Erik Hoel on AI
6 quotes from Erik Hoel on the wall, 2023 to 2026, newest first. Each one is verbatim and links to where it was said. Sources include The Intrinsic Perspective, interconnects.ai.
- 2026
We are well into the era of human cope
Erik Hoel in his newsletter The Intrinsic Perspective, May 2026, in an essay on OpenAI's disproof of an Erdős conjecture, commenting skeptically on humans celebrating marginal mathematical improvements over AI-generated proofs.
Approximation is not automation
Erik Hoel, in his essay "Bits In, Bits Out" in The Intrinsic Perspective, March 2026, argues that LLMs can mimic text well but not truly automate intellectual work like writing, six years after GPT-3.
Neuroscientist and writer Erik Hoel, in a January 2026 essay on The Intrinsic Perspective describing his new arXiv paper, states his proof's conclusion that large language models cannot be conscious.
The Intrinsic PerspectiveProving (literally) that ChatGPT isn't conscious
- 2024
Humans will be weirdly sticky in a world after Artificial General Intelligence
Erik Hoel, in his essay "Sticky humans in a post-AGI world" in The Intrinsic Perspective, October 2024, compares the survival of independent bookstores after Amazon to how humans may remain preferred over AI even after AGI.
- 2023
The industries themselves are not very lucrative
Erik Hoel, in a December 2023 essay on The Intrinsic Perspective, arguing that the markets AI is best at disrupting, such as freelance writing and illustration, are not large or profitable enough to justify the industry's valuations.
The Intrinsic PerspectiveExcuse me, but the industries AI is disrupting are not lucrative
Corporations don't get to decide if we have competitors as a species
Quoted by Nathan Lambert in a 2023 essay on Interconnects discussing AI alignment and safety discourse.
interconnects.ai"AI alignment" and uncalibrated discourse on AI
Mentioned in other quotes 4
Quotes by others that refer to Erik Hoel
- 2026
LLMs are not people: They're counterfeit people
Philosopher Dan Dennett, quoted by Erik Hoel in a June 2026 Intrinsic Perspective essay rebutting Matthew Yglesias, argued that large language models should not be mistaken for conscious or sentient beings.
Mathematicians are having a Lee Sedol moment
Pedro Domingos reacting on Twitter/X in May 2026 to OpenAI's claim that an internal model had disproved a geometry conjecture of Paul Erdős from 80 years ago. Quoted by Erik Hoel in The Intrinsic Perspective.
- 2024
They are good tests of memorization, not of intelligence
Podcaster Dwarkesh Patel, quoted in Erik Hoel's essay on AI progress plateauing, argued that benchmarks like MMLU and BigBench mainly test memorization rather than reasoning.
The Intrinsic PerspectiveAI progress has plateaued below GPT-5 level
- 2023
A particular kind of religious discourse in a secular age
Economist Tyler Cowen, quoted in Erik Hoel's 2023 essay on AI extinction-risk warnings, described the AI safety debate as resembling religious discourse in a secular age.
The Intrinsic PerspectiveThe White House agrees you have a small brain