Elena' s AI Blog
  Browse by topic  · Browse by tag

All posts
AI Core▶
AI 157 Generative AI 35 Large Language Models 63 ChatGPT 17 Agentic AI 24 AI Development 27 AI Safety 16 AI Law 20 Ethics 11 Edge AI 1 Weekly AI Signals 61 Robots 2 AI Apps 41
Machine Learning▶
Machine Learning 15 Deep Learning 8 TensorFlow 14 CNN 1 Transfer Learning 1 Mixed Precision 1 Preprocessing 6 Classification 4 Regression 2 Recommender Systems 1 Kaggle 8 Data 2 NLP 8 Fine Tuning 2 Research 8
Developer Tools▶
Git 29 GitHub 13 CLI 8 Conda 2 Setup 14 Automation 15 Operating System 3 Mac 6 Docker 9 Performance 2
Content and Site▶
Blogging 28 SEO 9
Python Series
Personal and Academic▶
Life 39 History 2 PhD 2 Study 4
Other Tags▶
Security 8 MCP 3 Open-Source 3 Troubleshooting 2 Beginners 1 Email 1 Flask 1 OOP 1 claude-code 1 security 1

Composio Alternatives: A Map of the MCP Integration Landscape


Elena Daehnhardt ·

Composio is not the only way to connect AI agents to your tools, and 'Composio alternative' can mean two quite different things. Here is an honest map of the connector platforms that supply the tools and the MCP gateways that govern them, with a clear recommendation for each type of reader. Read more...

npx @modelcontextprotocol/inspector: npm error could not determine executable to run


Elena Daehnhardt ·

The documented npx command for the MCP Inspector broke on version 2.10.0 with 'could not determine executable to run'. Here is why npx gets confused when a package ships two bins, and the one-line workaround. Read more...

Composio MCP: connect Claude, Cursor and AI agents to 1,500+ toolkits


Elena Daehnhardt ·

Running one MCP server per service works until you have a dozen of them. Composio puts GitHub, Gmail, Notion, Slack and 1,500+ other toolkits behind a single MCP connection, handles authentication, and lets the agent discover tools on demand instead of loading them all into the context window. Here is how it works, how to set it up, and when it is worth the extra dependency. Read more...

MCP Error: Invalid outputSchema, JSON Schema Declares an Unsupported Dialect


Elena Daehnhardt ·

Every tool from every server built on the MCP TypeScript SDK gets rejected the moment it declares an outputSchema, because the SDK still emits JSON Schema draft-07 and strict clients only accept 2020-12. Here is the exact error, why it happens, and two workarounds I tested against SDK 1.30.1. Read more...

Operational Polish: Human Reports and Draft Preview Endpoints


Elena Daehnhardt ·

Operational polish for a working LangGraph pipeline: replacing raw JSON run reports with a human-readable summary, adding a FastAPI preview endpoint for artifacts, and sending Slack approval messages with an actual preview link instead of a truncated wall of text. Read more...

Build Your First MCP Tool: AI News Search and Newsletter in Python


Elena Daehnhardt ·

Build your first MCP-style Python tool with FastAPI, structured tool boundaries, and a repeatable newsletter generation workflow. Read more...

How to build a website with AI in 2026: tools for beginners and coders


Elena Daehnhardt ·

In 2026 you can describe a website in plain English and watch AI build it in minutes. I compare the AI website builders worth trying for beginners, the more advanced AI coding tools developers are using, and — more usefully — how to work out which one you actually need. Read more...

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


Elena Daehnhardt ·

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 ·

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 ·

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 ·

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 ·

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 ·

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 ·

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 ·

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 ·

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 ·

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 ·

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 ·

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 ·

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 ·

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 ·

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 ·

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 ·

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 ·

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 ·

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 ·

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...

1 2 3
Copyright © Elena Daehnhardt
Privacy and Terms