I don't think there’s any stopping it - only slowing it down. To stay ahead in the AI arms race, you need to attract the best talent, build the strongest education system, and invest in top-tier infrastructure. Isolation and restrictions alone aren’t a guarantee of success and could even backfire.
By getting rid of department of education and defunding public education on top of give participation grades, remove competition in the education system, the US will no doubt win the AI race! This is a combined effort of the left and right!
Think of it more as a flex and eye opener to how much in thrall we are to Silicon Valley and everything they claim. The parallels between our worship of Wall Street in the 90s and tech today should give some pause. You have a Chinese company producing a model that can "think" and "reason" far beyond what current available models have done and open sourcing it for everyone to steal copy. Meanwhile OpenAI and Claude, which are not open at all, relied on training that used rote learning (simplification) that we trash Chinese schools in order to rationalize the amount of stem engineers and PHDs they churn out. Not to mention the export controls imposed by a free-trade loving country from both presidential parties. Values and principles don't mean a thing when **** gets real... So yes, everyone knows how to do it now and for cheaper, which is diverting constant attention that no one 100% knows what OpenAI did with o1 or anything else in the last 2 years, even when that was good enough for Trump to plaster his face on a $$$ deal that's just as fugazi as the hype behind LLMs/AI's inevitability.
The threat to open AI and other tech companies is that developers and users will migrate to the open source community, so no one will be able to control the AI and gain massively by monopolizing it. All the countries in the world will be able to use it, not just the rich and powerful ones, democratization of AI is where this is heading.
Did DeepSeek-linked group improperly obtained OpenAI data? Trump’s AI tsar has claimed there’s ‘substantial evidence’ that DeepSeek leaned on OpenAI’s models to develop its own technology
That is most likely true, but they can do it at a fraction of the cost. Didn't open AI crap the internet when it build its model?
https://news.yahoo.com/news/trump-ai-tsar-substantial-evidence-095334172.html David Sacks, Donald Trump’s AI adviser, says China’s DeepSeek may have trained its technology by “sucking knowledge” out of OpenAI’s model in a method known as “distilling”. OpenAI backed up Mr Sacks’s claims, saying there was some evidence of “distillation”, a technique that essentially allows AI models to become much more advanced by using learnings from other models. This would mark a breach of OpenAI’s intellectual property. my 2 cents, this intellectual property theft enabled DeepSeek to replicate OpenAi's work
If they were training based on open AI feeds, that would be wrong, but if it was just craping the internet and there were tons of open AI information on it, then it would be fine. Since it is open source, other teams would be able to verify this fairly quickly, within a few months.
Worlds biggest violins for b****boy Altman whining about an AI company stealing other people's IP in order to get ahead.
"Person of Interest" is a TV show where a billionaire tech genius, Harold Finch, develops a powerful AI surveillance system called "The Machine" that can predict potential crimes by identifying individuals, known as "Persons of Interest," who are about to be involved in violent acts; he then recruits a former CIA operative, John Reese, to secretly intervene and prevent these crimes from happening, raising questions about morality and the limits of surveillance technology in the process. Key points about the show: The Machine: The central element of the show, this AI constantly analyzes data to identify potential criminals and victims, providing their Social Security numbers to Finch and Reese. Harold Finch (Michael Emerson): A reclusive billionaire who created the Machine and is deeply concerned about the ethical implications of its power. John Reese (Jim Caviezel): A former CIA operative who takes on various identities to investigate and intervene in the lives of Persons of Interest. Moral dilemmas: The show explores complex ethical questions about whether it's justified to interfere in people's lives based on predictions of future crimes, even if it means violating their privacy. Rocket River
Not so sure about everyone is betting only on Silicon Valley. I’ve read that some big money is sitting on the sidelines, waiting. DeepSeek shows that there will be many disruptors. It shouldn’t be surprising that a new field has many players. The assumption that only giants like Meta can do this is probably wrong. Since OpenAI, Claude, etc., are private, we don’t really know what they’re doing. They might have highly efficient models that they don’t want to expose to the public yet - why give investors reason to not give more money if they don’t need it as much? Free trade has to be balanced with security. Traditionally (before the current admin), the U.S. has always made exceptions to free trade for national security - nuclear power, for example. AI could arguably be the next "nuclear" power. So, I don’t blame the U.S. for trying to stay ahead in AI through export controls. That said, it might not be effective and could backfire, as mentioned earlier. I didn’t follow the $500B AI infrastructure announcement, so no comment there. But if it turns out to be all marketing and no substance, that wouldn’t be a surprise - we’ve seen plenty of that for decades.
David Sacks is an idiot for the most part. That’s not the definition of IP theft. Data theft? Sure - but hey, whose data is it now?
Tell me that you don't own enough Mag Seven stocks to be a part of the "everyone" without telling me.
care to give your definition OpenAI says it has uncovered evidence that DeepSeek used its proprietary models to train a competing open-source model, potentially violating the company's terms of service.
Data theft - unauthorized (or not approved) access, copying, use of data IP theft - unlawfully access, copying and use of protected IP such as patents Model distillation, according to OpenAI is leveraging "the outputs of a large model to fine-tune a smaller model". Output of a model is data, not IP. https://platform.openai.com/docs/guides/distillation
