Connect with us

Published

on

The logo of OpenAI is shown on a mobile phone in front of a computer screen displaying the photographs of Sam Altman, left, and Elon Musk, March 14, 2024.

Muhammed Selim Korkutata | Anadolu | Getty Images

OpenAI on Friday clapped back against Elon Musk, one of its co-founders, after the billionaire’s request last month for a federal court to stop the ChatGPT-maker from converting to a fully for-profit business.

In a blog post titled “Elon Musk wanted an OpenAI for-profit,” the startup alleged that in 2017, Musk “not only wanted, but actually created, a for-profit” to serve as the company’s proposed new structure.

“When he didn’t get majority equity and full control, he walked away and told us we would fail,” OpenAI wrote in the blog post. “Now that OpenAI is the leading AI research lab and Elon runs a competing AI company, he’s asking the court to stop us from effectively pursuing our mission.”

Musk and xAI did not immediately respond to requests for comment.

Since Musk announced the debut of xAI, his OpenAI competitor, in July 2023, the startup has released its Grok chatbot and is raising up to $6 billion at a $50 billion valuation, in part to buy 100,000 Nvidia chips, CNBC reported last month.

Musk was questioning OpenAI’s nonprofit model from day one, a member of OpenAI’s legal team told CNBC.

OpenAI’s “structure doesn’t seem optimal,” Musk wrote in a November 2015 email to OpenAI CEO Sam Altman, according to screenshots shared in the blog post. He added that receiving a “salary from the nonprofit muddies the alignment of incentives,” and that it’s “probably better to have a standard C corp with a parallel nonprofit.”

In a text conversation with former board member Shivon Zilis, OpenAI co-founder Greg Brockman wrote that a conversation he had with Musk “turned into talking about structure” and that Musk “said non-profit was def the right one early on, may not be the right one now,” according to blog screenshots.

Musk forwarded an article about China’s strategy for AI research facilities to Brockman and fellow OpenAI co-founder Ilya Sutskever. Musk wrote that China “will do whatever it takes to obtain what we develop. Maybe another reason to change course,” per the blog post.

Brockman agreed, and he wrote that starting in 2018, OpenAI’s path would need to be a “Al research + hardware for-profit,” according to the blog post. Musk wrote back, “Let’s talk Sat or Sun. I have a tentative game plan that l’d like to run by you.”

Altman, Brockman, Musk and others negotiated terms for the planned OpenAI for-profit in the fall of 2017, but the talks fell apart due to disagreements about equity, control and who would be CEO, according to the blog. Musk initially proposed that he should “unequivocally have initial control of the company” but said “this will change quickly” when the board has 12 to 16 members, per screenshots.

Musk created a public benefit corporation called “Open Artificial Intelligence Technologies, Inc” in September 2017, according to screenshots included in OpenAI’s blog post. A few days later, OpenAI rejected Musk’s proposed terms for the for-profit and offered to keep the conversation going, but Musk responded that his offer was “no longer on the table” and that “discussions are over,” per screenshots.

In January 2018, Musk proposed that OpenAI spin into Tesla, his electric vehicle company, according to the blog.

“The only paths I can think of are a major expansion of OpenAl and a major expansion of Tesla Al. Perhaps both simultaneously. The former would require a major increase in funds donated and highly credible people joining our board. The current board situation is very weak,” Musk wrote, according to the blog. He added that “OpenAI is on a path of certain failure relative to Google.”

Brockman responded with a lengthy plan, including the idea that the company should “try our best to remain a non-profit,” according to screenshots. In February 2018, Musk resigned as co-chair of OpenAI.

OpenAI’s complex history

OpenAI originally debuted in 2015 as a nonprofit and then in 2019 converted into a “capped-profit” model, in which the OpenAI nonprofit was the governing entity for its for-profit subsidiary. Altman claimed onstage last week at the DealBook Summit that the company decided to go to a capped-profit structure in part because Musk stopped funding them.

Thanks largely to the viral spread of ChatGPT, which debuted in November 2022, OpenAI has become one of the hottest, and at times one of the most controversial, startups on the planet. The company’s valuation has climbed to $157 billion since it launched ChatGPT. OpenAI has raised about $13 billion from Microsoft, and it closed its latest $6.6 billion round in October, led by Thrive Capital and including participation from chipmaker Nvidia, SoftBank and others.

The company also received a $4 billion revolving line of credit, bringing its total liquidity to more than $10 billion. OpenAI expects about $5 billion in losses on $3.7 billion in revenue this year, CNBC confirmed in September with a person familiar with the situation.

OpenAI is now in the midst of a potentially two-year process of converting into a fully for-profit public benefit corporation, which could make it more attractive to investors. The restructuring plan would also allow OpenAI to retain its non-profit status as a separate entity, CNBC previously reported.

OpenAI has faced increasing competition from startups such as Musk’s xAI and Anthropic, as well as tech giants such as Google, Amazon and Meta. The generative AI market is predicted to top $1 trillion in revenue within a decade, and business spending on generative AI surged 500% this year, according to recent data from Menlo Ventures.

A thorny legal battle

Attorneys representing Musk, his AI startup xAI and Zilis filed for a preliminary injunction against OpenAI on Nov. 29.

In their motion for preliminary injunction, attorneys for Musk argued that OpenAI should be prohibited from “benefitting from wrongfully obtained competitively sensitive information or coordination via the Microsoft-OpenAI board interlocks.”

The latest court filings represent an escalation in the legal feud between Musk, OpenAI and Altman, as well as other long-involved parties and backers including tech investor Reid Hoffman and Microsoft.

Musk in March 2024 sued OpenAI — and co-founders Altman and Brockman — in a San Francisco state court, alleging breach of contract and fiduciary duty. In the suit, Musk claimed that the early OpenAI team had set out to develop artificial general intelligence “for the benefit of humanity,” but that the project had been transformed into a for-profit entity that’s largely controlled by principal shareholder Microsoft.

In June, Musk withdrew that complaint and later refiled in federal court. Attorneys for Musk in the federal suit, led by Marc Toberoff in Los Angeles, argued in their complaint that OpenAI had violated federal racketeering, or RICO, laws.

In November, they expanded their complaint to include allegations that Microsoft and OpenAI had violated antitrust laws when the ChatGPT maker allegedly asked investors to agree to not invest in rival companies, including Musk’s xAI.

“Microsoft and OpenAI now seek to cement this dominance by cutting off competitors’ access to investment capital (a group boycott), while continuing to benefit from years’ worth of shared competitively sensitive information during generative AI’s formative years,” the lawyers wrote in the November filing. They added that the terms OpenAI asked investors to agree to amounted to a “group boycott” that “blocks xAI’s access to essential investment capital.”

Altman denied that OpenAI investors can’t invest in competitors during an onstage interview last week at The New York Times’ DealBook Summit. Altman said that investors are welcome to do so but that the company will stop their “information rights,” such as sharing its research road map and other materials.

Microsoft has invested nearly $14 billion in OpenAI but revealed in October that it would record a $1.5 billion loss in the current period largely due to an expected loss from the AI startup. Microsoft gave up its observer seat on OpenAI’s board in July, although CNBC reported that the Federal Trade Commission would continue to monitor the influence of the two companies over the AI industry.

— CNBC’s Lora Kolodny contributed reporting.

WATCH: OpenAI releases AI video generation tool Sora

OpenAI releases AI video generation tool Sora

Continue Reading

Technology

Just Eat shares soar 54% after Prosus offers to buy food delivery firm for $4.3 billion

Published

on

By

Just Eat shares soar 54% after Prosus offers to buy food delivery firm for .3 billion

Just Eat Takeaway said it was delisting its shares from the London Stock Exchange due to the “low liquidity and trading volumes” of its shares on the exchange.

Mike Kemp | In Pictures | Getty Images

European food delivery giant Just Eat Takeaway.com is poised to be acquired by Dutch technology investor Prosus in an all-cash deal worth roughly 4.1 billion euros ($4.3 billion).

The offer values Just Eat’s shares at 20.3 euros each, representing a premium of 63% when compared to the firm’s closing price on Friday.

Prosus, which is majority owned by South Africa’s Naspers, already holds a 28% stake in leading food delivery company Delivery Hero.

Shares of Just Eat soared as much as 54% on Monday morning, notching a new 52-week high. The stock price was last seen trading 52.8% higher on the news.

Prosus shares fell 8.3%, tumbling to the bottom of the pan-European Stoxx 600 index, while Delivery Hero rose 5.4%.

“We are very excited for Just Eat Takeaway.com to join the Prosus group and the opportunity to create a European tech champion,” Fabricio Bloisi, CEO of Prosus and Naspers group, said in a statement.

“We believe that combining Prosus’ strong technical and investment capabilities with Just Eat Takeaway.com’s leading brand position in key European markets will create significant value for our customers, drivers, partners, and shareholders,” Bloisi said.

Stock Chart IconStock chart icon

hide content

Just Eat

The offer comes after a rocky few years for Just Eat. Like many other food delivery companies, the company’s stock price collapsed in the wake of the coronavirus pandemic, which had initially boosted the firms as consumers turned to these platforms during lockdowns.

A stark shift in consumer habits since, however, led to a sharp deceleration in growth rates.

The Dutch multinational delisted from the London Stock Exchange late last year, citing an effort to “reduce the administrative burden, complexity and costs associated with the disclosure and regulatory requirements of maintaining the LSE listing.” The move made Amsterdam the firm’s sole trading venue.

In November, Just Eat Takeaway.com said it would sell its GrubHub arm to New York-based online takeout startup Wonder for $650 million — a huge discount compared to the $7.3 billion the firm paid for the U.S. food delivery app.

“Prosus fully supports our strategic plans and its extensive resources will help to further accelerate our investments and growth across food, groceries, fintech and other adjacencies. We are looking forward to an exciting future together,” Jitse Groen, CEO and founder of Just Eat Takeaway.com, said in a statement on Monday.

— CNBC’s Ryan Browne contributed to this report.

Continue Reading

Technology

Chinese medical devices are in health systems across U.S., and the government and hospitals are worried

Published

on

By

Chinese medical devices are in health systems across U.S., and the government and hospitals are worried

A popular medical monitor is the latest device produced in China to receive scrutiny for its potential cyber risks.  However, it is not the only health device we should be concerned about. Experts say the proliferation of Chinese health-care devices in the U.S. medical system is a cause for concern across the entire ecosystem. 

The Contec CMS8000 is a popular medical monitor that tracks a patient’s vital signs.  The device tracks electrocardiograms, heart rate, blood oxygen saturation, non-invasive blood pressure, temperature, and respiration rate.  In recent months, the FDA and the Cybersecurity and Infrastructure Security Agency (CISA) both warned about a “backdoor” in the device, an “easy-to-exploit vulnerability that could allow a bad actor to alter its configuration.”  

CISA’s research team described “anomalous network traffic” and the backdoor “allowing the device to download and execute unverified remote files” to an IP address not associated with a medical device manufacturer or medical facility but a third-party university — “highly unusual characteristics” that go against generally accepted practices, “especially for medical devices.”

“When the function is executed, files on the device are forcibly overwritten, preventing the end customer—such as a hospital—from maintaining awareness of what software is running on the device,” CISA wrote.

The warnings says such configuration alteration could lead to, for instance, the monitor saying that a patient’s kidneys are malfunctioning or breathing failing, and that could cause medical staff to administer unneeded remedies that could be harmful. 

The Contec’s vulnerability doesn’t surprise medical and IT experts who have warned for years that medical device security is too lax. 

Hospitals are worried about cyber risks

“This is a huge gap that is about to explode,” said Christopher Kaufman, a business professor at Westcliff University in Irvine, California, who specializes in IT and disruptive technologies, specifically referring to the security gap in many medical devices.

The American Hospital Association, which represents over 5,000 hospitals and clinics in the U.S., agrees. It views the proliferation of Chinese medical devices as a serious threat to the system. 

As for the Contec monitors specifically, the AHA says the problem urgently needs to be addressed. 

“We have to put this at the top of the list for the potential for patient harm; we have to patch before they hack,” said John Riggi, national advisor for cybersecurity and risk for the American Hospital Association.  Riggi also served in FBI counterterrorism roles before joining the AHA. 

CISA reports that no software patch is available to help mitigate this risk, but in its advisory said the government is currently working with Contec. 

Contec, headquartered in Qinhuangdao, China,  did not return a request for comment. 

One of the problems is that it is unknown how many monitors there are in the U.S. 

“We don’t know because of the sheer volume of equipment in hospitals. We speculate there are, conservatively, thousands of these monitors; this is a very critical vulnerability,” Riggi said, adding that Chinese access to the devices can pose strategic, technical, and supply chain risks. 

In the short-term, the FDA advised medical systems and patients to make sure the devices are only running locally or to disable any remote monitoring; or if remote monitoring is the only option, to stop using the device if an alternative is available. The FDA said that to date it is not aware of any cybersecurity incidents, injuries, or deaths related to the vulnerability.

The American Hospital Association has also told its members that until a patch is available, hospitals should make sure the monitor no longer has access to the internet, and is segmented from the rest of the network.

Riggi said the while the Contec monitors are a prime example of what we don’t often consider among health care risk, it extends to a range of medical equipment produced overseas. Cash-strapped U.S. hospitals, he explained, often buy medical devices from China, a country with a history of installing destructive malware inside critical infrastructure in the U.S.  Low-cost equipment buys the Chinese potential access to a trove of American medical information that can be repurposed and aggregated for all sorts of purposes. Riggs says data is often transmitted to China with the stated purpose of monitoring a device’s performance, but little else is known about what happens to the data beyond that. 

Riggi says individuals aren’t at acute medical risk as much as the information being collected and aggregated for repurposing and putting the larger medical system at risk. Still, he points out that, at least theoretically, is can’t be ruled out that prominent Americans with medical devices could be targeted for disruption. 

“When we talk to hospitals,  CEOS are surprised, they had no idea about the dangers of these devices, so we are helping them understand.  The question for government is how to incentivize domestic production, away from overseas,”  Riggi said. 

Chinese data collection on Americans

The Contec warning is similar at a general level to TikTok, DeepSeek, TP-Link routers, and other devices and technology from China that the U.S. government says are collecting data on Americans. “And that is all I need to hear in deciding whether to buy medical devices from China,” Riggi said. 

Aras Nazarovas, an information security researcher at Cybernews, agrees that the CISA threat raises serious issues that need to be addressed. 

“We have a lot to fear,” Nazarovas said. Medical devices, like the Contec CMS8000, often have access to highly sensitive patient data and are directly connected to life-saving functions.  Nazarovas says that when the devices are poorly defended, they become easy prey for hackers who can manipulate the displayed data, alter vital settings, or disable the device completely.  

“In some cases, these devices are so poorly protected that attackers can gain remote access and change how the device operates without the hospital or patients ever knowing,” Nazarovas said. 

The consequences of the Contec vulnerability and vulnerabilities in an array of Chinese-made medical devices could easily be life-threatening.  

“Imagine a patient monitor that stops alerting doctors to a drop in a patient’s heart rate or sends incorrect readings, leading to a delayed or wrong diagnosis,” Nazarovas said. In the case of the Contec CMS8000, and Epsimed MN-120 (a different brand name for the same tech), warning from the government, these devices were configured to allow remote code execution by the remote server.  

“This functionality can be used as an entry point into the hospital’s network,” Nazarovas said, leading to patient danger.  

More hospitals and clinics are paying attention. Bartlett Regional Hospital in Juneau, Alaska, does not use the Contec monitors but is always looking for risks. “Regular monitoring is critical as the risk of cybersecurity attacks on hospitals continues to increase,” says Erin Hardin, a spokeswoman for Bartlett.  

However, regular monitoring may not be enough as long as devices are made with poor security. 

Potentially making matters worse, Kaufman says, is that the Department of Government Efficiency is hollowing out departments in charge of safeguarding such devices. According to the Associated Press, many of the recent layoffs at the FDA are employees who review the safety of medical devices. 

Kaufman laments the likely lack of government supervision on what is already, he says, a loosely regulated industry. A U.S. Government Accountability Office report as of January 2022, indicated that 53% of connected medical devices and other Internet of Things devices in hospitals had known critical vulnerabilities. He says the problem has only gotten worse since then. “I’m not sure what is going to be left running these agencies,” Kaufman said.

“Medical device issues are widespread and have been known for some time now,” said Silas Cutler, principal security researcher at medical data company Censys. “The reality is that the consequences can be dire – and even deadly. While high-profile individuals are at heightened risk, the most impacted are going to be the hospital systems themselves, with cascading effects on everyday patients.”  

Continue Reading

Technology

Substack boosts video capabilities amid potential TikTok ban

Published

on

By

Substack boosts video capabilities amid potential TikTok ban

Rafael Henrique | SOPA Images | AP

After posting almost 200 videos, amassing hundreds of thousands of followers and racking up millions of views, Carla Lalli Music is quitting YouTube. Substack is her new focus. 

Music is a cookbook author and food content creator, and she is shifting her focus to Substack, a subscription platform that lets creators charge users subscriptions for access to their content. Music told CNBC she came to that decision after earning more in one year of using Substack, nearly $200,000 in revenue, than she did by posting videos on YouTube since 2021. 

Music is the exact kind of content creator that Substack is trying to lure to its platform as TikTok’s future in the U.S. remains in limbo. 

San Francisco-based Substack launched in 2017 as a tool for newsletter writers to charge readers a monthly fee to read their content. The platform allows creators to connect to their followers directly without having to navigate algorithmic models that control when their content is shown, as is the case on TikTok, Google’s YouTube and other social platforms. Substack has raised about $100 million, most recently at a post-money valuation of more than $650 million, the company told CNBC.

This year, Substack has broadened its focus beyond newsletters, and on Thursday, it announced that creators can now post video content directly through the Substack app and monetize these videos.

“There’s going to be a world of people who are much more focused on videos,” Substack Co-founder Hamish McKenzie told CNBC. “That is a huge world that Substack is only starting to penetrate.”

Substack began this push after the social media landscape was thrown into flux as a result of the effective ban of TikTok in January that caused the popular Chinese-owned service to go offline for a few hours. TikTok was also removed from Apple and Google’s app stores for nearly a month. 

The disruption to TikTok in January happened as a result of a law signed by former President Joe Biden to force a sale of the Chinese-owned app or have it effectively banned in the U.S. On his first day in office, President Donald Trump signed an executive order extending TikTok’s ability to operate in the U.S., but that order expires on April 5. 

Days after TikTok went offline, Substack launched a $20 million fund to court creators to its platform.

“If TikTok gets banned for political reasons, there’s nothing to do with the work you’ve done, but it really affects your life,” McKenzie said. “The only and surefire guard against that is if you don’t place your audience in the hands of some other volatile system who doesn’t care about what happens to your livelihood.”

Moving beyond newsletters

McKenzie says that they are going after creators on competing social media platforms to start sharing their video content on Substack.

“Video-first creators, people who are mobile oriented, there’s a whole lot of new possibility waiting to be unlocked once they meet this model in the right place,” McKenzie said. 

Already, Substack has more than 4 million paid subscriptions with over 50,000 creators who make money on the platform, the company said. Substack says that 82% of its top 250 revenue-generating creators have already integrated audio or video into their content, reflecting a growing emphasis on multimedia content.

Prior to the video announcements, Substack allowed creators to post videos on the app to Notes, which is the platform’s front-facing feed format. But the feature did not allow creators to publish video content behind Substack’s paywalls. 

The update enables creators to put video content behind a paywall and it provides data on estimated revenue impact. It also allows them to track viewership and new subscribers.

Carla Lalli Music is a cookbook writer and food creator.

Carla Lalli Music

Our base case for TikTok is that it gets banned in the U.S.: Lead Edge Capital's Mitchell Green

Continue Reading

Trending