All topics

LLM stories

Large language models have moved from research curiosity to production infrastructure. This topic focuses on LLM APIs, fine-tuning, RAG pipelines, agent frameworks, and the economics of running models at scale.

HN discussions here tend to be unusually practical — benchmark comparisons, cost breakdowns, and real deployment war stories from teams shipping LLM features.

898 stories archived · Page 26 of 30

RSS feed for LLM

GM Backs Sodium Ion Batteries for U.S. Grid Storage

spectrum.ieee.org · 207 points · 105 comments

Becoming a Research Engineer at a Big LLM Lab

maxmynter.com · 54 points · 21 comments

Awsmux – Multi-account AWS CLI, up to 5.4x faster, 7.4x fewer tokens

github.com · 10 points · 4 comments

LLM Usage in Debian: Three Proposals

debian.org · 215 points · 212 comments

Running a 28.9M parameter LLM on an $8 microcontroller

github.com · 286 points · 74 comments

What is the status on continual learning for LLMs?

5 points · 14 comments

HotPin – lossless 120B MoE inference on 24GB RAM (CPU, 50 loc)

7 points · 0 comments

A system prompt to get AI to stop pretending to be human

swiftrocks.com · 35 points · 14 comments

Politician reads AI prompt during assembly

youtube.com · 70 points · 48 comments

Brazilian farmers tokenized dairy cows to get loans, bypassing bank limits

coindesk.com · 60 points · 54 comments

The AI Boom Made Average People More Interesting

alec.is · 7 points · 0 comments

Will it be possible to do LLM Pool Training?

2 points · 2 comments

What happens behind the scenes when we change effort for same LLM models?

11 points · 8 comments

2x, not 10x: coding with LLMs in 2026

obryant.dev · 278 points · 243 comments

Has anyone figured out how to guard LLMs on kernel level?

2 points · 1 comments

TS Compiler Knowledge Graph reducing AI tokens about 90%

github.com · 3 points · 0 comments

The Productivity Mirage

frantic.im · 15 points · 0 comments

Opposing ICE Might Save the Country. It Could Also Ruin Your Life

wired.com · 7 points · 2 comments

"We removed over 80% of Claude Code's system prompt for Opus 5 and Fable 5"

twitter.com · 20 points · 2 comments

AI Meter – Local token usage with energy and water estimates

ai-meter.app · 2 points · 2 comments

GR: Ban LLM Contributions from Debian

lists.debian.org · 7 points · 0 comments

Falsities of LLM Negationists

xlii.space · 5 points · 1 comments

GPL Infection of LLMs?

4 points · 0 comments

Debian launches competing General Resolutions on LLM usage in Debian code

debian.org · 15 points · 1 comments

AMD and Cerebras Launch AI Inference Solution

cerebras.ai · 28 points · 9 comments

Pizauth: An OAuth2 token requester daemon

github.com · 4 points · 0 comments

Why call it "inference" instead of just "model hosting"?

6 points · 9 comments

Sourceminder.org - token-efficient code search

sourceminder.org · 4 points · 2 comments

Is anyone giving out tokens for benchmarking LLMs?

3 points · 1 comments

China Wields Its Rare Earth Leverage over Europe with New Export Controls

nytimes.com · 6 points · 2 comments