America is getting an AI gold rush instead of a factory boom

AI vs Manufacturing Investment

• Many see the “AI gold rush” as soaking up capital and power that could have gone into factories and durable productive assets; others note data doesn’t yet show data-center capex crowding out overall equipment investment.
• Power consumption of AI datacenters and its impact on electricity prices is a recurring concern.
• Some argue US manufacturing value added is at record highs but growth is modest and jobs are down; others say unit output is flat and GDP hides industrial decline.

AI’s Role in Factories and Robotics

• Optimists: AI (especially vision and transformer-based control) could drastically expand what robots can do—handling messy, context‑rich tasks, lowering the minimum scale at which automation pays off, and enabling more flexible, “general-purpose” robot workcells.
• Skeptics: LLMs excel at flexible, fuzzy tasks—the opposite of mass manufacturing’s need for tiny, exact, repeatable instruction sets. Current industrial automation already uses “AI” (ML, vision) where it helps.
• Some see LLMs’ main manufacturing impact as assisting engineers (design, programming, workflow), not running robots directly.
• Several commenters dislike that “AI” is used to lump together control/robotics and LLM chatbots, which drives confusion and hype.

Jobs, Wages, and Desirability of Factory Work

• Fears: AI plus automation could further hollow out the middle class and crush new‑grad and creative jobs, without delivering widely shared gains.
• Others say current hiring weakness is macro (rates, tariffs, politics), not AI, though belief in AI makes managers more willing to cut headcount.
• Long subthread on why US factories struggle to hire: monotonous, physically demanding, often unsafe work versus similar or lower-paid service jobs with easier conditions.
• Disagreement over whether “higher pay and better benefits” claims from employers are substantial or illusory; unions and working conditions (breaks, music, respect) are central themes.

US vs China Industrial Capabilities

• Multiple commenters argue China has quietly built deep process knowledge, heavy automation, and broad tech leadership, while the US financialized and offshored its industrial base.
• Others note many Chinese factories still rely on labor‑intensive assembly, and demographic decline will force more automation globally.
• Debate over whether US can realistically rebuild manufacturing capacity at scale after losing tooling ecosystems and skills, versus targeted, highly automated reshoring (EVs, chips, defense).

Trade, Tariffs, and National Security

• Competing views: tariffs as necessary to preserve strategic industries vs tariffs as a regressive tax that makes everyone poorer.
• Strong argument that some domestic manufacturing is essential for leverage and security (we must be able to “build it ourselves” if trade breaks down), but not everything can or should be onshore.
• Japan’s protectionist playbook and China’s import substitution are cited as examples where import barriers worked only with long‑term, coordinated industrial policy.

AI Bubble, ROI, and AGI Bets

• Widespread unease that AI resembles past bubbles: enormous capex into rapidly obsoleting hardware, unclear sustainable business models, and “too big to fail” political backing.
• Practitioners report real but modest productivity gains (coding help, summarization) alongside new costs: reviewing AI-generated “slop,” hallucinations, and brittle integrations.
• Intense argument over the “AGI race”: some claim whoever reaches AGI first will dominate geopolitically, justifying massive overinvestment; many others doubt LLMs can reach or safely control AGI and question the wisdom of betting an economy on that assumption.

Structural Barriers to a Factory Boom

• Experienced founders describe capital markets, startup culture, and exit environment as heavily biased toward software and asset‑light “middleman” models; financing bus‑sized machines in rich countries is hard, exits are scarce, and supply‑chain fragility is rising.
• Even where demand exists, regulatory delay, permitting, and fragmented policy make standing up new plants slow and risky compared with software or Chinese manufacturing.