karfly

Joined 6 years ago

Karma
37
Submissions
12
Member since
Apr 8
Building stuff Bio: karfly.github.io
View on Hacker News

Recent submissions

Comment · 1 year ago
tl;dr:

1. Copy the ChatGPT share link

2. Prefix with "pdf": "chatgpt.com/share/..." -> "pdfchatgpt.com/share/..."

3. Instantly get clean PDF, Markdown, or HTML

Story: I often share ChatGPT dialogs with colleagues, but share links usually fail: loading issues, cookie popups, or temporary access. PDF/Markdown sharing solves this completely for me.

Existing Chrome extensions generate messy PDFs and aren't mobile-friendly.

So I built a reliable ChatGPT-to-PDF exporter, spending a fun week reverse-engineering OpenAI APIs and vibing on the PDF readability and design. The app handles code blocks (with wrapping), LaTeX, long tables, inline citation links, images, emojis, special symbols, and more.

It also generates Markdown for easy dialog copy-pasting into Cursor or other tools.

How to use:

A. Go to pdfchatgpt.com and paste your ChatGPT share link

B. Use URL prefix lifehack from "tl;dr" section above

Feel free to report edge cases or parsing errors in the comments.

And it's free.

pdfchatgpt.com

Comment · 3 years ago
Comment · 3 years ago
IMO all RLHF stuff is mainly about aligning model not to reply with offensive and inappropriate answers, but NOT about making model's answers better.
Comment · 3 years ago
You're totally true. But it turned out, that proper prompting (like packing dialog context into the prompt) worked great.

I spent many hours with orig ChatGPT and with this recreated version. The main difference I have found is that the recreated ChatGPT is more inclined to ask questions on questions (maybe can be fixed with more prompt engineering). I didn't find any major differences in the quality or usefulness of the answers.

Comment · 3 years ago
It's true – you need to pay for OpenAI API. With my typical usage frequency of ChatGPT it's going to cost me about 3$/month.
Comment · 3 years ago
1. You’re right, it’s not ChatGPT exactly. ChatGPT = GPT-3.5 + finetuning on conversational data + alignment (RLHF). But I found out, that GPT-3.5 + proper prompting gives comparable results and works really good.

2. The context of the conversation is preserved by adding whole (or some parts of) dialog to the prompt.

Comment · 3 years ago
We all love chat.openai.com, but...

It's TERRIBLY laggy, has daily limits, and is only accessible through an archaic web interface.

This repo is ChatGPT re-created with GPT-3.5 LLM as Telegram Bot. And it works great.

In addition it supports special modes like "Code Assistant" and "Movie Expert".

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton