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.
Flock Safety layoffs: about 270 jobs, 18% of staff, per Reuters. Buyout timeline, cities leaving, misuse cases, and the AI license-plate camera debate.
A US charity plans AI classroom monitoring in Gaza with 72-hour recordings and alerts to its founder. What AP confirmed, what is unclear, and the criticism.
Gigabyte W775-V10-L01 GB300 workstation: 252GB HBM3E, 496GB LPDDR5X, 1.6kW, 26,000+ tokens/s. Who it is for, versus DGX Spark, Mac Studio M5 Ultra and RTX PRO.
Gods Don’t Give Gifts AI movie: first MPA rating (R), December 4 release, Best Animated Feature campaign. Facts, caveats and what filmmakers can learn.
Google AMIE was tested on 100 real patients at Beth Israel Deaconess and published in The Lancet. The 90% figure, the 56% top-1 figure, and what they mean.
Grok Imagine Video 1.5 Lite costs $0.02 per second at 480p up to $0.14 at 1080p and ranks 17 on Artificial Analysis text-to-video. Pricing, speed and limits.
Japan cyberattacks in September 2026: Times Car (6.6M), Gyazo (23.6M), Keio ransomware, Monogatari (10.8M). Confirmed vs claimed, plus a protection checklist.
LangChain Restock shows how to build an AI agent that pays with Stripe Link and MPP, with Slack review and Link approval. Setup, flow, limits and risks.
LightOnOCR-3 adds grounding boxes, image descriptions and chart tables to 0.8B, 1B and 4B Apache 2.0 OCR models. Benchmarks, speed on H100, and the caveats.
Torvalds says he likes AI, calls vibe coding a way to find joy in programming, and warns about care in the kernel. What he said and what it means.
Long-WAM from NVIDIA scales world-action model context to 19.2 s: 78.7% on RoboCasa GR-1, 99.5% LIBERO-Long, 107.4 ms per action chunk. Paper and project page.
Lumentum sold out through 2029: which optical parts AI data centers need, the reported 30% shortfall, the Nvidia deal, and what it means for GPU buildouts.
