Sweatshop Data Is Over

https://www.mechanize.work/blog/sweatshop-data-is-over/

27•whoami_nr•2h ago

Comments

jrimbault•1h ago

> This meant that while Google was playing games, OpenAI was able to seize the opportunity of a lifetime. What you train on matters.

Very weird reasoning. Without AlphaGo, AlphaZero, there's probably no GPT ? Each were a stepping stone weren't they?

phreeza•1h ago

Transformers/Bert yes, alphago not so much.

vonneumannstan•1h ago

>Very weird reasoning. Without AlphaGo, AlphaZero, there's probably no GPT ? Each were a stepping stone weren't they?

Right but wrong. Alphago and AlphaZero are built using very different techniques than GPT type LLMs. Google created Transformers which leads much more directly to GPTs, RLHF is the other piece which was basically created inside OpenAI by Paul Cristiano.

msp26•44m ago

OpenAI's work on Dota was also very important for funding

jimbo808•42m ago

Google Brain invented transformers. Granted, none of those people are still at Google. But it was a Google shop that made LLMs broadly useful. OpenAI just took it and ran with it, rushing it to market... acquiring data by any means necessary(!)

losteric•1h ago

> Despite being trained on more compute than GPT-3, AlphaGo Zero could only play Go, while GPT-3 could write essays, code, translate languages, and assist with countless other tasks. The main difference was training data.

This is kind of weird and reductive, comparing specialist to generalist models? How good is GPT3’s game of Go?

The post reads as kind of… obvious, old news padding a recruiting post? We know OpenAI started hiring the kind of specialist workers this post mentions, years ago at this point.

9rx•57m ago

> This is kind of weird and reductive, comparing specialist to generalist models

It is even weirder when you remember that Google had already released Meena[1], which was trained on natural language...

[1] And BERT before it, but it is less like GPT.

rcxdude•37m ago

Also, the main showcase of the 'zero' models was that they learnt with zero training data: the only input was interacting with the rules of the game (as opposed to learning to mimic human games), which seems to be the kind of approach the article is asking for.

rob74•1h ago

It's kind of reassuring that the old adage "garbage in, garbage out" still applies in the age of LLMs...

Flat design vs. realistic ("skeuomorphic") design

The Work of Raj Chetty

Show HN: MIT License Rust Accessibility-Based Computer Use SDK+mcp

Algorithmic Collusion of Pricing and Advertising on E-Commerce Platforms

What to Watch (Or Not): Ballard, Perfect Days, Billy Joel

Open SWE: An Open-Source Asynchronous Coding Agent

Show HN: Create DJ-like transitions between any two songs

A defense of learning Latin and Greek

Show HN: Train – AI Workout Companion That Helps You Train Smarter

What is it? White balloon object spotted over Anchorage

A new worst coder has entered the chat: vibe coding without code knowledge

Legendary GPU architect Raja Koduri's startup leverages RISC-V and targets CUDA

Retab: The developer starter pack for document processing

Ditching GitHub

Is Universal Basic Income Effective? Not Really

PyPI: Preventing ZIP parser confusion attacks on Python package installers

Live: GPT-5

The Sunlight Budget of Earth

Show HN: FocusTree – a simple task app (prototype), free open source

GPT-5 Coding Examples

I built a site to surface mind-blowing, underrated websites

ZFSBootMenu

1h The NDA for the Framework Desktop reviews has been lifted

Think Linux desktop market share isn't over 6%? This scan says otherwise

Is your brain necessary for consciousness?

US to levy 100% tariff on imported chips, but some firms exempt

EyJaafCsubstantially: Cramming English words into JSON web tokens

Go-2025-3849: Incorrect results returned from Rows.Scan in database/SQL

TimescaleDB 2.21 – 42× Faster DELETEs

Password Pusher: Share secrets securely with self-deleting links and audit logs