Extra, extra: on May 14 I will be speaking on AI Security and release my agentic security scanner, VASO, at The AI Security Summit, London https://aisecuritysummit.com/events/london-2026 Come and say hi 😊

Cursor to start offering security code reviews. Everybody is jumping on the bandwagon I guess. Is good everyone gets access to security reviews, let’s hope software gets better. 🤔 #appsec #cybersecurity

My voice to agent interface is working, prototype 0.1. Now I can speak and command my agents. This so amazing! 🤩 #AgenticAI #AI

Claude Code is making so many basic mistakes, mixing data, repeating itself, etc. The model degradation is real.

So how many Claude Code or Codex instances you run in parallel? I usually go between 2 and 4 😆

May the 4 be with you 😊 #starwars #starwarsday

Hermes with gpt-5.5 is a killer. IMO, ChatGPT subscription is the way to go with agents OpenClaw, Hermes, etc. works like a charm!

OpenAI vs Anthropic, Codex vs Claude Code, OpenClaw vs Hermes, Frontier models vs Open Source models. Don’t choose, learn all and use them wisely. Use these technologies to improve your life & work. Just saying…

Vibe coded a voice interface for my agents (agnostic) using Hermes and codex 5.5 over the weekend. Really impressive! Codex is slower than Claude but builds way better!! #codex #openai #hermes #agenticAI #AI