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.
Calacanis: open≈frontier. Musk: world of difference. Task difficulty, Kimi K3, pacing/antitrust side-threads — an explainx.ai builder read.
Late July–Aug 2026: Opus 5 one-prompt Three.js games went mainstream — FPS, Minecraft clones, karts — while human WebGPU demos like RVO2 zombies show the other track.
Cloudflare Agents Week Day 1: @cloudflare/computer gives each agent a shared SQLite filesystem with isolate + container backends. Why it matters for scale.
Comp AI’s MIT CRM puts a research agent on Eve ahead of forms. Stack, quick start, sandbox rules, and real dispatch.ts durability gaps.
OpenCode reports DeepSeek Flash at 8T tokens on Aug 1 (5T free + 3T Go). See the cost math, MIT 284B MoE context, and builder takeaways.
Musk’s unlabeled AI tsunami chart went viral. Axes missing, “it’s so over” marked Nov 2024, then a 2026 spike — rhetoric vs measurable progress.
explainx.ai is building a practitioner-facing AI Safety Score. Early signal: everyday dilemmas, model behavior under real phrasings — details soon.
Use pdf-inspector for native-text extraction, inspect reading order and tables, preserve page references, and test OCR fallback on mixed PDFs.
FROG in a Bowl = Format, Role, Objective, Goal, plus Context. The prompting checklist labs recommend, named so you remember it.
Genspark open-sourced GenOffice Alpha: Docs, Sheets, Slides, PDF on Electron for Mac/Windows. Free core, AI via credits, Apache-2.0 — builder read.
Inkling-Small: 12B active MoE open weights that match Inkling on coding and reasoning at lower cost. Benchmarks, Tinker, trade-offs vs factuality.
EvoLib: test-time learning via consolidation + dynamic weighting of skills and insights. MIT code, arXiv 2605.14477, vs memory archives and SkillOpt.
