/// BLOG

LOG: BLOG

Ruminations on code, intelligence, and the architectural voids of the modern web.

2026-10-04

Human-in-the-Loop in Practice - How to Design Human Oversight of AI Agents

Too little oversight ends in incidents like Replit's deleted database; too much ends in clicking without reading - Claude Code users approve 93% of agent prompts. Four levels of autonomy, a risk and reversibility matrix, the approval card, escalations and AI Act Article 14 in architecture.

13 min

Coding Agents in a Development Team - How to Roll Out Claude Code and Cursor Without Losing Quality

2026-10-03

AI & Automation

Coding Agents in a Development Team - How to Roll Out Claude Code and Cursor Without Losing Quality

In METR's study, experienced developers were 19% slower with AI while feeling 20% faster, and DORA shows AI raises throughput but also delivery instability. How to roll out coding agents in a team: working models, metrics, AGENTS.md, tests as a gate, security and senior pushback.

13 min

Allegro and Amazon in the AI Era - How to Increase Product Visibility on Marketplaces

2026-10-01

AI & SEO

Allegro and Amazon in the AI Era - How to Increase Product Visibility on Marketplaces

Allegro's AI Assistant has been open to everyone since 7 May 2026, more than 300 million customers used Amazon's Rufus in 2025, and Amazon blocks ChatGPT's bots. How Allegro relevance works, how to prepare listings for AI assistants, why to combine a marketplace with your own store and how to keep one source of product data.

13 min

Context Engineering - How to Manage Model Context So Your AI Agent Doesn't Lose the Thread

2026-09-29

AI & Automation

Context Engineering - How to Manage Model Context So Your AI Agent Doesn't Lose the Thread

A Chroma study showed that all 18 models tested lose quality as context length grows, and in the Manus agent every output token comes with about 100 input tokens. What context engineering is, why long context hurts, the four techniques (write, select, compress, isolate) and a context budget in practice.

13 min

Google Discover in 2026 - How to Earn Feed Traffic When Search Sends Fewer Clicks

2026-09-28

AI & SEO

Google Discover in 2026 - How to Earn Feed Traffic When Search Sends Fewer Clicks

In 2025 Discover referrals to US publishers fell 29%, less than search (38%), and in February 2026 Google ran its first Discover-only update. How the feed works, what changed in 2026, technical requirements, content that gets shown, titles without clickbait and measurement in Search Console.

12 min

AI Video and Image Generation in Marketing - Tools, Costs and Copyright in 2026

2026-09-27

AI & Automation

AI Video and Image Generation in Marketing - Tools, Costs and Copyright in 2026

OpenAI shut down Sora less than a year after the Sora 2 launch, and Veo 3.1 now costs $0.05 to $0.40 per second of video. Which AI video and image tools work in 2026, what an accepted asset really costs, who owns the copyright, how to label content since August 2026 and how to build a process with human review.

13 min

2026-09-27

AI & Automation

Social Media Automation with AI - One Piece of Content, Many Channels, Your Brand Voice Intact

LinkedIn's "Seems like AI slop" button was clicked over a million times in three weeks, and posts copy-pasted from a chatbot lose around 40% of their views. How to build an n8n pipeline that turns one piece into content for LinkedIn, a newsletter and X, keeps your brand voice and goes through human approval.

13 min

2026-09-26

AI & SEO

Brand Reputation in AI - What to Do When ChatGPT Says Something Untrue About Your Company

In a July 2026 Searchable study, 93% of 1,704 brands had at least one fact wrong or missing in AI answers, and small businesses got fabrications far more often. Where models get untrue facts about a brand, how to audit them, five correction paths and what the law says when an error costs a contract.

13 min

TOTAL_RECORDS: 0110

PAGE 2/14

SYSTEM_STABLE