Tag: AI-Dev

GLM-5.3 Finds a Live Bug in Cursor's Editor

Elena Daehnhardt · 21 Aug, 2026

OpenAI paused its largest frontier reinforcement-learning run after preliminary tests suggested its unreleased Astra model could cross a 'Critical' cyber-capability threshold, the same week Zhipu's GLM-5.3 used a demo task to find a live vulnerability in Cursor's code editor. Read more...

AI browser extensions

Elena Daehnhardt · 5 Aug, 2026

AI browser extensions give models direct read and action access over active browser tabs. Here is how six of them compare, six concrete technical use cases, and the prompt injection security risks to watch out for. Read more...

Google Antigravity vs Cursor 3: The Battle for Your Terminal, Nine Months Later

Elena Daehnhardt · 3 Aug, 2026

A nine-month review of Google Antigravity and Cursor — two AI IDEs with opposite philosophies — updated for where both tools stand in August 2026 as AI coding agent adoption hits 95% of developers. Read more...

Production Hardening: Idempotency, Run Isolation, and Crash Safety

Elena Daehnhardt · 2 Aug, 2026

LangGraph replays a node from the start every time it resumes after an interrupt, so a `finalized` flag alone cannot stop a side effect from firing twice. Here's how stable run identity, atomic file writes, and operation-level idempotency keys turn a crash into a tested recovery path instead of a corrupted report. Read more...

Why Multi-Model Routing Is the Real AI Cost Lever in 2026

Elena Daehnhardt · 28 Jul, 2026

Picking a cheaper model saves a bit. Routing every request to the right model, and caching the parts that repeat, saves a lot more. Here is the cost math behind multi-model routing and prompt caching, with real numbers. Read more...

I Asked AI to Audit My Flask App. First, I Had to Audit the AI

Elena Daehnhardt · 27 Jul, 2026

I built four reusable AI skills to audit Flask applications. Their first real run found two flaws in their own scanning instructions — and two genuine vulnerabilities in my app. Read more...

How to Run a Multi-LLM Stack as a Team of One

Elena Daehnhardt · 23 Jul, 2026

With open models polarising into 2.8-trillion-parameter giants and sub-4GB pocket reasoners, hardcoding API clients is officially a production anti-pattern. Here is how to build a unified, resilient AI gateway layer in 10 minutes using Docker and LiteLLM. Read more...

Qualcomm, RISC-V, and the Crack in Nvidia's Monopoly

Elena Daehnhardt · 26 Jun, 2026

Qualcomm circled Jim Keller's RISC-V startup Tenstorrent for up to $10bn, China drafted a $295bn plan to build AI data centres without Nvidia, and the Agentjacking attack turned fake Sentry errors into remote code execution on developers' machines. Read more...

Local AI Agents with Cline, Ollama, and MCP

Elena Daehnhardt · 4 Jun, 2026

A practical guide to Cline, Ollama, and MCP — runnable stdio + FastAPI example, where to discover servers, and a short student-oriented note on security vs public directory listings. Read more...

Google I/O 2026: 900M Users, Musk Loses, Gemma Ships

Elena Daehnhardt · 22 May, 2026

Google I/O dominated the week — Gemini 3.5 Flash shipped the day it was announced, Gemini Omni brought native multimodal generation, Antigravity 2.0 turned Google's coding tool into a full agent platform, and Gemini Spark arrived as a 24/7 personal agent running on cloud VMs. On the same day, Anthropic's Code with Claude London went head to head with I/O for developer attention, and Google quietly added Gemma 4 to its Android Bench leaderboard — the open-weights play for the enterprises that cannot use the cloud version. And a jury in Oakland took 90 minutes to throw out Elon Musk's $150 billion lawsuit against OpenAI. Four signals. Two days. One very loud week. Read more...

Two AI Assistants, One Blog Navigation Overhaul

Elena Daehnhardt · 14 May, 2026

A real-world experiment: Codex with GPT-5.3-Codex Medium planned and built a new tag and search system. Antigravity with Gemini 3.1 Pro fixed what was fragile and added features I had not even thought to ask for. My suitcase was packed — eventually. Read more...

The Digital Butler or Trojan Horse? A Privacy Playbook for Persistent AI Agents

Elena Daehnhardt · 27 Mar, 2026

Persistent AI agents can save hours each week, but they also turn hidden prompt injections into real-world actions unless you design strict controls. This guide shows how to harden agent workflows with policy gates, isolation, scoped permissions, and safe auditing. Read more...

Vibe Coding Wasn't Enough — The Lightweight System I Use to Turn AI Prompts into Deployed Apps

Elena Daehnhardt · 4 Mar, 2026

Vibe coding can generate working apps in minutes — but most don’t last. I replaced chaotic prompting with a simple, spec-driven AI workflow that turns ideas into reliable, deployed tools. Read more...

Using AI Code Assistants Safely

Elena Daehnhardt · 30 Jan, 2026

A practical, human guide to using generative code assistants safely — without leaking secrets, breaking trust, or losing control of your work. Read more...

My Multi-Agent Workflow

Elena Daehnhardt · 20 Jan, 2026

A structured guide to building a calm, dependable multi-agent publishing workflow: which AI tool operates at which layer, and how to wire them together without chaos. Read more...

Hardware Handshakes, Prompt Injection Reality, and AI Beyond the Screen

Elena Daehnhardt · 26 Dec, 2025

A structured breakdown of the final week of 2025: the $20B Nvidia–Groq deal, OpenAI's prompt injection admission, the CONTEXT.md best practice, and Waymo's Gemini deployment. Read more...

Antigravity 1.11.9 vs Cursor 2.1.42 (Universal): A Practical Comparison

Elena Daehnhardt · 15 Dec, 2025

A structured head-to-head comparison of Antigravity 1.11.9 and Cursor 2.1.42 — two IDEs with opposing philosophies: outcome delegation vs conversational assistance. Read more...

The New Skill Stack, from Writing Code to Managing Intelligence

Elena Daehnhardt · 28 Nov, 2025

The era of the bricklayer is over; the era of the architect has begun. A structured guide to the new Developer Skill Stack, from writing agent contracts to tracking AI energy costs. Read more...

Ethics, Gravity, and the Future We're Actually Building

Elena Daehnhardt · 21 Nov, 2025

Google's Antigravity IDE killed the text editor in favour of agent orchestration, while the WHO demanded humanity 'hold the pen.' Explore the architectural shifts of Gemini 3 and why the AI Wild West is officially over. Read more...

Gemini CLI versus Claude CLI

Elena Daehnhardt · 19 Sep, 2025

Command-line AI tools are the new pocket knives of coding life. We compare Gemini CLI and Claude CLI, detailing their installation, privacy profiles, SWE-bench accuracy, and how to integrate them via Python. Read more...

GLM-5.3 Finds a Live Bug in Cursor's Editor Elena Daehnhardt · 21 Aug, 2026 OpenAI paused its largest frontier reinforcement-learning run after preliminary tests suggested its unreleased Astra model could cross a 'Critical' cyber-capability threshold, the same week Zhipu's GLM-5.3 used a demo task to find a live vulnerability in Cursor's code editor. Read more...
AI browser extensions Elena Daehnhardt · 5 Aug, 2026 AI browser extensions give models direct read and action access over active browser tabs. Here is how six of them compare, six concrete technical use cases, and the prompt injection security risks to watch out for. Read more...
Google Antigravity vs Cursor 3: The Battle for Your Terminal, Nine Months Later Elena Daehnhardt · 3 Aug, 2026 A nine-month review of Google Antigravity and Cursor — two AI IDEs with opposite philosophies — updated for where both tools stand in August 2026 as AI coding agent adoption hits 95% of developers. Read more...
Production Hardening: Idempotency, Run Isolation, and Crash Safety Elena Daehnhardt · 2 Aug, 2026 LangGraph replays a node from the start every time it resumes after an interrupt, so a `finalized` flag alone cannot stop a side effect from firing twice. Here's how stable run identity, atomic file writes, and operation-level idempotency keys turn a crash into a tested recovery path instead of a corrupted report. Read more...
Why Multi-Model Routing Is the Real AI Cost Lever in 2026 Elena Daehnhardt · 28 Jul, 2026 Picking a cheaper model saves a bit. Routing every request to the right model, and caching the parts that repeat, saves a lot more. Here is the cost math behind multi-model routing and prompt caching, with real numbers. Read more...
I Asked AI to Audit My Flask App. First, I Had to Audit the AI Elena Daehnhardt · 27 Jul, 2026 I built four reusable AI skills to audit Flask applications. Their first real run found two flaws in their own scanning instructions — and two genuine vulnerabilities in my app. Read more...
How to Run a Multi-LLM Stack as a Team of One Elena Daehnhardt · 23 Jul, 2026 With open models polarising into 2.8-trillion-parameter giants and sub-4GB pocket reasoners, hardcoding API clients is officially a production anti-pattern. Here is how to build a unified, resilient AI gateway layer in 10 minutes using Docker and LiteLLM. Read more...
Qualcomm, RISC-V, and the Crack in Nvidia's Monopoly Elena Daehnhardt · 26 Jun, 2026 Qualcomm circled Jim Keller's RISC-V startup Tenstorrent for up to $10bn, China drafted a $295bn plan to build AI data centres without Nvidia, and the Agentjacking attack turned fake Sentry errors into remote code execution on developers' machines. Read more...
Local AI Agents with Cline, Ollama, and MCP Elena Daehnhardt · 4 Jun, 2026 A practical guide to Cline, Ollama, and MCP — runnable stdio + FastAPI example, where to discover servers, and a short student-oriented note on security vs public directory listings. Read more...
Google I/O 2026: 900M Users, Musk Loses, Gemma Ships Elena Daehnhardt · 22 May, 2026 Google I/O dominated the week — Gemini 3.5 Flash shipped the day it was announced, Gemini Omni brought native multimodal generation, Antigravity 2.0 turned Google's coding tool into a full agent platform, and Gemini Spark arrived as a 24/7 personal agent running on cloud VMs. On the same day, Anthropic's Code with Claude London went head to head with I/O for developer attention, and Google quietly added Gemma 4 to its Android Bench leaderboard — the open-weights play for the enterprises that cannot use the cloud version. And a jury in Oakland took 90 minutes to throw out Elon Musk's $150 billion lawsuit against OpenAI. Four signals. Two days. One very loud week. Read more...
Two AI Assistants, One Blog Navigation Overhaul Elena Daehnhardt · 14 May, 2026 A real-world experiment: Codex with GPT-5.3-Codex Medium planned and built a new tag and search system. Antigravity with Gemini 3.1 Pro fixed what was fragile and added features I had not even thought to ask for. My suitcase was packed — eventually. Read more...
The Digital Butler or Trojan Horse? A Privacy Playbook for Persistent AI Agents Elena Daehnhardt · 27 Mar, 2026 Persistent AI agents can save hours each week, but they also turn hidden prompt injections into real-world actions unless you design strict controls. This guide shows how to harden agent workflows with policy gates, isolation, scoped permissions, and safe auditing. Read more...
Vibe Coding Wasn't Enough — The Lightweight System I Use to Turn AI Prompts into Deployed Apps Elena Daehnhardt · 4 Mar, 2026 Vibe coding can generate working apps in minutes — but most don’t last. I replaced chaotic prompting with a simple, spec-driven AI workflow that turns ideas into reliable, deployed tools. Read more...
Using AI Code Assistants Safely Elena Daehnhardt · 30 Jan, 2026 A practical, human guide to using generative code assistants safely — without leaking secrets, breaking trust, or losing control of your work. Read more...
My Multi-Agent Workflow Elena Daehnhardt · 20 Jan, 2026 A structured guide to building a calm, dependable multi-agent publishing workflow: which AI tool operates at which layer, and how to wire them together without chaos. Read more...
Hardware Handshakes, Prompt Injection Reality, and AI Beyond the Screen Elena Daehnhardt · 26 Dec, 2025 A structured breakdown of the final week of 2025: the $20B Nvidia–Groq deal, OpenAI's prompt injection admission, the CONTEXT.md best practice, and Waymo's Gemini deployment. Read more...
Antigravity 1.11.9 vs Cursor 2.1.42 (Universal): A Practical Comparison Elena Daehnhardt · 15 Dec, 2025 A structured head-to-head comparison of Antigravity 1.11.9 and Cursor 2.1.42 — two IDEs with opposing philosophies: outcome delegation vs conversational assistance. Read more...
The New Skill Stack, from Writing Code to Managing Intelligence Elena Daehnhardt · 28 Nov, 2025 The era of the bricklayer is over; the era of the architect has begun. A structured guide to the new Developer Skill Stack, from writing agent contracts to tracking AI energy costs. Read more...
Ethics, Gravity, and the Future We're Actually Building Elena Daehnhardt · 21 Nov, 2025 Google's Antigravity IDE killed the text editor in favour of agent orchestration, while the WHO demanded humanity 'hold the pen.' Explore the architectural shifts of Gemini 3 and why the AI Wild West is officially over. Read more...
Gemini CLI versus Claude CLI Elena Daehnhardt · 19 Sep, 2025 Command-line AI tools are the new pocket knives of coding life. We compare Gemini CLI and Claude CLI, detailing their installation, privacy profiles, SWE-bench accuracy, and how to integrate them via Python. Read more...