'Fair play' is a dimension of competition. They aren't orthogonal attributes.
You can compete fairly or unfairly.
In any event, distillation seems to be in a grey enough area that reasonable people can disagree (IMO). It strikes me as more akin to theft than to fair competition, but it doesn't seem like it should be treated as criminal. Rather, I think the onus should be on the labs to defend themselves against it.
All that said, Jensen's take is clearly quite biased.
Hasn’t it already been ruled by a court that LLM outputs are not subject to copyright? It opens a can of worms to say that LLM companies can dictate what their outputs are used for.
I think copyright is mostly not relevant. Contracts operate mostly independent of copyright law (in the U.S.). If OpenAI puts in their license something to the effect of "you may not train your LLMs on these outputs" and/or "your access is limited in these ways", but you violate the license, they can and should block your access and sue you. These are things that can be monitored and enforced under existing law.
And, in fact, they already do this:
- OpenAI [1]: "[you may not] Use Output to develop models that compete with OpenAI."
- Google [2]: "You may not use the Services to develop machine learning models or related technology."
- Anthropic [3]: "[You may not use our services] to develop any products or services that compete with our Services, including to develop or train any artificial intelligence or machine learning algorithms or models or resell the Services"
Papazsazsa · · focus · HN ↗
next_xibalba · · focus · HN ↗
You can compete fairly or unfairly.
In any event, distillation seems to be in a grey enough area that reasonable people can disagree (IMO). It strikes me as more akin to theft than to fair competition, but it doesn't seem like it should be treated as criminal. Rather, I think the onus should be on the labs to defend themselves against it.
All that said, Jensen's take is clearly quite biased.
edgyquant · · focus · HN ↗
next_xibalba · · focus · HN ↗
And, in fact, they already do this:
- OpenAI [1]: "[you may not] Use Output to develop models that compete with OpenAI."
- Google [2]: "You may not use the Services to develop machine learning models or related technology."
- Anthropic [3]: "[You may not use our services] to develop any products or services that compete with our Services, including to develop or train any artificial intelligence or machine learning algorithms or models or resell the Services"
[1] <a href="https://openai.com/policies/row-terms-of-use/" rel="nofollow">https://openai.com/policies/row-terms-of-use/
[2] <a href="https://policies.google.com/terms/generative-ai/archive/20230314" rel="nofollow">https://policies.google.com/terms/generative-ai/archive/2023...
[3] <a href="https://www.anthropic.com/legal/consumer-terms" rel="nofollow">https://www.anthropic.com/legal/consumer-terms