TheChek News Flow

Trade officials promise details of ‘agreement’ on Monday in bid to de-escalate tensions

Treasury secretary Scott Bessent on Sunday said the US made “substantial progress” over two days of trade talks with Chinese officials in Geneva, in the first sign that Washington and Beijing may start to ratchet down tensions.

“We will be giving details tomorrow, but I can tell you that the talks were productive,” Bessent told reporters after he and US trade representative Jamieson Greer finished their meetings with Chinese vice-premier He Lifeng. 

Greer said it was “important to understand how quickly we were able to come to agreement, which reflects that perhaps the differences were not so large as maybe thought” and added that there had been a “lot of groundwork”. 

The optimistic comments from the US negotiating team were the first sign that the countries might de-escalate the trade war that has roiled financial markets and triggered concerns about global supply chains.

The US has placed a 145 per cent tariff on goods from China while Beijing has retaliated with its own 125 per cent levy. 

The Chinese embassy in Washington did not respond to a request for comment about China’s stance on the talks. The US and Chinese negotiation teams met at the Geneva residence of the Swiss ambassador to the UN. 

Following the first day of negotiations on Saturday, Trump posted on his Truth Social platform that the US and China had made “great progress”. He added: “A total reset negotiated in a friendly, but constructive manner.”

Treasury secretary said substantial progress had been made with details to come Monday

Key Points

GENEVA—Senior economic officials from the U.S. and China concluded their weekend trade talks Sunday with no deal announced to end the continuing trade war between the countries.

Treasury Secretary Scott Bessent and U.S. Trade Representative Jamieson Greer, who led the U.S. delegation, told reporters that substantial progress was made during the marathon discussions with their Chinese counterparts, led by Chinese Vice Premier He Lifeng. The talks were “productive,” Bessent said, adding that the U.S. side will share more details Monday.

The Chinese Embassy in Washington didn’t immediately respond to questions.

The talks spanned at least eight hours Saturday and several hours Sunday.

President Trump posted on social media after the first day of talks concluded, saying great progress was made. “A very good meeting today with China, in Switzerland. Many things discussed, much agreed to. A total reset negotiated in a friendly, but constructive, manner,” he wrote, without offering details.

Chinese leader Xi Jinping decided to dispatch his top aides, The Wall Street Journal previously reported. Notably, in addition to He, his economic right-hand man, Xi’s public-security minister, Wang Xiaohong, was also expected to be part of the Chinese entourage.

The talks set the stage for a potential thawing of trade relations between the world’s two largest economies. Since starting his second term, Trump has slapped 145% tariffs on Chinese goods while Beijing has hit back with 125% duties on American products.

Though some U.S. and China officials departed the Saturday meeting early, Bessent and Greer remained for at least an extra hour to continue talks with the remaining Chinese delegation, according to one of the people briefed on the discussions. Bessent, Greer and other officials from both countries who remained were expected to have dinner at the meeting site. Earlier, Greer and Bessent were spotted having lunch with their respective teams, separate from the Chinese delegation at a nearby Italian restaurant.

The talks took place at the residence of the Swiss ambassador to the United Nations. The sprawling property is behind a massive gate on a small, one-way street. Dozens of law-enforcement officials, including the U.S. Secret Service, provided security. Both U.S. and Chinese officials were seen being escorted to the site in black vehicles with tinted windows.

China’s state-run news agency Xinhua described Saturday’s meetings as “an important step toward resolving the dispute” between the two world powers.

Trump has opened the door for lower tariffs. He said in recent days that they couldn’t get much higher than the current 145%, so that it was likely they would eventually come down. And Friday, he suggested lowering tariffs to 80%. On Friday night, Commerce Secretary Howard Lutnick told Fox News that the president is expected to keep significant so-called reciprocal tariffs on trade with China but may settle near 34%, which is the rate Trump announced on April 2.

Trump has said he wants fairer trade with Beijing—the president has often criticized China’s manufacturing power—and that he wants to curb its role in fentanyl trafficking into the U.S. It is likely the U.S. would want to see some commitments from China on either front before lowering tariffs.

Some administration officials viewed Trump’s 80% idea as a public message to encourage Bessent to try to get closer to completing an arrangement that would lower China’s tariff rate to around 80%, the Journal previously reported.

Billy Evans has two children with the Theranos founder, who is in prison for fraud. He’s now trying to raise money for a testing company that promises “human health optimization.”

Elizabeth Holmes is in prison for defrauding investors through her blood-testing company, Theranos. In the meantime, her partner is starting one of his own.

Billy Evans, who has two children with Ms. Holmes, is trying to raise money for a company that describes itself as “the future of diagnostics” and “a radically new approach to health testing,” according to marketing materials reviewed by The New York Times.

If that sounds familiar, it’s because Theranos similarly aimed to revolutionize diagnostic testing. The Silicon Valley start-up captured the world’s attention by claiming, falsely as it turned out, to have developed a blood-testing device that could run a slew of complex lab tests from a mere finger prick.

Mr. Evans’s company is named Haemanthus, which is a flower also known as the blood lily. It plans to begin with testing pets for diseases before progressing to humans, according to two investors pitched on the company who spoke on the condition of anonymity because they had agreed to keep the plans secret. Mr. Evans’s marketing materials, which lay out hopes to eventually raise more than $50 million, say the ultimate goal is nothing short of “human health optimization.”

A photo provided to potential investors of the start-up’s prototype bears more than a passing physical resemblance to Theranos’s infamous blood-testing machine, variously known as the Edison or miniLab. The device that Mr. Evans’s company is developing is a rectangular contraption with a door, a digital display screen and what the investor materials describe as tunable lasers inside.

Haemanthus says its device will test blood as well as saliva and urine.

The marketing documents provided with the photo say there is “no regulatory oversight — U.S.D.A. confirmed in writing.”

It’s not clear what the company means by that. A spokesman for the U.S. Department of Agriculture, Seth W. Christensen, said he was not able to confirm whether the agency had corresponded with Haemanthus. “U.S.D.A. does regulate vet diagnostics,” including blood testing, Mr. Christensen said.

Mr. Evans responded in an interview, “When you’re in stealth, you’re trying to be in stealth. They aren’t going to find anything associated with the name Haemanthus.” Mr. Evans sent a partially redacted document from the U.S.D.A. that said, “It does not appear that the proposed product is within the regulatory jurisdiction" of the Center for Veterinary Biologics, which is a part of the U.S.D.A.

Mr. Evans, the 33-year-old heir to a California hotel fortune who met Ms. Holmes while federal authorities were investigating her, has not publicly discussed the new venture. The documents indicate he has already assembled roughly 10 employees. He describes his employment on social media simply as working for a “stealth start-up.”

James W. Breyer, the well-known venture capitalist and early investor in Facebook, said his team had been asked to put in money and decided against it “for many of the same reasons we passed twice on Theranos.”

“In diagnostics, we’ve long held that the difference between a compelling story and a great company lies in scientific defensibility and clinical utility,” he wrote in an email.

If sequels are de rigueur in the so-called disruptive world of technology, this one is particularly bold. Theranos became one of the most celebrated start-ups in the globe last decade and attracted both big-time investors (Rupert Murdoch, Larry Ellison) and a board of advisers that included Henry Kissinger.

Ms. Holmes, often clad in a black turtleneck that invited comparisons to the Apple founder Steve Jobs, was feted on magazine covers, and at the White House.

Few knew that Theranos’s technology could not diagnose hundreds of conditions it claimed it could. As was chronicled in The Wall Street Journal, a best-selling book, a podcast, television series and later criminal proceedings, Theranos was largely using third-party technology to run rudimentary assays — when it did any testing at all. Patients received false diagnoses. The company crumbled ahead of Ms. Holmes’s indictment for fraud.

Ms. Holmes, who has always maintained that she is innocent, was convicted of fraud in 2022 and sentenced to 11 years in prison. She is incarcerated in a Bryan, Texas, federal prison.

Mr. Evans’s idea for Haemanthus traces back at least a year and a half, when he incorporated the company in Delaware, according to public corporate filings. Documents filed in Delaware and Texas show that its offices have been at various addresses in the trendy South Lamar neighborhood of Austin, Texas, where Mr. Evans lives with his and Ms. Holmes’s two children.

Haemanthus began by soliciting $3.5 million in funding from friends and family and this spring began reaching out to other well-to-do backers in Austin and the San Francisco Bay Area for an additional $15 million, according to the investor materials.

The billionaire Michael Dell’s investment firm turned down the effort, according to two people briefed on the outreach.

The one investor who could be identified in public records is Matthew E. Parkhurst, the part owner of a Mediterranean tapas bar in downtown Austin and other investments. Mr. Parkhurst did not respond to requests for comment.

Much of the Haemanthus executive team hails from Luminar, a struggling self-driving car company where Mr. Evans worked for two years, according to his LinkedIn profile.

Pet health care is the first market Mr. Evans’s company aims to address. The start-up has thus far received one patent.

According to the company’s marketing materials and patent, the Haemanthus device will use a laser to scan blood, saliva or urine from pets and analyze the samples on a molecular level. In a matter of seconds, the marketing material said Mr. Evans’s machine would be able to identify and qualify biomarkers such as glucose and hormones, and deploy what the company calls deep learning models to detect cancer and infections.

Animal medicine has grown into a colossal industry as private-equity firms have increasingly acquired and consolidated independent veterinary practices.

Pet cancer screenings alone are a multibillion-dollar market. Edgemont Partners, a health care investment bank, describes it as a “recession-proof industry.”

Haemanthus told investors that it had roughly two dozen advisers, including veterinarians and diagnosticians, though it did not name them.

Haemanthus’s materials say the long-term goal is to develop a stamp-size, wearable version of the product for humans. “Based on our experience and partner input,” it says, that will require three years and $70 million.

The investor presentation makes no mention of Mr. Evans’s connection to Ms. Holmes.

A Chinese team has solved a decades-old mystery that stumped the mathematical world – by resolving the ‘doomsday hypothesis’

Chinese mathematicians have resolved the loose end of a “doomsday hypothesis” that had long puzzled the mathematical community.

The feat was made possible through major computational methods which could be applied to further problems in the field.

As a result of the breakthrough, the mathematical world finally has proof that manifolds of Kervaire invariant one do exist in dimension 126, ending a decades-long mystery.

The Kervaire invariant is a function that measures whether a smooth framed manifold, or a topological space or shape that can have curvature but locally appears flat, can be converted into a sphere through “surgery”, a concept introduced by American mathematician John Milnor in 1950.

If it can be converted into a sphere, the invariant evaluates to zero. The Kervaire invariant problem seeks to discover dimensions in which the answer is non-zero, or one, meaning they can host strange shapes that do not convert into a sphere.

The paper, which has not undergone peer review, was written by Wang Guozhen and Lin Weinan from the Fudan University Shanghai Centre for Mathematical Sciences and Xu Zhouli from the University of California Los Angeles (UCLA).

“We establish the existence of smooth framed manifolds with Kervaire invariant one in dimension 126, thereby resolving the final case of the Kervaire invariant problem,” the team said in a preprint paper published on the arXiv repository in December.

“We conclude that smooth framed manifolds with Kervaire invariant one exist in and only in dimensions 2, 6, 14, 30, 62, and 126.”

The problem has stumped mathematicians for decades. In 1963, work published by Michel Kervaire and Milnor led to proof that the manifolds of the Kervaire invariant one did exist in dimensions 6 and 14.

By 1984, it was proved that this was also the case in dimensions 30 and 62, following a pattern of numbers in the function 2 less than 2 to the power of n.

Mathematicians believed this should be the case for the next numbers that followed in this pattern, such as 126 and 254, but progress stopped at dimension 62 for decades.

The assumption that manifolds of Kervaire invariant one must exist in higher dimensions was used to build mathematical propositions about exotic shapes, but the fact that this assumption could be false became known as the doomsday hypothesis.

In 2009, American mathematician Michael Hopkins from Harvard University and his collaborators showed that manifolds of Kervaire invariant one “exist only in dimensions at most 126”, and do not exist in dimensions 254 and higher. This meant that the doomsday hypothesis had come true.

While their proof had solved a long-standing problem in algebraic topology, the mystery of whether it existed in dimension 126 remained over the next 15 years.

Now the mathematical world finally has proof that manifolds of Kervaire invariant one do exist in dimension 126, tying up the loose ends of a decades-long mystery.

The “jaw-dropping” proof was made possible through both theoretical insights and computer calculation methods developed by the researchers, Hopkins told Quanta Magazine, which is funded by Simons Foundation, a mathematics and basic sciences organisation.

Before their proof was released, Hopkins said that mathematicians had thought such a “heroically computational” achievement was still far out of reach.

All three authors of the study are Peking University alumni with a focus on algebraic topology. Both Xu and Wang completed their bachelor’s and master’s degrees at the university in 2011, and Lin completed his bachelor’s in 2015.

Xu went on to receive his doctorate in mathematics from the University of Chicago in 2017, after which he was an instructor at the Massachusetts Institute of Technology (MIT) until 2020.

He was a professor at the University of California San Diego from 2020 to 2024, before joining UCLA as a professor last November, according to his website.

When Xu arrived at the University of Chicago, his adviser proposed the problem of dimension 126 and introduced him to Mark Mahowald, who was a professor at Northwestern University at the time.

According to Quanta Magazine, Mahowald rejected his proposal and suggested Xu study related problems in lower dimensions instead.

Wang received his doctorate from MIT in 2015, before becoming a postdoctoral fellow at Copenhagen University until 2016.

He joined Fudan University as a postdoctoral fellow in 2016 and became a professor in 2020, according to his website.

Xu and Wang, who studied together, are long-time collaborators who have published several papers together.

Lin, who is currently a tenure-track young investigator at Fudan University, received his doctorate from the University of Chicago in 2021, after which he spent some time as an instructor at Peking University.

Solving the issue of dimension 126 involved studying the stable homotopy groups of spheres, which describe how points on high-dimensional spheres can be deformed into lower ones.

The Adams spectral sequence is a tool or atlas of dots used to help mathematicians understand this complicated world.

It has long been known that if one specific dot in the 126th column of the atlas can survive until the final page of the atlas, then there are manifolds in this dimension that do not convert into a sphere.

The issue was that there are 105 hypothetical ways for it to disappear before reaching it, which had to be eliminated, according to Quanta Magazine.

Xu and Wang developed new computational techniques and passed them to Lin, who wrote a programme to rule out 101 of the possibilities. After a year of work, the authors were able to rule out the last four.

Stefan Hoops says trade war has triggered long overdue spending push on the continent

The chief executive of Deutsche Bank’s asset manager DWS has argued that US President Donald Trump’s tariff policies have proved a powerful, if uncomfortable, catalyst for long-overdue economic reforms in Europe.

In an interview with the Financial Times, Stefan Hoops said that while rising tariffs and geopolitical tension had rattled global markets, they had also jolted Europe out of “complacent, lethargic behaviour” surrounding infrastructure and defence spending.

“Europe has always been about ability versus willingness to really compete . . . We have plenty of savings, we have great innovation, and we have fiscal space. There are lots of things we could do, we just never did them,” Hoops added.

“The decisions, the announcements by the US president, have simply pushed Europe out of its pretty complacent, lethargic behaviour.”

Hoops, who took over at the €1tn asset manager in 2022, said the EU’s rearmament push and Germany’s plan to unleash defence spending and overhaul its infrastructure were irreversible shifts following Trump’s threats to end US security guarantees. “This is putting Germany on a fundamentally different growth path than we otherwise would have been on,” he said.

Trump’s geopolitical realignment, which includes an ambiguous position towards defending European allies if they came under attack, has led to major defence spending plans from countries including Germany, France and Britain.

Indicating that Trump’s antagonistic approach may be crude but effective, he said: “You might question the method, but the outcome is that Europe is finally starting to move.”

Hoops acknowledged that Trump’s behaviour could have long-term costs, potentially damaging international co-operation on climate, research, and trade. But on balance, he said Europe was likely to perform better economically as a result of new spending plans, avoiding the kind of recession that many economists fear the US could fall into as a result of trade-related turmoil.

The US Federal Reserve said on Wednesday that the uncertainty about the future of the economy had increased as the risk of higher unemployment and inflation had risen.  

“I think that people have become less optimistic on the US and less pessimistic on Europe,” said Hoops. 

The German executive said that it was too soon to determine whether the US would suffer a recession, however, he added: “I’m more worried about a lower growth path than the US otherwise would have been on, as opposed to a recession, which is typically going to lead to a rebound.”

The market volatility thrown off by trade turmoil has forced asset managers to reallocate portfolios, with many pulling money from US equities. But Hoops said the uncertainty would sift out the best investors from the worst.

“This is the year of truth for active asset management,” he said. “If you can’t find ways to differentiate in this environment — marked by dispersion, divergence, and structural change — then when can you?”

He added that volatility would also likely drive dealmaking among asset managers. “I feel that the current market behaviour will simply force consolidation, and not by choice . . . which would be interesting.”

OpenAI negotiates with Microsoft to unlock new funding and future IPO

High-stakes talks to revise terms of multibillion-dollar partnership are key to plans to adopt for-profit model

OpenAI and Microsoft are rewriting the terms of their multibillion-dollar partnership in a high-stakes negotiation designed to allow the ChatGPT maker to launch a future IPO, while protecting the software giant’s access to cutting-edge artificial intelligence models.

Microsoft, OpenAI’s biggest backer, is a key holdout to the $260bn start-up’s plans to undergo a corporate restructuring that moves the group further away from its roots as a non-profit with a mission to develop AI to “benefit humanity”.

A critical issue in the deliberations is how much equity in the restructured group Microsoft will receive in exchange for the more than $13bn it has invested in OpenAI to date.  

According to multiple people with knowledge of the negotiations, the pair are also revising the terms of a wider contract, first drafted when Microsoft first invested $1bn into OpenAI in 2019. 

The contract currently runs to 2030 and covers what access Microsoft has to OpenAI’s intellectual property such as models and products, as well as a revenue share from product sales. 

Three people with direct knowledge of the talks said Microsoft is offering to give up some of its equity stake in OpenAI’s new for-profit business in exchange for accessing new technology developed beyond the 2030 cut off. 

That deal is critical to OpenAI’s restructuring efforts and could dictate the future of a company which has been in the vanguard of tech groups building large language models, a transformative technology that is beginning to disrupt global industries. 

OpenAI’s chief executive Sam Altman has said his goal is to go further and build artificial general intelligence, systems that surpass the abilities of humans.

Last week OpenAI ditched controversial plans that would have removed ultimate control of the group by its non-profit board. However, it retained plans for its business arm to become a public benefit corporation (PBC), a body focused on social good in addition to making profits.

That corporate model, adopted by rivals such as Anthropic and Elon Musk’s venture xAI, would still allow OpenAI to offer investors equity in the business. A person close to the company said the change is a key demand of investors and would ensure that an “IPO becomes possible” in the future.

Negotiations between OpenAI and Microsoft are complicated by a cooling between the companies, according to multiple people with direct knowledge of their relationship. 

The groups remain close collaborators. Microsoft has embedded OpenAI’s technology into its software products, while providing it with huge amounts of computing power to train AI models. 

But OpenAI’s ambitions have increased competition with its biggest benefactor. The start-up has targeted enterprise customers with AI products, while seeking partners such as Japan’s SoftBank and Larry Ellison’s Oracle to build its own vast computing infrastructure dubbed “Stargate”.

“The friction comes partly due to style. OpenAI says to Microsoft ‘gives us money and compute and stay out of the way: be happy to be on the ride with us.’ So naturally this leads to tensions,” said one senior employee at Microsoft. “To be honest, that is a bad partner attitude, it shows arrogance.” 

One person close to OpenAI said: “Microsoft still wants [this conversion] to succeed. It’s not like it’s all gone to hell and it’s open warfare. There’s a tough negotiation but we’re confident we’ll get it done.”

OpenAI was founded as a non-profit research lab in 2015 by Altman, Musk and nine others. The group launched a for-profit subsidiary in 2019 into which outside groups could invest in exchange for a share of future profits, up to a certain cap. 

At the time, the group told investors including Microsoft to regard such funding “in the spirit of a donation” and warned them its mission would take precedence over profits. 

Recent investors have not regarded their backing as a donation, however. 

In October last year, OpenAI raised $6.6bn raise from SoftBank, Microsoft and venture capitalists including Thrive Capital and Altimeter Capital. In March, it raised a further $40bn in a round led by SoftBank.

As part of those deals, provisions in the investors’ contracts lay out how much equity they will receive when OpenAI converts to a new structure. 

Those contracts mean that the investors have the option to recoup some or all of the cash they have committed if OpenAI fails to convert into a PBC. OpenAI’s executives are confident that their backers will remain committed even if there is a delay to the restructuring.

The requirement to convert into a more conventional for-profit group is “a high level recognition of what’s required to raise this amount of money,” said one person close to OpenAI, who added that raising “$40bn under a capped profit structure is not achievable.” 

Even if a deal can be reached with Microsoft, OpenAI faces further hurdles. It pledged on Monday to ensure its business arm would still ultimately be controlled by a non-profit board by giving the board a substantial equity stake in the PBC and the power to nominate its directors. 

But that has failed to satisfy critics who claim OpenAI is imperilling its mission by putting profit over purpose. 

Musk, who left OpenAI after falling out with Altman, has vowed to continue his legal action seeking to stop any corporate restructuring.

“The charity is still turning over its assets and technology to private persons for private gain — including Sam Altman — while moving all of the charity’s actual work on AI/AGI into a giant for-profit corporation,” wrote Musk’s attorney Marc Toberoff in a statement. 

Page Hedley, a former OpenAI employee, said that the proposed changes undermined OpenAI’s mission and created “the potential for extraordinary wealth and power from artificial general intelligence [to] be reallocated from the public to OpenAI’s investors.”

OpenAI must also persuade the authorities in California and Delaware — the states in which it is headquartered and incorporated — that its proposal will uphold the group’s mission to benefit the public.

Delaware’s attorney-general Kathy Jennings said on Monday she would review OpenAI’s new plan “for compliance with Delaware law by ensuring that it accords with OpenAI’s charitable purpose and that the non-profit entity retains appropriate control over the for-profit entity.”

Industry insiders said a failure of OpenAI’s new plan to make its business arm a PBC could prove a critical blow. That would hit OpenAI’s ability to raise more cash, achieve a future float, and obtain the financial resources to take on Big Tech rivals such as Google.

That has left OpenAI’s future at the mercy of investors, such as Microsoft, who want to ensure they gain the benefit of its enormous growth, said Dorothy Lund, professor of law at Columbia Law School.

“When you’re a mission driven company which needs money from investors, you are in a dangerous position,” she said. “You have to walk this line: you want your investors to keep giving you huge billion dollar cheques, so you need to keep them happy.”