And this is why OpenAI should have remained a non-profit and gone full opensource on everything and try to get others to do the same at least with respect to AI/ML software. Proprietary software...
Having these conversations is difficult, as it involves talking candidly about proprietary systems and it’s unclear who to reach out to in specific organizations to discuss such models and what the appropriate processes are for inter-org discussion about unreleased research
And this is why OpenAI should have remained a non-profit and gone full opensource on everything and try to get others to do the same at least with respect to AI/ML software. Proprietary software and "trade secrets" belong in the realm of business, not in the realm of AI research.
The University of Oregon is developing a series of “bias probes” to analyze bias within GPT-2.
The University of Texas at Austin is studying the statistical detectability of GPT-2 outputs after fine-tuning the model on domain-specific datasets, as well as the extent of detection transfer across different language models.
At least two universities are working on some kind of defense against misinformation and bias generated from GPT-2. That's nice.
Judging from their release of MuseNet and this latest update on GPT-2, we're rapidly approaching a world where creative people such as composers, musicians, and writers are paid far worse than they are now ("why should I pay you $100 for an article when I can prompt GPT-2 to write it for $0?", "why do I need to pay you royalties of $1000s for the music we use in our movie when we can pay MuseNet $0?") due to competition with these AI systems, or they're paid far better like fine works of art. I am hoping these systems become complements and are used by creatives to augment their skills, rather than the sad image of turning music and writing into uber-commodities.
And this is why OpenAI should have remained a non-profit and gone full opensource on everything and try to get others to do the same at least with respect to AI/ML software. Proprietary software and "trade secrets" belong in the realm of business, not in the realm of AI research.
At least two universities are working on some kind of defense against misinformation and bias generated from GPT-2. That's nice.
Judging from their release of MuseNet and this latest update on GPT-2, we're rapidly approaching a world where creative people such as composers, musicians, and writers are paid far worse than they are now ("why should I pay you $100 for an article when I can prompt GPT-2 to write it for $0?", "why do I need to pay you royalties of $1000s for the music we use in our movie when we can pay MuseNet $0?") due to competition with these AI systems, or they're paid far better like fine works of art. I am hoping these systems become complements and are used by creatives to augment their skills, rather than the sad image of turning music and writing into uber-commodities.