Cake and Eat It?

It's extraordinary how the proponents of a free, entrepreneurial system, untrammelled by regulation and in their eyes, entirely meritocratic in nature, are often the first to cry 'foul' when faced with competitors who have second-guessed them and upset their 'own' playground. This week saw the disruptors themselves disrupted: OpenAI [et al] suddenly found that an Asian upstart had become he new kid on the block and had upset the cosy little applecart they had hitherto assumed to be entirely their own. DeepSeek, the Chinese AI that I mentioned last night, has really gotten under the skin of the erstwhile King of AI, which now is accusing it of intellectual property theft; to quote a Tech Radar piece today: 'OpenAI claims to have evidence that DeepSeek, the Chinese startup that has thrown the US tech market into financial turmoil, used the company's proprietary models to train its own open-source LLM, called R1. This would represent a potential breach of intellectual property, as it goes against the OpenAI terms of service agreement.'

How ironic, eh? When the LLM's of the chief players in this 'market' have been training their models by scraping data willy-nilly from the internet [often your data], a considerable proportion of which is the absolute intellectual property of others. Hypocrisy, I say. Whatever the methodology the creators of DeepSeek employed in their data gathering for their LLM training, it is in no way dissimilar from the methodologies employed by the people who are now bitching about them: let's face it, their data is as up for grabs as anyone else's, right? You either believe in totally free enterprise - in which case you just have to accept the downsides as well as the upsides - or you apply the rules and regulations of socialised commerce to your own business. You can't have your cake and eat it, as everyone's Grandma would say; and what, I wonder, is the 'Open' in OpenAI supposed to stand for anyway? Go figure...