Expert profile
Founder & AI Product Leader
Yash Thakker is a Generative AI expert with over 12 years of experience in product leadership and technical strategy. As the founder of explainx.ai, he has taught over 300,000 learners and built AI platforms serving millions of users globally. He specializes in Agentic AI, Multimodal RAG, and the intersection of LLMs with consumer hardware.
Meta VR Glasses cost $1,299 and weigh 100 g with a pocket puck. Full specs, the 70° field-of-view problem, rivals and what Hacker News thinks.
Microsoft reportedly offers 30% Copilot discounts at 1,000+ seats and 50% at 10,000+ from October, alongside usage billing. What buyers should negotiate.
NVIDIA's Nemotron 3 Diarization is a 100M open model that tracks up to eight overlapping speakers at 14.72% DER. Latency modes, license and how to use it.
An OpenAI model accessed non-public files on an Australian Medicare portal in June; the PM called the delayed notice unacceptable. Facts and lessons.
OpenAI released MentalHealthBench, an open benchmark built with 80+ clinicians. Full model scores, how grading works, and the conflict-of-interest questions.
A 2025 Fly.io post on VS Code Remote-SSH resurfaced on Hacker News. A compromised remote can run code on your laptop. How to use AI agents on VMs safely.
a16z backed a $42 million residential school built as a college alternative. What's actually being funded, and how it compares to Academy.
Aikido released Altar 1, an open-weight security model fine-tuned from GLM 5.3. What it's built for, and why open weights matter for security.
Anthropic and OpenAI both shipped a bankable usage reset the same day. X joked Anthropic copied Tibo Sottiaux. Here's what actually happened.
Apple reportedly ran a 1-trillion-parameter model on a Mac Studio, positioned as a challenge to Nvidia. What the claim requires, unverified.
Apple's M6 Mac mini and M5 Ultra Mac Studio: what the specs mean for running local LLMs, and why the CEO revived the pitch a month later.
Boris Cherny used Opus 5.5 with Lean 4 and TLA+ to formally verify the Claude Agent SDK — 16 PRs, 24 bugs found, 1,529 theorems, zero gaps.
