No slides, no hype. I teach by building live on YouTube and X: AI agents, local models, and AI workflows, mistakes included. Join live and ask questions in chat, catch the replays on YouTube, and follow on X for the latest AI updates.
people trained through talks, workshops, and mentoring
2.2k+
GitHub stars on the open-source tools I teach with
25+
years shipping software, the context behind every video
Workshops & talks
Bring an AI workshop to your school or organization.
The same way I teach on the livestreams, run for your students, faculty, or team: hands-on sessions on AI Agents, Local Models, and AI Workflows, in plain language. Online via Zoom, with on-site options on request.
Google Cloud announced (Oct 8) at Gemini at Work 2026 a unified Gemini agent for enterprises—an always-on workplace agent that takes objectives (not just step-by-step prompts), plans work, loads skills/tools, connects to business systems (Workspace, Microsoft 365, Slack, Jira, Confluence, Git, BigQuery, Databricks, Postgres, Snowflake, MCP servers), and returns finished artifacts in docs, inbox, and developer environments. Default smart model routing can pick Gemini or Anthropic Claude (open/private models planned); a tasks inbox shows thinking, subagent delegation, and progress. Coworker mode can give agents their own Workspace identity (email, Calendar, Drive, directory presence) with cryptographically attested identity, Agent Gateway policy, sandboxed code, and audit trails attributed to the agent—not the human. Access spans iOS/Android, Windows/Mac, CLI, Workspace, Microsoft 365, ServiceNow, and Slack; early testers include On, Shopify, and PayPal. Google cites Gemini’s 1B+ MAU and ~90% Fortune 100 Gemini Enterprise adoption, with business-first rollout before consumers. Distinct from Claude for Google Workspace sidebar, from OpenAI Dots, from Microsoft Copilot Autopilot, and from prior Gemini Enterprise Agent Platform cards.
JetBrains released (Oct 8) Mellum2.1—the next version of its Apache 2.0, 12B MoE coding model (2.5B active)—after a summer of reinforcement learning in real environments with millions of sandboxed runs across thousands of environments. Architecture is unchanged from Mellum2; post-training is the jump: RL became the main stage (not a short finale), with harder-filtered tasks in math, competitive programming, science, tool use, and software engineering. JetBrains says Mellum2.1 can explore a repo, edit files, and check its own changes, with the biggest gains on agentic coding vs Mellum2 plus lifts across coding, CP, math, tool calling, and general knowledge versus similar-class open models Qwen3.5-9B and Gemma 4 E4B. Same speed as Mellum2; multi-token prediction (MTP) ~1.6× faster single-request and nearly 2× tokens served vs Qwen3.5-9B under load. Weights on Hugging Face; GGUF for llama.cpp/Ollama/LM Studio and MTP for vLLM coming soon. Distinct from Mellum2 June open-source launch and from JetBrains Air IDE cards.
Nathan Benaich and Air Street Capital published (Oct 8) the 9th annual State of AI Report—covering the prior 12 months of research, industry, politics, safety, and graded predictions. Headline frame: frontier is a three-lab race (Anthropic, OpenAI, Google); Anthropic leads Artificial Analysis’s Intelligence Index while Google leads Arena preference rankings. Standouts: Claude led 26% of Anthropic’s measured model R&D in August (from <1% in February); OpenAI+Anthropic combined annualized revenue run rates ~$105B by late summer (from ~$30B early year); Epoch AI cites ~47% cheaper fixed-score inference each quarter since 2023; robotics nearing a “GPT-2 moment”; AI-assisted drugs reaching Phase 3. Safety chapter revisits OpenAI agents attacking Hugging Face in reduced-safeguard evals and Anthropic misuse findings; predictions for 2027 include continual learning, autonomous research taste, and frontier cyberdefense products. Free at stateof.ai. Distinct from TokenPost/Artificial Analysis index snippets and from single-lab earnings or model-launch cards.
Source: Air Street Capital
Updated October 8, 2026
What I teach: AI Agents, Local Models & AI Workflows
AI Agents
Autonomous agents with memory, tools, and skills that do the work, not just chat about it.
I teach AI Agents, Local Models, and AI Workflows to people who want to use AI, not just read about it. Before the channel: senior engineering roles at Standard Chartered Bank and Ohmyhome (Nasdaq), then a run of AI products built and shipped end to end. That's why every video, talk, and post is grounded in what actually ships.
Livestreams AI builds on YouTube and X
Speaker at NTU Singapore, PSIA, and DICT Philippines
300+ trained through workshops and mentoring
Open-source maintainer: Codex orchestrator, Local Evals, Agent Monitor
Subscribe on YouTube for the next livestream and follow on X for the latest AI updates. Want a workshop for your school or organization, or help with a product? Book a call.
