Andrew Ng

Founder, DeepLearning.AI · Managing GP, AI Fund · Founder & Exec Chairman, Landing AI · Expert Watch · Fast Watch, weekly scan · our rolling analysis

Why we watch him. Andrew Ng is among the clearest translators of frontier AI research into deployable enterprise practice, and having coined and popularised "agentic" workflows, his read tends to set the next quarter of board-level AI conversation. He sits squarely on Dominic's core junction: coding gets 10 to 100x faster while value and the bottleneck shift to judgement, product taste and everything that is not the code. His pragmatic, adoption-first stance ("AI is the new electricity", 8M+ learners) makes him a useful counterweight to both the doom camp and the pure vendor voices, and a live tripwire on where enterprise AI practice is actually heading.

Current headline view: when building software gets 10 to 100x faster with coding agents, everything that is not building (deciding what to build, product judgement, feedback loops) becomes the new constraint. Coding gets much cheaper and faster while the value moves to judgement, product taste, and everything that is not the code.

Latest 1-page summary (as at 11 Jul 2026) 🟠 new thinking

The eight core theses we track him against:

  1. "AI is the new electricity". A general-purpose technology transforming every industry (his signature framing).
  2. Agentic workflows are the real unlock. Iterative agent loops beat single-shot prompting, and getting a prototype into production is where the value lands (Interrupt 2026, No Priors 2025).
  3. Speed is the startup advantage. Build and validate faster; execution velocity, not idea secrecy, wins (YC talk, Jun 2025).
  4. Democratize building. More people, including non-developers, should learn to direct AI to build (The Batch).
  5. AGI is decades away, not imminent. The nearer-term bubble risk sits in the training layer (Feb 2026 interview).
  6. Adoption over doom. Real applications in education, healthcare and enterprise over speculative existential risk.
  7. Regulation should not crush open source or small firms. Over-restrictive model rules burden smaller players most.
  8. The 3 loops for 0-to-1 with coding agents. Agentic coding (minutes), developer feedback (hours), external feedback (days) (The Batch, Jun 2026).

Method note. This v1 baseline is distilled from his newsletter "The Batch", recent talks and interviews (Interrupt 26, No Priors Ep. 128, the YC "Building Faster with AI" talk, and a Feb 2026 interview), not yet a full transcript or X history read. The weekly scan refines it from here and flags any deviation.

Analysis backlog (newest first)

Deep source docs are held locally in the workspace at expert-watch/, not on this page.

Notable reactions (operators & researchers)

As at 11 Jul 2026.

None logged yet. This is a v1 baseline; the weekly scan will capture notable operator and researcher reactions to his output from here.

Recent media

As at 11 Jul 2026.

Quick links

X (@AndrewYNg) LinkedIn The Batch DeepLearning.AI Latest YouTube (Interrupt 26)