A damning analysis by GPTZero revealed that 53 papers accepted to NeurIPS 2025, the world’s most prestigious AI conference, contained over 100 hallucinated citations.

Light vs. heavy agents miss the real axis: durability vs. disposability. I was thinking about this while reading the latest post of Hamza from ZenML. He makes a compelling distinction between Light Agents (chat wrappers) and Heavy Agents (orchestrated workflows).

1/ 🚨 Hot take: We’re not early on AI anymore. 10% of humanity already uses ChatGPT regularly. 2026 is year 4 of the AI revolution. And that changes everything. 👇

AI feels like the 19th century railroad boom. We're overbuilding infrastructure (datacenters, energy, etc) ahead of real demand. Many projects will fail. But just like those "redundant" rail lines later became London's metro, today's excess may quietly power tomorrow's society

More than 200k banking jobs in Europe are at risk by 2030. Morgan Stanley estimates a 10% workforce reduction due to AI. The cuts will affect central services, including back-office, middle-office, risk, and compliance. What's your plan to stay relevant professionally in 2026?

Last night, we hosted a Rasa x Rime event in San Francisco at Continue and focused on one question: Why does Voice AI still feel frustrating in 2025, despite all the progress and all the investment? Here is what I learned at the event: medium.com/@profrodai/v...

Last week in Paris we debated voice stacks. Conclusion: “Open vs Commercial” is the wrong question. The real constraints are engineering ones. What’s changing 👇 1️⃣ Latency ≠ tokens/sec It’s jitter, interruption & VAD. Graceful barge-in is the hardest problem in Voice AI.

1/ Most voice AI apps fail. And they don’t just fail on hard tasks. They fail on normal human speech. Try saying a credit card number naturally: "My card ends in four five three two." Watch how fast things fall apart. --- #VoiceAI #Rasa #RimeAI #Deepgram #AgentEngineering

Automate -> Dominate -> Deactivate -> Meditate History doesn’t repeat, it just reboots. 💡 Follow Prof.Rod for the most relatable (and educational) AI content on the internet. Inspired from MonkeyUser

1/ The next AI Agents London Meetup is happening on October 15th! 🚀 After what we saw at the last one, you definitely don't want to miss it. RSVP: www.meetup.com/london-agent...

Is Standard Agent all we need for agentic AI? I'm hosting a live AMA with Rishikesh Ajit (Founding Engineer @ Jentic, maintainer of Standard Agent). We’ll walk the repo, run code, and show how you can onboard + contribute. 📅 Oct 2 · 4pm IST 🔗 RSVP: luma.com/hplglc1u

Agentic Happy Hours after MCP Dev Summit 🍻 @Apify + @jenticai hosting a casual pub meetup. Limited space. 📅 Thursday, Oct 2 🕡 6:30pm-10:00pm 📍 Ye Olde London, London EC4M RSVP: luma.com/MCP-Dev-Summ...

1/ Do agents really need evals? That was the question I asked Hamza Tahir, CTO of ZenML, in our first live webinar. The internet treated it as binary. The reality is more nuanced. 🧵

🚀 London, we’re live! On Sept 17, join us + Xcede for the 1st London Agentic AI Meet-Up. Speakers from Microsoft + Tomoro AI. 📍 111 Buckingham Palace Rd 🕕 6pm doors, 6:30 talks AI agents, best practices, networking. Seats are limited! RSVP: www.meetup.com/london-agent...

🔥 Want to know what it takes to build AI agents and raise millions for your startup? This Thu, Aug 28 (16:00 UTC+1) join our Ask Me Anything with @seanblanchfield.bsky.social, CEO of Jentic (€4M pre-seed) moderated by @rodriveraai.bsky.social Sign Up Link Below! ⬇️ #AI #AIAgents #Startups #AMA

1/ I just published my first article on Dev.to 🎉 "Stop Building Brittle Agent Workflows" why AI agents break when chaining API calls, and how the Arazzo Engine makes workflows deterministic and reliable. 🧵

1/ Evals discourse is loud right now: "Evals or you're shipping vibes." "Ship, monitor, A/B – evals slow you down." Both are right in context. What matters is systematic measurement of quality, sized to your stage.

Starting soon - 16:00 UTC+1! AMA with Michael Cordner on building agents that are simple, open, and actually useful! Jump in + get the ping: luma.com/et0is0mj