TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

The True Nature of LLMs

10 pointsby eschnou8 months ago

4 comments

moffkalast8 months ago
The old classic [0]. Sufficiently advanced linear algebra is indistinguishable from magic.<p>[0] <a href="https:&#x2F;&#x2F;www.reddit.com&#x2F;r&#x2F;LocalLLaMA&#x2F;comments&#x2F;1bgh9h4&#x2F;the_truth_about_llms" rel="nofollow">https:&#x2F;&#x2F;www.reddit.com&#x2F;r&#x2F;LocalLLaMA&#x2F;comments&#x2F;1bgh9h4&#x2F;the_tru...</a>
yawnxyz8 months ago
I think knowing what part of the knowledge base to delete — to get to adequately small reasoning model — is the hard part.<p>Doesn&#x27;t &quot;reasoning&quot; rise from the knowledge? How much of a brain can you cut away before you affect the reasoning? When do you know what you&#x27;ve cut away, and what aspects did you miss &#x2F; forget about?<p>We can probably train &#x2F; fine-tune, w&#x2F; synthetic data, and we&#x27;ll get reasonably close, but the &quot;reasoning&quot; will always hit rough patches, bc our training didn&#x27;t include <i>that</i> kind of reasoning... and if we had to give it examples of every single kind of reasoning, then it can&#x27;t move past all the already-established kinds of reasoning, so it&#x27;s still pattern matching
eschnou8 months ago
Hi HN, I&#x27;d love to get your thoughts on this one! Anyone using LLM, hidden inside an app, just as a reasoning &#x27;brick&#x27; to progress some workflows, decide on best math, etc.
avereveard8 months ago
Eh, this is just reinventing decision trees from first principles. There&#x27;s a reson why we cant have an universal decision tree, and it&#x27;s that the universal concepts would need to be described somehow to the model to take action, and this somehow is language, and our current sota for getting a model to understand language is to feed it gazillion combinations of sentences and their valid continuations.<p>But. There&#x27;s indeed some challenging aspect in making these model plan for solving unexpected, novel problems not in the training set. Possibly a model that produces axioms and relations and a constrain solver that evaluates if the solution is coherent, non conflicting and complete.