I think @JubbaOnJeans may enjoy this clip. Best coffee in Bangalore.

RT @EXM7777: f*ck it, i'm leaking my FULL notion library. you get access to all of my resources:.- over 200 prompts for ChatGPT, Claude and….

Hey @upekkhaBe this is totally spoiling the experience. Happening multiple times. I am talking about the camera people just coming from different angles and in the way.

Any suggestions on how to handle this?. We are getting a lot of automated Pull Requests on GitHub on bunch of our deliberately vulnerable applications. These applications were created so that folks can learn about applicaiton security. Not so that we get them patched!.

Just found out about Dire Straits - A Night in London concert. Some call the Sultans of Swing outro played by Mark Knopfler the greatest ever. If you find the magic starts at 7:20. Feeling euphoric while listening to it.

Image attribution - Gartner Research's Hype Cycle diagram made by Jeremy Kemp. Wikimedia.

I keep wondering where we are in the AI tech adoption curve.

RT @AlexGroberman_: I created my LinkedIn just 29 weeks ago. Absolute easiest social media algo to crack by a country mile. Since then, I….

6/The real insight: Your AI is only as secure as your weakest YAML file. #AI #CloudSecurity #AISecuirty #DevSecOps.

5/The loop continues:. - Secure AI needs secure cloud.- Secure cloud needs proper configuration.- Proper configuration needs expertise.- Expertise is scarce. So we automate with AI.Which needs to be securely configured.Back to the beginning.

4/The irony: AI can detect misconfigurations. But only if it's properly configured. Which requires understanding cloud security. Which brings us back to: AI security IS configuration security. Just with more layers, more complexity, more room for error.

3/Each AI service has its own settings. Azure OpenAI ≠ AWS Bedrock ≠ Google Vertex. Different IAM rules .Different network policies .Different compliance controls. One misconfigured setting exposes everything.

2/AI systems inherit these vulnerabilities. Plus AI adds new configuration layers. - Model endpoints need access controls.- Training data needs proper permissions.- API keys need secure storage.- Inference logs need encryption.

1/The AI Security Configuration Loop 🧵. We think AI security is about protecting the models. But AI runs on cloud infrastructure. Cloud infrastructure security is about configuration. Misconfigurations are the #1 cloud vulnerability.

Anyone here using @perplexity_ai comet browser?. I am on the waiting list. Wanted to know if any one has first hand experience with it?.

I use AI every day. Still googled What is RAG in AI? like it’s a new health fad. So I made this. 16 AI terms explained like you’re 10. Hallucinations, agents, vector databases etc. Watch this if you’re tired of pretending to know what “agentic” means. #kloudle #agentic #ai

10/. The worst feeling isn’t being hacked. It’s knowing you saw it coming.and still ran out of time.

9/. That’s what Kloudle is now. For every founder racing to ship. Every team juggling speed and security. Every startup that hears the ticking clock.

8/. That experience shaped how I built Kloudle. What I needed back then. A way to scan my cloud infra in minutes. No setup. No delay. Just clear visibility.