Current artificial intelligence models from labs like OpenAI are unlikely to lead to major scientific breakthroughs, a tech co-founder said, pouring cold water on some of the hype around the technology and claims by major figures in the field.

The comments by Thomas Wolf, co-founder of $4.5 billion AI startup Hugging Face, are in stake contrast to those by major names in AI including OpenAI boss Sam Altman and Anthropic CEO Dario Amodei.

When Wolf talks about scientific breakthroughs, he means novel ideas like those at a Nobel Prize level. Examples including Nicolaus Copernicus who theorized the sun was at the center of the universe and other planets move round it.

Wolf explained a couple of issues with chatbots right now. The first is that these products like ChatGPT and others often agree or align with the person prompting it. Think back to if you’ve asked a chatbot a prompt and it will tell you how interesting or great that question is.

The second is that the models underpinning these chatbots are designed to “predict the most likely next token” or “word” in a sentence.

However, he noted two key traits of scientists. The first is that scientists who make major breakthroughs are often contrarian and question what others are saying.

“The scientist is not trying to predict the most likely next word. He’s trying to predict this very novel thing that’s actually surprisingly unlikely, but actually is true,” Wolf said.

The Hugging Face co-founder has been thinking about this topic for the last few months. His interest was sparked after he read an essay penned by Anthropic’s Amodei, who posited that “AI-enabled biology and medicine will allow us to compress the progress that human biologists would have achieved over the next 50-100 years into 5-10 years.”

That got Wolf thinking about the state of AI and how this won’t be possible, in his view, with the current crop of models.

Wolf said that these chatbots and tools will likely be used as a sort of “co-pilot for a scientist” where they are used for research to help the human generate new ideas.

To some extent, this has been happening already. Google DeepMind’s AlphaFold product has helped to analyze protein structures which the company has promised could aid scientists in discovering new drugs.

But there are some new startups that are hoping to take AI one step further into being able to make scientific breakthroughs, including Lila Sciences and FutureHouse.

Taiwan will not accept Washington’s proposal to locally manufacture half the chips it currently supplies to the U.S., the island’s top trade negotiator said.

Speaking to reporters, Cheng Li-chiun, also the country’s vice premier, said on Wednesday that the proposal for a “50-50” split in semiconductor production was not even discussed, as she returned from trade talks in the U.S., according to Taiwan’s Central News Agency.

Cheng said the talks were focused on lowering tariff rates, securing exemptions from tariff stacking — additional duties — and reducing levies on Taiwanese exports. Taiwan currently faces a “reciprocal” tariff rate of 20%.

Washington has held discussions with Taipei about the “50-50” split in semiconductor production, which would cut American reliance on Taiwan, Commerce Secretary Howard Lutnick said last weekend in an interview to NewsNation, adding that currently 95% of the U.S. demand was met via chips produced within Taiwan.  

“My objective, and this administration’s objective, is to get chip manufacturing significantly onshored — we need to make our own chips,” Lutnick said. “The idea that I pitched [Taiwan] was, let’s get to 50-50. We’re producing half, and you’re producing half.” 

U.S. President Donald Trump had also taken aim at the island’s dominance in chips earlier this year, accusing it of “stealing” the U.S.’ chip business.

The Office of the U.S. Trade Representative and Taiwan’s Ministry of Economic Affairs did not immediately respond to CNBC’s request for comments.

Lutnick’s proposal has been condemned by Taiwan’s politicians, with Eric Chu, chairman of the island’s principal opposition party Kuomintang, calling it “an act of exploitation and plunder,” according to the Central News Agency report.

“No one can sell out Taiwan or TSMC, and no one can undermine Taiwan’s silicon shield,” Chu said, referring to Taiwan Semiconductor Manufacturing Company, the world’s leader in advanced chip manufacturing.

Taiwan’s critical position in global chips production is believed to have assured the island nation’s defense against direct military action from China, often referred to as the “Silicon Shield” theory.

In his NewsNation interview, Lutnick downplayed the “Silicon Shield,” arguing that Taiwan would be safer with more balanced chip production between Washington and Taipei. Beijing views the democratically governed island of Taiwan as its own territory and has vowed to reclaim it by force if necessary, while Taipei rejects those claims. 

Taiwan People’s Party Chairman Huang Kuo-chang reportedly called Lutnick’s proposal an attempt to “hollow out the foundations of Taiwan’s technology sector.”

Shares of South Korean chip heavyweights Samsung Electronics and SK Hynix surged Thursday, a day after the two companies partnered with artificial-intelligence major OpenAI as part of the U.S. firm’s Stargate initiative.

Samsung shares hit their highest since January 2021, rising over 4%, while SK Hynix stock surged more than 9% — highest since 2000.

OpenAI said in a statement that this partnership will “focus on increasing the supply of advanced memory chips essential for next-generation AI and expanding data center capacity in Korea.”

The ChatGPT-maker said the two South Korean firms plan to scale up production of advanced memory chips, which are critical to power its AI models.

The announcement came as OpenAI CEO Sam Altman met with South Korean President Lee Jae Myung in Seoul, and the top leaders at Samsung and SK Hynix.

OpenAI has also signed a series of agreements to explore developing next-generation AI data centers in South Korea, including with the Korean Ministry of Science and ICT, telecommunications operator SK Telecom, as well as with Samsung subsidiaries.

Earlier this month, SK Hynix announced that it was ready to mass-produce its next-generation high-bandwidth memory chips, cementing its leading position in the AI value chain. HBM is a type of memory that is used in chipsets for artificial-intelligence computing, including in chips from global AI giant Nvidia — a major client of SK Hynix. 

HBM4 chips are expected to be the main AI memory chip needed for Nvidia’s next-generation Rubin architecture — a more powerful AI chip for global data centers.

SK Hynix has been a main chip supplier to Nvidia, while rival Samsung has reportedly been working to get its HBM4 chips certified by Nvidia. 

Samsung has traditionally been the market leader in memory, but its position has been threatened by SK Hynix that has taken a lead in the HBM space. A report from Counterpoint Research in July found that SK Hynix had caught up with Samsung’s memory revenues in the second quarter, with both now vying for the top position in the global memory market. 

Samsung’s second-quarter earnings missed expectations, as profits from its chip business declined almost 94% year on year, although its Chief Financial Officer Soon-cheol Park said that the company expects a rebound in the second half of the year.

— CNBC’s Dylan Butts contributed to this report.