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.
Subquadratic’s SubQ pairs a sub-quadratic sparse-attention stack (SSA) with a 12M-token positioning; official SSA post cites 52× prefill at 1M vs dense FA...
HY-World 2.0 from Tencent Hunyuan: multi-modal 3D worlds (3DGS/meshes) vs pixel-only video world models, WorldMirror 2.0 reconstruction, pipeline...
Immich: AGPL-3 self-hosted photo & video manager (NestJS/Svelte/Flutter), ML search & faces, mobile backup—Google Photos alternative. immich.app docs.
LangChain’s Deep Agents jumped Terminal-Bench 2.0 with the same GPT‑5.2‑Codex—harness-only. Plus harness definitions (Hashimoto), Stanford IRIS...
Cofounder 2 coordinates department agents with roadmap milestones, approvals, and MCP/skills. Official intro: cofounder.co/resources/introducing-cofounder-2.
DeepSeek V4-Pro MoE (1.6T/49B), 1M context: SWE Verified 80.6% (HF Table 6), CSA/HCA; official API pricing & promos—DeepSeek Models & Pricing + PDF report.
Pre-mortem agent skill (parcadei/continuous-claude-v3): verified risks for coding agents—two-pass workflow, npx skills install. Canonical explainx.ai listing.
Runway Characters on GWM-1: one image → 24fps HD, ~37ms/frame & ~1.75s server turn; vision, tools, RAG, meetings. runwayml.com/news/building-runway-characters.
Saperly gives AI agents real numbers with voice + SMS (hosted, webhook, audio modes) and npx @saperly/mcp. Pricing & zones: saperly.com + docs.saperly.com.
What project-level skills-lock.json records—GitHub sources, sourceType, computedHash—and why teams commit it for npx skills workflows, CI, and supply-chain...
Comprehensive guide to AI benchmarks in 2026: language models (MMLU, HellaSwag), reasoning (GPQA, Humanity's Last Exam), coding (SWE-bench, LiveCodeBench),...
A May 2026 X post claimed an email after mocking Claude. Here is what Anthropic actually documents: Opus can end abusive threads in-app, Usage Policy...
