Show HN: I was curious about spherical helix, ended up making this visualization (visualrambling.space)

Most LLM benchmarks measure accuracy and coherence, but not whether the intended meaning survives. I’ve been calling this gap fidelity, the preservation of purpose and nuance. Has anyone else seen drift like effects in recursive generations or eval setups?

Comments (8)

Mallowram · 4h ago

What is intended meaning when words are arbitrary and meaning is always relative to arbitrariness?

realitydrift · 4h ago

That’s a fair point. Words themselves are arbitrary symbols, but meaning isn’t only in the symbols. It’s in the intent behind them and the use they’re put to.

For example, if I say “the meeting is at 3pm” and a model rewrites it as “planning sessions are important,” the words are fine, the grammar is fine, but the purpose (to coordinate time) has been lost. That’s the gap I’m calling fidelity: whether the output still serves the same function, even if the surface form changes.

Mallowram · 4h ago

There is no intent in words in and of themselves. Intent always comes from something specific tied to neural syntax, which is lost in words. That's an illusion. There is intensionality, which is different, which is what you're actually talking about. Intensionality is vague, it's not meaningful without context. The problem automating words is AI can't solve the conduit metaphor, which is the idea the words alone code meaning. They can't. This is the achilles heel of AI.

realitydrift · 3h ago

I agree. Words don’t carry intent by themselves. Intention is always embedded in use. That’s why I frame fidelity as about whether a system’s continuation still serves the same human purpose. The “conduit metaphor” you mention is exactly the trap: treating words as if they inherently encode meaning. Models fall into this because they optimize surface probabilities rather than checking whether the function of the exchange was preserved.

Mallowram · 3h ago

In 1973 Basil Bernstein studied how UK scores in math defied class boundaries while scores in reading comprehension/essays stayed tied to them, and he developed a theory with Halliday that language embeds so much that we cannot decipher easily: dominance, status, control, land-centering, gender/mate-selection, and much more, that they assumed language was far more a social system of primate-simian signaling rather than what Shannon and classical linguists took as "communication". My guess is, LLMs are really unresolvable revelations of these hidden nuances. In a way, LLMs demand a specific language, which doesn't exist, in order to function.

realitydrift · 3h ago

That’s a really interesting reference. Bernstein and Halliday were basically pointing out that language is never just propositional, it’s always smuggling social structure with it. That’s exactly why drift matters: when an LLM compresses or rewrites, it isn’t just shifting words, it’s rebalancing those embedded cues of power, context, and purpose. Humans keep the “extra baggage” because it carries meaning beyond the literal. Models optimize it away. That gap between statistical surface and lived function is what I’ve been calling semantic drift.

docsorb · 4h ago

You're touching upon a very nuanced point of identifying the true "intent" behind these words. Do you think the way these models are trained should be different to correctly map the potential intent vs the true meaning?

like your example, "the meeting is at 3pm", _we got enough time_ intends something else with "the meeting is at 3pm" _where the hell are you?_ intends something else. It is not so obvious to get that intent without a lot of context (like, time, environment, emotion etc.)

realitydrift · 4h ago

Exactly. That’s the hard part. Meaning is often carried less by the literal words and more by context (time, environment, emotion, shared knowledge). My point with fidelity is that current benchmarks don’t check whether outputs preserve that function in context. An AI can echo surface words but miss the intended role: coordination, reassurance, accountability. And that’s where drift shows up.

Google will allow only apps from verified developers to be installed on Android (9to5google.com)

AWS CEO says using AI to replace junior staff is 'Dumbest thing I've ever heard' (theregister.com)

FFmpeg 8.0 (ffmpeg.org)

Show HN: I was curious about spherical helix, ended up making this visualization (visualrambling.space)

Why are anime catgirls blocking my access to the Linux kernel? (lock.cmpxchg8b.com)

Mark Zuckerberg freezes AI hiring amid bubble fears (telegraph.co.uk)

What are OKLCH colors? (jakub.kr)

A German ISP changed their DNS to block my website (lina.sh)

DeepSeek-v3.1 (api-docs.deepseek.com)

AI tooling must be disclosed for contributions (github.com)

Show HN: Base, an SQLite database editor for macOS (menial.co.uk)

Go is still not good (blog.habets.se)

Comet AI browser can get prompt injected from any site, drain your bank account (twitter.com)

A visual introduction to big O notation (samwho.dev)

U.S. government takes 10% stake in Intel (cnbc.com)

Waymo granted permit to begin testing in New York City (cnbc.com)

Ban me at the IP level if you don't like me (boston.conman.org)

Zedless: Zed fork focused on privacy and being local-first (github.com)

Ask HN: Why does the US Visa application website do a port-scan of my network?

Weaponizing image scaling against production AI systems (blog.trailofbits.com)

Io_uring, kTLS and Rust for zero syscall HTTPS server (blog.habets.se)

How to build a coding agent (ghuntley.com)

What makes Claude Code so damn good (minusx.ai)

Sequoia backs Zed (zed.dev)

Line scan camera image processing for train photography (daniel.lawrence.lu)

Pixel 10 Phones (blog.google)

Building the mouse Logitech won't make (samwilkinson.io)

Gemini 2.5 Flash Image (developers.googleblog.com)

Manim: Animation engine for explanatory math videos (github.com)

Home Depot sued for 'secretly' using facial recognition at self-checkouts (petapixel.com)

Gemma 3 270M re-implemented in pure PyTorch for local tinkering (github.com)

We put a coding agent in a while loop (github.com)

Burner Phone 101 (rebeccawilliams.info)

Everything I know about good API design (seangoedecke.com)

95% of Companies See 'Zero Return' on $30B Generative AI Spend (thedailyadda.com)

4chan will refuse to pay daily online safety fines, lawyer tells BBC (bbc.co.uk)

Ghrc.io appears to be malicious (bmitch.net)

Code review can be better (tigerbeetle.com)

AWS in 2025: Stuff you think you know that's now wrong (lastweekinaws.com)

macOS 26 Tahoe's Dead Canary Utility App Icons (daringfireball.net)

Google's Liquid Cooling (chipsandcheese.com)

FCC bars providers for non-compliance with robocall protections (docs.fcc.gov)

What is going on right now? (catskull.net)

Code formatting comes to uv experimentally (pydevtools.com)

Is 4chan the perfect Pirate Bay poster child to justify wider UK site-blocking? (torrentfreak.com)

Making games in Go: 3 months without LLMs vs. 3 days with LLMs (marianogappa.github.io)

Leaving Gmail for Mailbox.org (giuliomagnifico.blog)

Ask HN: Why hasn't x86 caught up with Apple M series?

Temporary suspension of acceptance of mail to the United States (post.japanpost.jp)

An illustrated guide to OAuth (ducktyped.org)

Semantic drift: when AI gets the facts right but loses the meaning

Comments (8)