LLMs, Innovation, and Capitalism - by Belinda Li - Reboot
joinreboot.org · 3,053 words · saved by 1 readers
What happens to innovation when science gets privatized?
For a long time, much of AI research was default-open. OpenAI’s GPT is an innovation on the Transformer model, which Google researchers open-sourced in 2017. Two years later, OpenAI set a new precedent by initially choosing not to release GPT-2 to the public—describing the model as simply “too dangerous.” Today, both AI research and corporate gatekeeping has accelerated massively. NLP researcher Belinda Li attributes these locked doors not to “safety,” but to capitalism. In today’s guest essay, she makes the democratic case for open science, explaining how the closed model shapes—and…
related reading
- Large Language Models, Innovation, and Capitalism - Belinda Zou Libelindal.github.io
- Google "We Have No Moat, And Neither Does OpenAI"semianalysis.com
- What comes next with open models - by Nathan Lambertinterconnects.ai
- Society's Technical Debt and Software's Gutenberg Momentskventures.substack.com
- AI Revolution - Transformers and Large Language Models (LLMs)blog.eladgil.com
- Training Imperativesdan.io
- Why didn't DeepMind build GPT3?rootnodes.substack.com
- Report: The Openness of AI | A Contrary Research Deep Dive | Contrary Researchresearch.contrary.com
- Bernie’s bad bet on OpenAIsubstack.com
- Thoughts on AI in academiatheinfinitesimal.substack.com
- Things we learned about LLMs in 2024simonwillison.net
- Open models in perpetual catch-up - by Nathan Lambertinterconnects.ai