Currently at the hairdresser and I'm talking to my AI on WhatsApp which is fixing on my computer a deployment that broke in my CI/CD because of the push I made before leaving. But what a pleasure to be a developer in 2026.
SOFTWARE
-
Modern Python Cookbook: 818 Pages with 130+ Data Science Recipes
By
–
Modern Python Cookbook (818 pages, 130+ recipes, 3rd Edition): http://
amzn.to/47l2lz1 v/ @PacktDataML ——————
#DataScience #DataScientist #MachineLearning #ML #ComputationalScience #Simulation -
Me and my agents after 83 commits in a day
By
–
Me and my agents after I clocked 83 commits in a day
-
OpenFang: Full Operating System for Autonomous AI Agents
By
–
Someone built a full operating system for autonomous AI agents! OpenFang is an agent OS that runs agents for you on schedules instead of waiting for you to prompt them. Written entirely in Rust, compiles to a single 32MB binary. Here's the key difference: Traditional agent
-
xAI Releases Grok 4.20 Beta Models
By
–
xAI released 3 new Grok 4.20 models on their APIs – grok-4.20-beta-0309-non-reasoning
– grok-4.20-beta-0309-reasoning
– grok-4.20-multi-agent-beta-0309 "Grok 4.20 Beta is our newest flagship model with industry-leading speed and agentic tool calling capabilities. It combines -
New Tool Automates Backend Development for AI Agents
By
–
The biggest bottleneck in Agentic dev?
— Charly Wargnier (@DataChaz) 11 mars 2026
The backend.
@insforge_dev 2.0 is aiming to fix this 👀
→ Instantly turns Cursor/Claude into full-stack builders
→ Auto-generates your DBs, auth, storage
→ Provides a context-optimized layer for edge functions
Free & open-source 🧵 ↓ https://t.co/NCz8TU0QzgThe biggest bottleneck in Agentic dev? The backend. @insforge 2.0 is aiming to fix this → Instantly turns Cursor/Claude into full-stack builders
→ Auto-generates your DBs, auth, storage
→ Provides a context-optimized layer for edge functions Free & open-source ↓ -
10000 drones production requires only 4000 full-time employees
By
–
Pour fabriquer 10000 drones de ce genre il ne faut que 4000 employés à plein temps, 800 par site sur 5 sites, de la rigolade.
-
Infrastructure Behind AI Models: Deployment at Scale
By
–
Thank you. Models attract most of the attention, yet the ability to deploy and sustain them at scale tends to depend on the infrastructure behind the scenes.
-
Paged Attention: Kernel Level Overhead vs. System-Wide Throughput Gain
By
–
Good writeup. Here's an interesting fact about paged attention: It isn't free at the kernel level. As expected, the non-contiguous memory reads add roughly 20-26% overhead per attention kernel call. But the system-wide gain (2-4x throughput) overshadows that because you can
-
SSH Mode for Distributed ML Development Workflow
By
–
yes, solid work trending in a good direction, but almost all my work is across like 20 different machines (my local, my claw machine, my gpu machines). possibly they could add ssh mode, a bit like VS Code does (for the same reasons).