cd /news/large-language-models/llms-are-bad-at-novelty-but-that-is-… · home topics large-language-models article
[ARTICLE · art-54116] src=news.ycombinator.com ↗ pub= topic=large-language-models verified=true sentiment=↑ positive

LLMs are bad at novelty, but that is our chance to Singularity

A Hacker News user argues that large language models (LLMs) are inherently bad at generating novelty, which presents an opportunity for humans to focus on creative and innovative work. The author envisions a future where most people shift from repetitive jobs to creating new medicines, art, and scientific knowledge, potentially driving exponential economic growth and achieving a technological singularity through human ingenuity rather than AI.

read2 min views1 publishedJul 10, 2026

There is possibility that AI can automate all there can be automated. But still one area where it is notoriously bad is creation of novel things. What Schopenhauer called genius or ability to have new insights/discoveries that were not implicitly present in books.

"In most books, putting out of account those that are thoroughly bad, the author, when their content is not altogether empirical, has certainly thought but not perceived; he has written from reflection, not from intuition, and it is this that makes them commonplace and tedious. For what the author has thought could always have been thought by the reader also, if he had taken the same trouble; indeed it consists simply of intelligent thought, full exposition of what is implicite contained in the theme. But no actually new knowledge comes in this way into the world; this is only created in the moment of perception, of direct comprehension of a new." - Arthur Schopenhauer

I can envision a future where people are freed from doing repeatable jobs and 99% are shifted into enterprises of creating something new.

  • New life saving medicine - New life extension medicine - New art work - New novel - New product on the market - New engineering breakthrough - New scientific knowledge

Event though today only miniscule percentage of people are working in those kind of jobs it will very fast become dominant industry. Because that is the only loophole in LLMs that can not be fixed but is highly valuable and people with money will be ready to pay obscene amounts of money to extend their life. Who knows maybe there will be whole new big companies with sole purpose of trying to extend the life of Elon Musk or Jeff Bezos.

I work for Musk foundation and my job is to extend the life of Elon Musk - anonymous employee

Jokes aside, imagine the gigantic potential we can unlock if we transfer billions of people on a quest to unlock novelty. For millions of years people were working on repeatable jobs and our brain is not able to comprehend that there can be any other job out there that does not contain "skill", "productivity", "seniority" which are just another words for someone who mastered the art of repeatability which is not something humans should brag about.

I am saying that massive creation of novelty can cause double digit rates of GDP growth or even unlimited rates which means singularity. But it will not come from AI, but from people on a mission to discover something new.

Comments URL: [https://news.ycombinator.com/item?id=48858598](https://news.ycombinator.com/item?id=48858598)

Points: 1

── more in #large-language-models 4 stories · sorted by recency
federicopereiro.com · · #large-language-models
Fast legacy
── more on @elon musk 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/llms-are-bad-at-nove…] indexed:0 read:2min 2026-07-10 ·