Hacker News - Newest: "LLM"

Train Your Own LLM from Scratch

kristianpaul — Tue, 05 May 2026 04:09:17 +0000

Article URL: https://github.com/angelos-p/llm-from-scratch

Comments URL: https://news.ycombinator.com/item?id=48017948

Points: 214

# Comments: 22

Show HN: A tiny C program where an LLM rewires its DAG while running

mrkn1 — Tue, 05 May 2026 01:43:05 +0000

Article URL: https://github.com/kouhxp/liteflow

Comments URL: https://news.ycombinator.com/item?id=48017118

Points: 9

# Comments: 0

Show HN: Smile-Serve – Inference Server for ML, ONNX, and LLM

haifeng — Mon, 04 May 2026 22:02:55 +0000

SMILE Serve is a production-ready inference server built on [Quarkus](https://quarkus.io/) that brings together three complementary inference capabilities on the JVM:

  - **Classic ML**: `/api/v1/models` for serialized SMILE models (`.sml`)
  - **ONNX Runtime**: `/api/v1/onnx` for any model in the ONNX open format (`.onnx`)
  - **LLM Chat**: `/api/v1/chat` for Llama 3 chat completions

A React-based web UI is bundled and served from the same process.

Comments URL: https://news.ycombinator.com/item?id=48015597

Points: 5

# Comments: 0

LLM anomaly detectors are not a cause for concern despite Mythos

badcryptobitch — Mon, 04 May 2026 19:43:33 +0000

Article URL: https://www.magonia.io/research/why-a-decade-of-writing-detection-logic-makes-the-mythos-exploit-numbers-less-scary/

Comments URL: https://news.ycombinator.com/item?id=48013928

Points: 2

# Comments: 0

Exploring LLM biases to manipulate AI search overview

Brajeshwar — Mon, 04 May 2026 17:49:05 +0000

Article URL: https://arxiv.org/abs/2605.00012

Comments URL: https://news.ycombinator.com/item?id=48012208

Points: 1

# Comments: 0

A thermodynamic trust layer cutting LLM hallucinations by 52%

Dan23RR — Mon, 04 May 2026 17:46:40 +0000

Article URL: https://github.com/Dan23RR/snc-core

Comments URL: https://news.ycombinator.com/item?id=48012163

Points: 4

# Comments: 0

LLM-first document AI is missing a 50-year-old CS technique

bhavya2k03 — Mon, 04 May 2026 15:34:23 +0000

Article URL: https://bhavyagupta.dev/posts/llm-document-extractors-fixed-point

Comments URL: https://news.ycombinator.com/item?id=48010017

Points: 3

# Comments: 0

How HN: BibCrit – LLM analysis grounded in real manuscript corpus data

jossifresben — Mon, 04 May 2026 14:55:54 +0000

Article URL: https://bibcrit.app/

Comments URL: https://news.ycombinator.com/item?id=48009524

Points: 2

# Comments: 1

Show HN: The Cat Is Under Mayonnaise – Modifying LLM Behavior Without Retraining

andycufari — Mon, 04 May 2026 13:38:14 +0000

Article URL: https://github.com/andycufari/the-cat-is-under-mayonnaise-experiment

Comments URL: https://news.ycombinator.com/item?id=48008612

Points: 2

# Comments: 1

Show HN: Aurra – Bi-temporal memory for AI agents (with LLM auto-supersede)

akshayt2012 — Mon, 04 May 2026 12:53:49 +0000

Article URL: https://www.aurra.us/blog/level-2-auto-supersede-beta

Comments URL: https://news.ycombinator.com/item?id=48008086

Points: 3

# Comments: 0

Fluctuating Accuracy in LLM Responses

chris_explicare — Mon, 04 May 2026 12:46:42 +0000

Dear HN community, I’m brand new here and already feel right at home after just 5 minutes. I have a question for you about my theory:

I’m sure you’ve all experienced the wildly fluctuating quality of LLM responses. My theory: During peak times, the operators gradually reduce the depth of processing to take some of the load off the servers. I’ve noticed this a lot with Claude over the past few months. What do you think?

Comments URL: https://news.ycombinator.com/item?id=48008015

Points: 3

# Comments: 0