AI
The Best MCP Servers in 2026: Community’s Top Picks for AI Coding Workflows
Last updated: July 27, 2026 Open Graph preview card for the punkpeye/awesome-mcp-servers GitHub repository, showing its title ‘A collection of MCP servers’…
Practical AI and machine learning for developers and businesses.
Practical artificial intelligence for developers and businesses — using LLMs, automating real workflows and shipping AI features without the hype.
AI
Last updated: July 27, 2026 Open Graph preview card for the punkpeye/awesome-mcp-servers GitHub repository, showing its title ‘A collection of MCP servers’…
AI
Last updated: July 13, 2026 Comparison graphic showing logos of Cursor AI alternatives including Windsurf, GitHub Copilot, Claude Code, and Blackbox AI…
AI
Last updated: July 9, 2026 By the end of this article, you will know how to author reusable Claude Code skills using…
AI
Last updated: July 7, 2026 AI-generated illustration representing AI tooling and connectivity, thematically tied to MCP server discovery and AI coding assistants.…
AI
Last updated: July 1, 2026 AI-First Business Solutions LinkedIn post about AI redefining pair programming Image: AI-First Business Solutions / LinkedIn Three…
AI
Last updated: June 27, 2026 Picture this: last Tuesday, a team using an AI coding agent merged 23 pull requests before lunch.…
AI
Last updated: June 23, 2026 Claude Help centre article header for creating custom skills, showing the Claude branding Image: Claude Help centre…
AI
Last updated: June 17, 2026 LinkedIn post header image for Peppe Silletti’s article on AI and the death of pair programming Image:…
AI
Last updated: June 9, 2026 Diagram illustrating the three-phase evolution from TuringBots to agentic SDLC orchestration across the software development lifecycle Image:…
No. Most practical AI today is built on top of existing models and APIs — solid engineering skills are enough to get started.
It changes the job more than it removes it. Developers who use AI to work faster and focus on judgement are in strong demand.
It depends on the provider and configuration. We cover how to choose tools and set them up so client data stays protected.
It is strong at language work with a human in the loop: drafting, summarising, translating, pulling structure out of messy text, and writing or reviewing code. It is weak wherever being confidently wrong is expensive — precise arithmetic, current facts nobody gave it, and anything needing accountability. Design around that split and the technology is genuinely useful.
Pick one repetitive, low-risk task that already eats time and try it there first: summarising enquiries, drafting first-pass copy, tidying data. Keep a human approving the output, measure whether it actually saved time, and only widen once it has. Starting with a strategy and a budget tends to produce a project; starting with one annoying task tends to produce a result.