoogleDeepMind's AlphaEvolve (Gemini AI) 🤯 goes beyond optimization - it evolves new algorithms! Boosting innovation & improving Google's systems.
Thoughts? Follow for more updates!
https://deepmind.google/discover/blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-
Google's Gemini AI expands in Android: watches, cars, TVs, & Android XR (extended reality). 🤖⌚️🚗📺🕶️ AI embedding deeper in our tech. Cool, but how will this *really* change daily habits & our tech expectations?
Your take?
Post: https://shorturl.at/BQQ1F
#AI #Gemini
Do you know Dokploy? An impressive open-source tool that simplifies apps deployment using Docker and Traefik. If you're looking for more control and a Heroku/Vercel alternative, this one's for you.
Read more: https://shorturl.at/Q7L1j
🚀 Big news from Google Developers! Gemini 2.5 Pro (I/O ed) is out! Major coding boost: better front-end/UI, smarter code transformation & agentic workflows. Topping WebDev Arena! 💻 Access: Gemini API/Vertex AI. Details: https://shorturl.at/zPR7q
Hey everyone! 🎉
I just published my April 2025 AI Roundup, a deep dive into this month’s top AI releases.
📖 Read the full breakdown with links here: https://www.thedavestack.com/april-2025-ai-roundup/
Would love to hear which update you’re most excited about!
🧠✨ Microsoft just dropped Phi‑4‑reasoning, ‑plus & ‑mini—14 B/3.8 B open‑weight SLMs that beat huge LLMs on math, science & code. Blog: https://azure.microsoft.com/en-us/blog/one-year-of-phi-small-language-models-making-big-leaps-in-ai/
10/ Deep-dive article, install scripts, benchmarks → https://www.thedavestack.com/thudm-launches-glm-4-32b/ 🔄 reposts welcome!
9/ Both X & Bluesky just rolled out TikTok-style video tabs—next I’ll post a 30-sec demo there.
8/ Tip: add the 🧠 emoji + keywords “LLM” “AI” “open-source” so this note surfaces in ML Ranked custom feed.
7/ 32 k context window → giant RAG, full documents, audit logs.
6/ Local-run recipe: vLLM, 2×A6000 (48 GB) or GGUF Q4 on a 3090/4090 (~13 GB).
5/ Beats Llama-3-70B on code + GSM-8K without breaking your GPU budget.
4/ Supports structured tool calls out-of-the-box. Hook functions, feed schema, get JSON back.
3/ Trained on 15 T tokens + RLHF; bilingual zh-en by design.
