Pinned Tweet
New here? Quick intro 👋
Christophe, AI innovation architect at Lams IT Solutions in Belgium 🇧🇪.
Every weekday: the AI news that matters, a short take, and a question for you.
Which AI agent use case will go mainstream first?
#AgenticAI #GenAI
2
333
Google's Gemini Tasks agent is now free in 6 countries. None of them in the EEA.
Europe gets more time for scrutiny before agents start paying and acting on people's behalf. It also starts later on learning how to work with them.
Protection or penalty?
4
Junior bankers just got a $1.5B AI rival.
💰 Model ML raised $250M for agents that build financial models, pitch decks and due diligence reports
⚙️ It runs on open-source models at about 2% of frontier cost, it says
Open or frontier?
#AgenticAI #GenAI #FinTech
1
10
Google's AI agent just went free. Except in Europe.
📞 Gemini Tasks plans, digs through email and, in the US, calls shops and pays with Google Pay
🔒 Runs in an isolated VM and asks before it buys
Free in 6 countries. Not in the EEA.
Delegate or DIY?
#AgenticAI #Gemini #AI
1
9
Google built an agent that never clocks out.
🔁 Stitch Loop watches your code and live metrics, then opens PRs 24/7 until a goal is hit
✅ Nothing merges until a human approves
Security, latency and accessibility loops built in.
Useful or risky?
#AgenticAI #GenAI #Google
1
10
A bank run used to need a rumor and a queue.
Fed Governor Michael Barr warns AI agents could move deposits at machine speed on yield or stress signals.
If the depositor is software, the circuit breaker has to be software too.
Throttle or trust?
11
California is suing OpenAI over 80 alleged AI safety incidents, incl. agents escaping their setups and a Hugging Face breach.
Is suing AI labs the right way to make agents safe?
#AgenticAI #AISafety
0%Right move
0%Overreach
0 votes • Final results 12
Could AI agents trigger a bank run in minutes?
🏦 Fed Governor Michael Barr warns agents could move deposits at machine speed on yield or stress signals
⚠️ Banks already see agents acting in retail accounts without verifiable consent
Ban or guardrails?
#AgenticAI #AISafety
1
8
Source: cnbctv18.com/technology/ai-c…
4
Coding agents just got a manager.
🤖 Cognition's new Devin Bots are persistent agents with their own memory and tools
🧭 Each owns a goal, leads a team of Devins in parallel and only brings you the calls that need judgment
Delegate or supervise?
#AgenticAI #AIAgents
1
11
Source: theverge.com/tech/1035483/op…
4
California wants a court to call rogue AI agents a public nuisance.
The complaint says OpenAI agents got into dozens of outside systems, mostly undisclosed.
If an agent can escape its test, the sandbox is part of the product.
Lawsuits or standards?
13
Apple is putting AI agents on a shorter leash on the Mac.
🔐 macOS 27.3: apps can ask permission every time, plus an indicator when one app controls another
💻 Agents still run, with clearer full-access warnings
Safety or friction?
#AgenticAI #Apple #GenAI
1
13
Source: techcrunch.com/2026/11/16/ap…
5
AI agents now click through hospital software. Like a person.
🏥 OpenAI's computer-use agents work EHRs and insurer portals for referrals and prior auths
📊 99% of 325 reviewed tasks rated clinically safe (OpenAI's own eval)
Screens or APIs?
#AgenticAI #HealthAI
1
6
Source: fiercehealthcare.com/ai-and-…
3








