Bringing the best of AI to college students for free
blog.google·16h·
Discuss: Hacker News
🧩type theory
YC Startups Use AI: Agents, OCR, and Prompt Engineering with Mercoa (YC W23)
aiengineering.report·16h·
Discuss: Hacker News
🧩type theory
MLE-STAR: A state-of-the-art machine learning engineering agents
research.google·5d·
Discuss: Hacker News, Hacker News
🧩type theory
More scientific papers being written with help of ChatGPT—especially in computer science
phys.org·1d·
Discuss: Hacker News
🧩type theory
Some AI tools don’t understand biology yet
arstechnica.com·17h
🧩type theory
Taking the bitter lesson to heart for speech-to-speech models
metavoice.io·10h·
Discuss: Hacker News
🧩type theory
DSPy Simba Explained
blog.mariusvach.com·1d·
Discuss: Hacker News
🧩type theory
Fine-tuning with GPT-OSS and Hugging Face Transformers
cookbook.openai.com·1d·
Discuss: Hacker News
🧩type theory
Grok 4 Dominates AI Chess Tournament Day 1
chess.com·23h·
Discuss: Hacker News
🧩type theory
I Am Not a Calculator
massicotte.org·17h·
Discuss: Hacker News
🧩type theory
Eval-maxing an AI FFmpeg command generator
getkiln.ai·12h·
Discuss: Hacker News
🧩type theory
Speaking in "LLM Idioms"
blog.promptlayer.com·2d·
Discuss: Hacker News
🧩type theory
AI Plays Risk – Lessons from a silly benchmark
andreasthinks.me·18h·
Discuss: Hacker News
🧩type theory
GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning
news.ycombinator.com·2d·
Discuss: Hacker News
🧩type theory
Good context leads to good code: How we built an AI-Native Eng Culture
blog.stockapp.com·14h·
Discuss: Hacker News
🧩type theory
Unsupervised Model Improvement via Internal Coherence Maximization
huggingface.co·3d·
Discuss: Hacker News
🧩type theory
Testing LLM Responses: A Fast, Cost-Effective Alternative to LLM-as-Judge
joywrites.dev·4d·
Discuss: Hacker News
🧩type theory
NeuralMorse – Reinventing Morse Code with Neural Networks
masatohagiwara.net·1d·
Discuss: Hacker News
🧩type theory
Vibe coding to Chomksy: will linguistics pop off in a world full of prompting?
aymannadeem.com·16h·
Discuss: Hacker News
🧩type theory
We beat GPT-4o's baseline with a simple re-prompting loop
aimon.ai·1d·
Discuss: Hacker News
🧩type theory