Connect with us


Inside Amira Yahyaoui’s Claims about Mos, a Student Aid Start-Up



As a Tunisian human rights activist in the 2000s, Amira Yahyaoui staged protests and blogged about government corruption. In interviews, she described being beaten by police. When she was 18, she said, she was kidnapped from the street, dropped off at the Algerian border and placed in exile for several years.

Ms. Yahyaoui’s compelling background helped her stand out among entrepreneurs when she moved in 2018 to San Francisco, where she founded a student aid start-up called Mos. The app hit the top of Apple’s App Store and Ms. Yahyaoui raised $56 million from high-profile investors, including Sequoia Capital, John Doerr and Steph Curry, according to PitchBook, which tracks start-ups. Mos was valued at $400 million.

In podcasts, TV interviews and other media, Ms. Yahyaoui, 39, frequently discussed Mos’s success.

Among other things, she said the start-up had helped 400,000 students get financial aid. But internal company data viewed by The New York Times showed that as of early last year, only about 30,000 customers had paid for Mos’s student aid services. The rest of the 400,000 users included anyone who had signed up for a free account and may have gotten an email about applying for student aid, two people familiar with the situation said.

After Mos expanded into online banking in September 2021, Ms. Yahyaoui told publications such as TechCrunch that the company had more than 100,000 bank accounts. But those accounts had very small amounts of money in them, according to the internal data. Less than 10 percent of Mos’s roughly 153,000 bank users had put their own money into their accounts, the data showed.

Some employees tried to speak up about Ms. Yahyaoui’s claims, said Emi Tabb, who worked at Mos in operations and had roles such as head of financial aid before resigning in late 2022. But Ms. Yahyaoui dismissed and sometimes disparaged employees who tried pushing back against her public comments, five people who witnessed the incidents said.

“She created a culture of fear,” Mx. Tabb said.

Mos is among a class of tech start-ups that rose during the fast money era of the late 2010s and early in the pandemic, when young companies landed millions of dollars in funding with little more than promises. Now as the money has dried up and many tech start-ups grapple with a downturn, investors are pickier, customers are warier of bold claims and employees are more suspicious of founder pronouncements.

Last year, Mos laid off approximately half its staff of around 50 and shut down its banking service. The company reverted to its original business of helping students find financial aid and began emphasizing its use of artificial intelligence.

Ms. Yahyaoui referred questions to a Mos spokeswoman, who declined to comment. When Ms. Yahyaoui was asked last year about Mos’s number of users, she posted on social media that female founders were often presumed guilty while male founders were presumed innocent.

“Maybe today we should start applying presumption of innocence to also female founders,” she wrote.

This account of Mos was based on interviews with eight current and former employees, as well as internal communications, presentations and analytics. The internal documents go up to 2023.

Ms. Yahyaoui grew up in Tunisia and then lived in exile in France. After moving to San Francisco, she raised money for Mos from investors including Expa, the investment firm started by Garrett Camp, a founder of Uber. Mos provided a service to help students find sources of financial aid, charging $149 for each school year.

Deena Shakir, an investor at Lux Capital, which backed Mos in 2020, said she and the firm’s partners “deeply respect” Ms. Yahyaoui.

“We take pride in supporting companies and founders like Amira whose commitment to enabling access for students gives us hope for the future of higher education,” Ms. Shakir said.

Mos had a slow start, three people with knowledge of the company said. Some students who signed up learned about aid they already knew about, like a Cal Grant for California residents, they said.

An investor presentation viewed by The Times showed that Mos had monthly revenue of $340,000 in December 2019. The start-up allowed users to pay $1 upfront and the remaining $148 when they got their financial aid.

Mos ultimately did not collect most of that money. Seventy percent of users defaulted on their payments after the pandemic hit in 2020, Jess Lee, an investor at Sequoia who sits on Mos’s board, later said in an article about the company published on Sequoia’s website.

As of late 2022, roughly 6,500 of Mos’s paying customers, or 22 percent, got refunds for its financial aid service, according to internal data. The company had told customers that if they didn’t get five times the cost of Mos’s services in financial aid, they could get a refund.

Mos said it could help students access $160 billion in scholarships, but that amount included loans, three people familiar with the situation said. The company’s pitch was to help students avoid debt.

Ms. Yahyaoui also said students who used Mos “saved” an average of $16,000. That was the amount that the start-up determined they qualified for and not what the students received in aid, three people with knowledge of the company said.

Mos’s website includes a moving ticker of happy customers (“Jasmine got $12,237 for Cal Poly,” for example). Ms. Yahyaoui asked employees to use stock photos and to make up names, three people with knowledge of the company said.

By 2021, financial technology was hot with investors. Ms. Yahyaoui pushed Mos to become a bank, making its financial aid product free. That September, the start-up announced its move into banking with a promotion that gave people $5 to sign up and another $5 for every referral.

Sign-ups poured in. Mos turned off the $5 promotion on its first day. Two months later, it turned it back on for three days and signed up more than 100,000 accounts, spending around $1 million in the promotion and sending Mos to the top of the App Store.

The sign-ups piqued investor interest, including from the investment firm Tiger Global. Sequoia’s Ms. Lee wanted to see how many of the accounts that signed up during the promotion remained active before investing more, two people familiar with the situation said. Sequoia encouraged Ms. Yahyaoui to hire an outside firm to assess whether the accounts belonged to real people, the people said.

Some employees also had concerns that many accounts did not belong to real people, three people familiar with the situation said. As sign-ups continued, Mos analyzed the accounts for potentially fraudulent behavior in an internal working document. In November, Ms. Yahyaoui restricted Ms. Lee’s access to that document, two of the people said.

Soon after, in February 2022, Tiger Global announced it led a $40 million funding for Mos. Sequoia joined the deal. It is not clear what impact access to the document would have had on Sequoia’s decision to invest more in Mos. Two people familiar with the situation said Ms. Lee retained access to a broader data source regarding the accounts.

In a statement, Ms. Lee said, “The most successful founders are the ones who have grit and are willing to test new hypotheses and adapt. Amira is the embodiment of these qualities.”

Tiger Global declined to comment.

Alongside the funding announcement, Sequoia published an article on its website detailing Ms. Yahyaoui’s dramatic past and entrepreneurial vision. It said fewer than 1 percent of Mos’s bank accounts had been closed, “an unheard-of statistic for a money-based sign-up promotion.”

Few people used the bank accounts, according to internal data viewed by The Times. Of roughly 153,000 open accounts, 95 percent had less than $5 in them and a third had a balance of zero through 2022, the data showed. Just 9.5 percent of account holders deposited money into their accounts during that time.

Mos told its board that 74 percent of bank account holders were students, according to a presentation viewed by The Times. But only around 20 percent were 22 or younger, according to internal data, with about 45 percent over the age of 30. Mos’s revenue from transaction fees, which made up the vast majority of the company’s total income after it became a bank, was less than $70,000 for the first nine months of 2022, two people familiar with the finances said.

Ms. Yahyaoui sometimes berated her top managers and threatened to fire them if their performance didn’t improve, according to five people who witnessed such events.

Using expletives, she wrote in a January 2022 message to employees that the company’s mission was meaningless “because of how bad we are at getting” stuff done.

“I need people I can count on to beat my dreams not to lower them,” she wrote.

Ms. Yahyaoui’s treatment of employees — including workers hired in Tunisia and Algeria — ran counter to her image as an activist, Mx. Tabb said.

At an employee gathering in September 2022, a Mos employee asked Sequoia’s Ms. Lee about her biggest concern for the start-up, three people who attended said. Ms. Lee initially said she was surprised by how good morale was given the circumstances, then added that it wasn’t clear what Mos’s product would be.

The start-up was at more of a “seed stage,” or very early in its development, Ms. Lee said.

Source link

Continue Reading
Click to comment

Leave a Reply

Your email address will not be published. Required fields are marked *


The Paris Olympics’ One Sure Thing: Cyberattacks



In his office on one of the upper floors of the headquarters of the Paris Olympic organizing committee, Franz Regul has no doubt what is coming.

“We will be attacked,” said Mr. Regul, who leads the team responsible for warding off cyberthreats against this year’s Summer Games in Paris.

Companies and governments around the world now all have teams like Mr. Regul’s that operate in spartan rooms equipped with banks of computer servers and screens with indicator lights that warn of incoming hacking attacks. In the Paris operations center, there is even a red light to alert the staff to the most severe danger.

So far, Mr. Regul said, there have been no serious disruptions. But as the months until the Olympics tick down to weeks and then days and hours, he knows the number of hacking attempts and the level of risk will rise exponentially. Unlike companies and governments, though, who plan for the possibility of an attack, Mr. Regul said he knew exactly when to expect the worst.

“Not many organizations can tell you they will be attacked in July and August,” he said.

Worries over security at major events like the Olympics have usually focused on physical threats, like terrorist attacks. But as technology plays a growing role in the Games rollout, Olympic organizers increasingly view cyberattacks as a more constant danger.

The threats are manifold. Experts say hacking groups and countries like Russia, China, North Korea and Iran now have sophisticated operations capable of disabling not just computer and Wi-Fi networks but also digital ticketing systems, credential scanners and even the timing systems for events.

Fears about hacking attacks are not just hypothetical. At the 2018 Pyeongchang Winter Olympics in South Korea, a successful attack nearly derailed the Games before they could begin.

That cyberattack started on a frigid night as fans arrived for the opening ceremony. Signs that something was amiss came all at once. The Wi-Fi network, an essential tool to transmit photographs and news coverage, suddenly went down. Simultaneously, the official Olympics smartphone app — the one that held fans’ tickets and essential transport information — stopped functioning, preventing some fans from entering the stadium. Broadcast drones were grounded and internet-linked televisions meant to show images of the ceremony across venues went blank.

But the ceremony went ahead, and so did the Games. Dozens of cybersecurity officials worked through the night to repel the attack and to fix the glitches, and by the next morning there was little sign that a catastrophe had been averted when the first events got underway.

Since then, the threat to the Olympics has only grown. The cybersecurity team at the last Summer Games, in Tokyo in 2021, reported that it faced 450 million attempted “security events.” Paris expects to face eight to 12 times that number, Mr. Regul said.

Perhaps to demonstrate the scale of the threat, Paris 2024 cybersecurity officials use military terminology freely. They describe “war games” meant to test specialists and systems, and refer to feedback from “veterans of Korea” that has been integrated into their evolving defenses.

Experts say a variety of actors are behind most cyberattacks, including criminals trying to hold data in exchange for a lucrative ransom and protesters who want to highlight a specific cause. But most experts agree that only nation states have the ability to carry out the biggest attacks.

The 2018 attack in Pyeongchang was initially blamed on North Korea, South Korea’s antagonistic neighbor. But experts, including agencies in the U.S. and Britain, later concluded that the true culprit — now widely accepted to be Russia — deliberately used techniques designed to pin the blame on someone else.

This year, Russia is once again the biggest focus.

Russia’s team has been barred from the Olympics following the country’s 2022 invasion of Ukraine, although a small group of individual Russians will be permitted to compete as neutral athletes. France’s relationship with Russia has soured so much that President Emmanuel Macron recently accused Moscow of attempting to undermine the Olympics through a disinformation campaign.

The International Olympic Committee has also pointed the finger at attempts by Russian groups to damage the Games. In November, the I.O.C. issued an unusual statement saying it had been targeted by defamatory “fake news posts” after a documentary featuring an A.I.-generated voice-over purporting to be the actor Tom Cruise appeared on YouTube.

Later, a separate post on Telegram — the encrypted messaging and content platform — mimicked a fake news item broadcast by the French network Canal Plus and aired false information that the I.O.C. was planning to bar Israeli and Palestinian teams from the Paris Olympics.

Earlier this year, Russian pranksters — impersonating a senior African official — managed to get Thomas Bach, the I.O.C. president, on the phone. The call was recorded and released earlier this month. Russia seized on Mr. Bach’s remarks to accuse Olympic officials of engaging in a “conspiracy” to keep its team out of the Games.

In 2019, according to Microsoft, Russian state hackers attacked the computer networks of at least 16 national and international sports and antidoping organizations, including the World Anti-Doping Agency, which at the time was poised to announce punishments against Russia related to its state-backed doping program.

Three years earlier, Russia had targeted antidoping officials at the Rio de Janeiro Summer Olympics. According to indictments of several Russian military intelligence officers filed by the United States Department of Justice, operatives in that incident spoofed hotel Wi-Fi networks used by antidoping officials in Brazil to successfully penetrate their organization’s email networks and databases.

Ciaran Martin, who served as the first chief executive of Britain’s national cybersecurity center, said Russia’s past behavior made it “the most obvious disruptive threat” at the Paris Games. He said areas that might be targeted included event scheduling, public broadcasts and ticketing systems.

“Imagine if all athletes are there on time, but the system scanning iPhones at the gate has gone down,” said Mr. Martin, who is now a professor at the Blavatnik School of Government at the University of Oxford.

“Do you go through with a half-empty stadium, or do we delay?” he added. “Even being put in that position where you either have to delay it or have world-class athletes in the biggest event of their lives performing in front of a half-empty stadium — that’s absolutely a failure.”

Mr. Regul, the Paris cybersecurity head, declined to speculate about any specific nation that might target this summer’s Games. But he said organizers were preparing to counter methods specific to countries that represent a “strong cyberthreat.”

This year, Paris organizers have been conducting what they called “war games” in conjunction with the I.O.C. and partners like Atos, the Games’ official technology partner, to prepare for attacks. In those exercises, so-called ethical hackers are hired to attack systems in place for the Games, and “bug bounties” are offered to those who discover vulnerabilities.

Hackers have previously targeted sports organizations with malicious emails, fictional personas, stolen passwords and malware. Since last year, new hires at the Paris organizing committee have undergone training to spot phishing scams.

“Not everyone is good,” Mr. Regul said.

In at least one case, a Games staff member paid an invoice to an account after receiving an email impersonating another committee official. Cybersecurity staff members also discovered an email account that had attempted to impersonate the one assigned to the Paris 2024 chief, Tony Estanguet.

Millions more attempts are coming. Cyberattacks have typically been “weapons of mass irritation rather than weapons of mass destruction,” said Mr. Martin, the former British cybersecurity official.

“At their worst,” he said, “they’ve been weapons of mass disruption.”

Source link

Continue Reading


'Bad at almost everything': AI wearable panned by reviewers



A new AI-fuelled gadget has fallen foul of the tech world’s expectations.

Source link

Continue Reading


Microsoft Makes High-Stakes Play in Tech Cold War With Emirati A.I. Deal



Microsoft on Tuesday plans to announce a $1.5 billion investment in G42, an artificial intelligence giant in the United Arab Emirates, in a deal largely orchestrated by the Biden administration to box out China as Washington and Beijing battle over who will exercise technological influence in the Gulf region and beyond.

Under the partnership, Microsoft will give G42 permission to sell Microsoft services that use powerful A.I. chips, which are used to train and fine-tune generative A.I. models. In return, G42, which has been under scrutiny by Washington for its ties to China, will use Microsoft’s cloud services and accede to a security arrangement negotiated in detailed conversations with the U.S. government. It places a series of protections on the A.I. products shared with G42 and includes an agreement to strip Chinese gear out of G42’s operations, among other steps.

“When it comes to emerging technology, you cannot be both in China’s camp and our camp,” said Gina Raimondo, the Commerce Secretary, who traveled twice to the U.A.E. to talk about security arrangements for this and other partnerships.

The accord is highly unusual, Brad Smith, Microsoft’s president, said in an interview, reflecting the U.S. government’s extraordinary concern about protecting the intellectual property behind A.I. programs.

“The U.S. is quite naturally concerned that the most important technology is guarded by a trusted U.S. company,” said Mr. Smith, who will take a seat on G42’s board.

The investment could help the United States push back against China’s rising influence in the Gulf region. If the moves succeed, G42 would be brought into the U.S. fold and pare back its ties with China. The deal could also become a model for how U.S. firms leverage their technological leadership in A.I. to lure countries away from Chinese tech, while reaping huge financial awards.

But the matter is sensitive, as U.S. officials have raised questions about G42. This year, a congressional committee wrote a letter urging the Commerce Department to look into whether G42 should be put under trade restrictions for its ties to China, which include partnerships with Chinese firms and employees who came from government-connected companies.

In an interview, Ms. Raimondo, who has been at the center of an effort to prevent China from obtaining the most advanced semiconductors and the equipment to make them, said the agreement “does not authorize the transfer of artificial intelligence, or A.I. models, or GPUs” — the processors needed to develop A.I. applications — and “assures those technologies can be safely developed, protected and deployed.”

While the U.A.E. and United States did not sign a separate accord, Ms. Raimondo said, “We have been extensively briefed and we are comfortable that this agreement is consistent with our values.”

In a statement, Peng Xiao, the group chief executive of G42, said that “through Microsoft’s strategic investment, we are advancing our mission to deliver cutting-edge A.I. technologies at scale.”

The United States and China have been racing to exert technological influence in the Gulf, where hundreds of billions of dollars are up for grabs and major investors, including Saudi Arabia, are expected to spend billions on the technology. In the rush to diversify away from oil, many leaders in the region have set their sights on A.I. — and have been happy to play the United States and China off each other.

Although the U.A.E. is an important U.S. diplomatic and intelligence partner, and one of the largest buyers of American weapons, it has increasingly expanded its military and economic ties with China. A portion of its domestic surveillance system is built on Chinese technology and its telecommunications work on hardware from Huawei, a Chinese supplier. That has fed the worries of U.S. officials, who often visit the Persian Gulf nation to discuss security issues.

But U.S. officials are also concerned that the spread of powerful A.I. technology critical to national security could eventually be used by China or by Chinese government-linked engineers, if not sufficiently guarded. Last month, a U.S. cybersecurity review board sharply criticized Microsoft over a hack in which Chinese attackers gained access to data from top officials. Any major leak — for instance, by G42 selling Microsoft A.I. solutions to companies set up in the region by China — would go against Biden administration policies that have sought to limit China’s access to the cutting-edge technology.

“This is among the most advanced technology that the U.S. possesses,” said Gregory Allen, a researcher at the Center for Strategic and International Studies and a former U.S. defense official who worked on A.I. “There should be very strategic rationale for offshoring it anywhere.”

For Microsoft, a deal with G42 offers potential access to huge Emirati wealth. The company, whose chairman is Sheikh Tahnoon bin Zayed, the Emirates’ national security adviser and the younger brother of the country’s ruler, is a core part of the U.A.E.’s efforts to become a major A.I. player.

Despite a name whimsically drawn from “The Hitchhiker’s Guide to the Galaxy,” in which the answer to the “ultimate question of life” is 42, G42 is deeply embedded in the Emirati security state. It specializes in A.I. and recently worked to build an Arabic chatbot, called Jais.

G42 is also focused on biotechnology and surveillance. Several of its executives, including Mr. Xiao, were associated with a company called DarkMatter, an Emirati cyber-intelligence and hacking firm that employs former spies.

In its letter this year, the bipartisan House Select Committee on the Chinese Communist Party said Mr. Xiao was connected to an expansive network of companies that “materially support” the Chinese military’s technological advancement.

The origins of Tuesday’s accord go back to White House meetings last year, when top national security aides raised the question with tech executives of how to encourage business arrangements that would deepen U.S. ties to firms around the world, especially those China is also interested in.

Under the agreement, G42 will cease using Huawei telecom equipment, which the United States fears could provide a backdoor for the Chinese intelligence agencies. The accord further commits G42 to seeking permission before it shares its technologies with other governments or militaries and prohibits it from using the technology for surveillance. Microsoft will also have the power to audit G42’s use of its technology.

G42 would get use of A.I. computing power in Microsoft’s data center in the U.A.E., sensitive technology that cannot be sold in the country without an export license. Access to the computing power would likely give G42 a competitive edge in the region. A second phase of the deal, which could prove even more controversial and has not yet been negotiated, could transfer some of Microsoft’s A.I. technology to G42.

American intelligence officials have raised concerns about G42’s relationship to China in a series of classified assessments, The New York Times previously reported. Biden administration officials have also pushed their Emirati counterparts to cut the company’s ties to China. Some officials believe the U.S. pressure campaign has yielded some results, but remain concerned about less overt ties between G42 and China.

One G42 executive previously worked at the Chinese A.I. surveillance company Yitu, which has extensive ties to China’s security services and runs facial-recognition powered monitoring across the country. The company has also had ties to a Chinese genetics giant, BGI, whose subsidiaries were placed on a blacklist by the Biden administration last year. Mr. Xiao also led a firm that was involved in 2019 in starting and operating a social media app, ToTok, that U.S. intelligence agencies said was an Emirati spy tool used to harvest user data.

In recent months, G42 has agreed to walk back some of its China ties, including divesting a stake it took in TikTok owner ByteDance and pulling out Huawei technology from its operations, according to U.S. officials.

Edward Wong contributed reporting.

Source link

Continue Reading


Copyright © 2024 World Daily Info. Powered by Columba Ventures Co. Ltd.