AI & machine learning2020 · shared on Hacker News

GPT-3: Language Models Are Few-Shot Learners

Explore the few-shot learning idea behind GPT-3, then compare the community’s early reactions.

GO BEYOND THE ABSTRACT

The conversation around the idea.

Original Hacker News threads. Different perspectives, questions, and things worth thinking about.

Showing up to 100 related threads. A paper may appear in a comment rather than the original submission.

Up to 50 recent source links, in HN item order.

The conversation around the idea.

Links found in the discussion

Claude Sonnet 4 now supports 1M tokens of context

OpenAI's GPT-3 may be the biggest thing since Bitcoin

Terence Tao: The role of small organizations in society has shrunk significantly

GPT-Neo – Building a GPT-3-sized model, open source and free

The New York Times is suing OpenAI and Microsoft for copyright infringement

OpenAI API

The Internet Archive is under a DDoS attack

Apple M1 Max Geekbench Score

Nikola: How to Parlay an Ocean of Lies into a Partnership with GM

GPT-3: Language Models Are Few-Shot Learners

Past Performance is Not Indicative of Future Results (2020)

How GPT3 Works – Visualizations and Animations

Microsoft and OpenAI's close partnership shows signs of fraying

Teaching GPT-3 to Identify Nonsense

An understanding of AI’s limitations is starting to sink in

OpenAI Sold its Soul for $1B

GPT-3

Recent Advances in Natural Language Processing

Are we in an AI Overhang?

GPT-3 has no idea what it’s talking about

AI training method exceeds GPT-3 performance with fewer parameters

Llama 2 – Meta AI

Books in .txt format for AI training purposes

Why GPT-3 Matters

The Pile: An 800GB Dataset of Diverse Text for Language Modeling

Facebook managers trash their own ad targeting in unsealed remarks

Speech and Language Processing (3rd ed. draft)

Banned math book

Show HN: Semantic Calculator (king-man+woman=?)

I asked GPT-NeoX-20B a hundred arithmetic questions

I asked GPT-3 to make a presentation for me

Sundar Pichai: An important next step on our AI journey

Are Humans Intelligent? An AI Op-Ed

Show HN: Humans vs AI – A/B testing GPT-3

My AI Timelines Have Sped Up

Low-code, no-model recommendation system with GPT3

GPT-3 pricing for API usage announced to beta users to go in effect from October

AI language models are struggling to “get” math

The Inherent Limitations of GPT-3

GPT-4 is getting worse over time, not better

Why Mastering Language Is So Difficult for AI

Tesla pulled its latest ‘Full Self Driving’ beta after testers complained

GPT-3 and Scaling Trends

How good is Codex?

GPT‑5.4 Mini and Nano

CheatGPT

Bootstrapping Self Awareness in GPT-4: Towards Recursive Self Inquiry

OpenAI temporarily disables the Browse with Bing beta feature

Intel Announces Aurora GenAI, ChatGPT Competitor with 1T Parameters

Yann LeCun and Andrew Ng: Why the 6-Month AI Pause Is a Bad Idea [video]

Jailbreaking ChatGPT with Dan (Do Anything Now)

What we learned from using GPT for 500k+ classifications

Ask HN: Open source LLM for commercial use?

Sam Altman Exposes the Charade of AI Accountability

Reverse Engineering a Neural Network's Clever Solution to Binary Addition

Google's Live from Paris Event Private/Deleted Immediately

The new Bing will happily give you citations for a pile of nonsense

Ask HN: How does ChatGPT work?

What does the data engineering behind LLMs look like?

Show HN: Turn Any ArXiv Paper into a 200-Page Prerequisite Reading Book

A Systematic Investigation of Commonsense Understanding in Large Language Models

Michael J. Fox Foundation Announces Breakthrough for Parkinson's Biomarker

OpenLLaMA: An Open Reproduction of LLaMA

How Transformers Work

OpenAI Announces Leadership Transition

How LLMs Work

I Asked ChatGPT to Explain Some Jokes to Me

Show HN: The largest collection of ChatGPT jailbreaks on the internet

The entities enabling scientific fraud at scale are large, resilient and growing

Yann LeCun: ChatGPT is 'not particularly innovative'

What were the 40 research papers on the list Ilya Sutskever gave John Carmack?

Open Assistant – project meant to give everyone access to a great chat based LLM

Why China Didn’t Invent ChatGPT

Understanding ChatGPT

Scaling Transformer to 1M tokens and beyond with RMT

Study: ChatGPT Outperforms Physicians in Quality, Empathetic Answers to Patients

StarCoder and StarCoderBase: 15.5B parameter models with 8K context length

NVIDIA introduces TensorRT-LLM for accelerating LLM inference on H100/A100 GPUs