The post argues from biotech and medicine that intelligence is often not what keeps progress slow. Better models do not remove long waits for National Institutes of Health datasets, Food and Drug Administration endpoint validation, clinical trial enrollment, or the political and organizational friction that blocks execution. The point is not that intelligence is unimportant. It is that many fields already have plausible technical paths and still fail because incentives, regulation, data access, embodiment, and collective action are where things stall.
That framing landed with a lot of people. The strongest comments widened it from biotech to a general claim about real-world systems. Human behavior, communication, and institutions were treated as the actual choke points. Several commenters connected this to climate change, housing, healthcare, and public policy, where better plans already exist but getting people to coordinate around them is the hard part. A recurring version was that AI can generate options, but it cannot grant permission, build trust, win local politics, or make incumbents surrender power.
Where the conversation got sharper was over whether this really demotes intelligence at all. One camp said the article is directionally right but imprecise. Governance, persuasion, negotiation, and process design are themselves forms of intelligence, just applied at the system level rather than inside a lab or model. The more persuasive rebuttal was that this misses the practical point. We are not short on ideas in many domains. We are short on adoption, legitimacy, and execution under conflict. That is why commenters kept returning to clinical trials, climate policy, zoning fights, and antibiotic overuse instead of hypothetical breakthroughs.
There was also a smaller but lively side argument about "hyperpersuasion" and the Rationalist or
LessWrong belief that sufficiently advanced AI could talk humans into anything. Most people treated that as magical thinking. They were more worried about scaled distribution through feeds, botnets, and engagement systems than about one chatbot hypnotizing a skeptical operator. The broad mood was skeptical of intelligence-as-master-key stories and impatient with any AI narrative that quietly assumes politics, institutions, and human incentives will obediently fall into place once the model gets smart enough.
writingruxandrabio.com