Gary Marcus: Generative AI Is Terrible Engineering
GaryMarcus · x · 2026-07-15
Gary Marcus amplified an opinion piece arguing that "AI doesn't have to be built this way," calling the premise that "we are building AI wrong" mandatory reading.
The core arguments from the Atlantic article he cited include:
- Researchers note that generative AI scales incredibly inefficiently, making it hard to find comparable real-world software scaling behaviors.
- While traditional industries and software rely on economies of scale, generative AI might be "one of the worst-engineered technologies."
- The article doesn't simply oppose AI; it questions the engineering and economic sustainability of the current technological path.
Marcus uses the piece to reinforce his own stance: the current approach to generative AI suffers from structural flaws, not just short-term product experience issues.
Related event: Debate Flares Over The Atlantic’s Critique of Generative AI(6 posts)→
More from AGI Musings
- Gary Marcus says LLM math skills are like knowing only a car’s engine size — GaryMarcus · 2026-07-22
- AI may make digital work infinitely leveraged while offline life gets more human — illscience · 2026-07-22
- Better AI math could save researchers time by killing false conjectures earlier — prateekj · 2026-07-22
- AI’s economic forecasts are split by nearly a quadrillion dollars by 2035 — bittingthembits · 2026-07-22
- Open source is becoming tech’s soft power, says Kevin Xu — kevinsxu · 2026-07-22
- OpenAI should keep giving more people access to more powerful AI — jxnlco · 2026-07-22