While managing my home network, I realized how many people are "hands-off" with their own security. Why should it be a tech-only task? 🤔 I don’t want my family needing me every time they manage network access. Imagine a network that's actually fun to manage—gamified guest access, learn-to-earn, TTL and lease control, IP masking, and secure DNS blocking! 🎮 With Project Sentinel, built on Elixir and a Raspberry Pi, you can control device access without sharing personal network details. Admin control is seamless, secure, and bloat-free. I’ll be tweeting about each feature and the problems I'm solving—working toward an open-source project anyone can use! 💡 Think less “DNS sinkhole” and ad-blocking, and more seamless management with a fun, user-friendly UI—all from a pocket-sized SBC plugged into your main router to manage your home network. #ProjectSentinel #Elixir #DNS #Networking #BuildingInPublic

Do people not care about the why anymore? Why a decision was made. Why something exists in a build. When you’re hands-on, you can’t avoid those questions. But a lot of people just generate, let the visual output speak for itself, and never ask why. The irony: we’re social creatures. We use social skills to prompt, yet skip the conversation about why in the moments that are supposed to be collaborative.

If you’re building a agent Always go the harness route first. Do not build a complicated 20 step workflow using a bunch of if else statements or langgraph Use a harness as your agent eg. Claude code, codex, deep agents, glean etc You’re just shooting urself in the foot off the bat Trust me, prompting works surprisingly well

🚀 Introducing DeepInfer, our in-house inference framework, now fully open source! Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable. 🔗 github.com/deepseek-ai/DeepI… 1/4

The dust build on my Mini network setup is a obvious indicator of how much AI has taken over my day to day. Time for a clean.

Nothing beats a terse response from your agent.

I’ve gotten so much value out of even just the $20 @ollama plan. Being able to experiment with the latest open-weight models without ever worrying about uptime is amazing. Plus my room no longer doubles as a sauna every time I want to try a new model.

Finally ready to rework a old game idea I had in a repo for over 9 years. With the tooling now along with some of the research, this is a good time to experiment and see the cost of making something a little more "serious" while learning and tuning my workflow.

Genuine question: how do people post so much on here? I know some of it is automation and bots. I’d just like to share interesting things and have real conversations that don’t end in a call-to-action or a sales pitch. How do you find the good conversations?

A read through im looking forward to checking out from Google. github.com/google-research/h… I'm sure some valuable insights, to help model a better workflow with models.