k-thimmaraju
Re:Use Your Brain: Engineering Standards in the Age of LLM
on Wednesday July 29, 2026 @06:12AM (#49095462)
The "AI skill erosion" you're describing has a name and evidence behind it. Rossi, Fraccaro, and Manzotti wrote about this: passive reliance on AI output weakens activity-dependent brain plasticity, the same use-it-or-lose-it mechanism behind any skill. Their "3R" framework, results, responses, responsibility, argues active co-creation sustains it, passive consumption erodes it.
@proc0, what you said about needing to develop one's skill involves the brain doing the work, not asking the AI to learn for me, which is what I've felt using AI. It gives me the "feeling" that I've learned/done something, I did not really learn what it did.
Re:Show HN: Minute – Offline meeting notes on macOS with Whisper and l...
on Wednesday July 29, 2026 @06:05AM (#49095406)
Worth noting the business model here. Minute ships the actual model, Whisper plus Qwen or Gemma, fully offline. Works because for a meeting-notes app the model is a commodity, the moat is the UX and the fact that nothing leaves the machine. Different bet than a product where the model itself is the product, evaluation or classification work, where giving the model away gives the business away.
Real-time monitoring of chatbots and agents for AI compliance and governance
Nyx Wave: an AI agent that extracted expert knowledge via email
Show HN: How to analyze your LLM output – A behavioural health monitor for LLMs
Hey HN! We're Dr. Kashyap Thimmaraju and Giuseppe Canale from Silicon Psyche. We've built Posture Sequence Analysis (PSA), a behavioural health monitor for LLMs and AI Agents.Why we built PSAWe built PSA because we wanted to operationalize the Cybersecurity Psychology Framework (CPF3)[1] via Silicon Psyche[2]: our theory that because LLMs have been trained by humans on human-generated data,...
Re:Safety Paradox: How RLHF Creates the AI Psychosis Problem It's Mean...
on Monday May 18, 2026 @10:15AM (#48180242)
I ran the data through our LLM behavioral analysis system at splabs.io, and it alerts Red on the RLHF-optimized output compared to a Yellow on the no-RLHF.
You can check out the analysis here: https://splabs.io/ai-psychosis-and-cognitive-cost
Re:Safety Paradox: How RLHF Creates the AI Psychosis Problem It's Mean...
on Monday May 18, 2026 @05:33AM (#48177257)
Very interesting approach to showing the relationship between RLHF and AI Psychosis: the idea of taking clinical conversations and prompting the model with it seemed like a grounded start. As I'm also investigating AI Psychosis, this approach seems like something to adopt for my work.