Connect with us

America

Silicon Valley startups are turning to Chinese open-source AI models

Published

on

Misha Laskin, a theoretical physicist and machine learning engineer who contributed to the development of some of Google’s most powerful AI models, encountered a concerning picture when examining the American AI landscape earlier this year.

Laskin observed a growing interest among US AI companies in free, customizable, and increasingly powerful open-source AI models.

The vast majority of these models are produced in China and are rapidly gaining ground against their US competitors.

Assessing the current situation, Laskin stated, “These models are not far behind the frontier (the cutting edge of technology). In fact, they are surprisingly close to the frontier. What is coming now is noticeably close to the frontier.”

Following this development, Laskin founded a startup called Reflection AI to offer an open-source American alternative to the Chinese models gaining traction in Silicon Valley.

The founder of the company, which recently reached an $8 billion valuation, said, “You are starting to see signs that open model companies in China are actually pushing the frontier of intelligence and the limits of intelligence technology in general.”

Over the past year, a significant portion of America’s most popular AI startups have turned to Chinese open AI models, which compete with and sometimes replace expensive US systems as the foundation for American AI products.

More than 15 AI startup founders, engineers, and industry experts who spoke to NBC News stated that American companies’ models still hold the lead in terms of capability.

However, experts emphasized that many Chinese systems are cheaper to access, more customizable, and have become sufficiently competent for many use cases over the past year.

Cost and speed advantages are changing preferences

Investors have poured tens of billions of dollars into OpenAI and Anthropic with the expectation that leading American AI companies will dominate the global market.

But the increasing use of free Chinese models by American companies raises questions about how exceptional these models are and whether America’s insistence on a “closed model” approach is flawed.

Michael Fine, head of machine learning at the search company Exa, which is valued at $700 million and backed by established Silicon Valley investors Lightspeed Venture Partners and Nvidia, said that running Chinese models on their own hardware is, in many cases, much faster and cheaper than using large models like OpenAI’s GPT-5 or Google’s Gemini.

Fine described the process:

“We often launch a feature with a closed model, but then we realize it’s too expensive or too slow, and we ask, ‘What tricks do we have up our sleeve to make this faster and cheaper?'”

Fine stated that the solution is often to replace the closed model with an equivalent open model and then run it on their own infrastructure.

Chinese-origin systems like DeepSeek’s R1 and Alibaba’s Qwen models can be used for free because they are “open-source” or “open-weight,” meaning anyone can download, copy, modify, and run them.

These systems differ from “closed” systems accessed through data centers controlled by major tech giants, such as Anthropic’s Claude or OpenAI’s GPT models.

The technology gap is closing fast

For years, the closed-source models from OpenAI and Anthropic performed far better than both American and Chinese open alternatives.

Even open-source initiatives like BloombergGPT, trained by institutions with resources like Bloomberg on their own financial data, lagged behind OpenAI’s closed models in financial knowledge.

However, over the past year, Chinese companies like DeepSeek and Alibaba have made significant technological strides. According to metrics tracked by Artificial Analysis, an independent AI benchmarking company, their open-source products now approach or match the performance of leading closed American models in many areas.

“The gap is really narrowing,” said Lin Qiao, co-creator of PyTorch, the dominant framework for training AI models, and CEO of Fireworks AI, regarding the capability difference between American closed-source and Chinese open-source models.

As a result of this performance increase, platforms like OpenRouter, which allow users to choose between different models, are seeing a shift toward Chinese open-source models.

Jerry Liu, founder of the productivity app Dayflow, estimates that about 40% of his users now prefer to use open-source models.

Dayflow offers an application built on basic tasks like scanning screenshots and summarizing user activity.

Users can choose between Google’s Gemini model and smaller open-source options like Alibaba’s Qwen.

Liu noted that for tasks like describing a user’s screen, the Qwen model is extremely consistent, stating, “Qwen is as good as GPT-5 for my use case.”

Unlike GPT-5 or Gemini, a smaller version of Qwen can be run at a relatively low cost or for free.

Liu mentioned that paying for closed model usage could cost Dayflow up to $1000 per person, making cheaper open-source models critical for the application’s sustainability.

Privacy sensitivity encourages local processing

The open-source models used by Dayflow perform all processing on each user’s own computer. Liu stated that this is attractive to users who do not want to send their data to the cloud for privacy reasons.

Emphasizing his preference for using open-source models on his own device, Liu said, “Would I use a product where my entire screen is beamed to some random guy’s cloud? Never.”

In addition to increased performance, stronger privacy, and lower costs, open-source models are also gaining ground due to ecosystem advantages.

The rising adoption rate among developers and the open-source systems they create encourage more software engineers to use these models.

Antonio Vespoli, co-founder of the browser assistant startup Circlemind AI, said that Chinese models now dominate online developer resources.

There is a practical reason for this: Chinese models like Qwen, which Airbnb CEO Brian Chesky stated they rely on “heavily,” have abundant training guides and community support.

Charles Zedlewski, chief product officer at the AI infrastructure company Together AI, noted that developers now find it simpler and more efficient to start with open models and adapt them with their own data.

Zedlewski stated that companies understand their needs more clearly as they launch their first AI applications.

Of the top 20 models among users of Kilo Code, a popular application that helps software engineers write code, seven are of Chinese origin, and six of them are open-source.

Beijing’s strategic support and production speed

While most of America’s AI developments occur in the private sector and with a closed-model approach led by industry giants like OpenAI and Anthropic, the Chinese government plays a more active role in charting the country’s AI vision.

In a speech on November 1, Chinese President Xi Jinping called for “more cooperation in open-source technologies.”

In March, China’s top economic planning authority announced its intention to support an ecosystem of open-source models.

While Chinese labs generally release their models openly, American companies like OpenAI achieved early success with closed models and have remained committed to that approach.

Furthermore, many Chinese companies are releasing their products at a faster pace than their American competitors.

Alibaba has released a new model roughly every 20 days this year, while the average time between Anthropic’s releases has been 47 days.

Nathan Lambert, a senior research scientist at the Allen Institute for AI and an expert on the open model ecosystem, told NBC News that the recent progress of Chinese models is no coincidence.

“The Chinese are real innovators in AI,” Lambert said.

Lambert, who writes extensively about China’s AI developments on the Substack platform and is considered an expert on China’s open-source ecosystem, added that the balance of power has shifted rapidly in the last 12 months.

Some in Silicon Valley note that American models still hold a significant advantage at the cutting edge of AI capabilities and that closed American models offer a user-friendliness that cumbersome open models cannot match.

Tim Tully, a partner at Menlo Ventures, argued that closed models are still much more capable and generally more useful:

“The tools are better, the productivity is better, the agent frameworks being built and used by everyone are better with Anthropic and OpenAI. They just work better. So the ecosystem is strong in the closed-source environment.”

However, many companies may avoid using Chinese models due to the real or perceived risks of using a product built on a Chinese-origin foundation.

“There is a perceived risk that buyers, whether from the private or public sector, are hesitant to purchase a product based on a Chinese-origin open-weight model,” said Tully, an investor in Anthropic, one of the world’s leading closed-model companies.

The US open-source ecosystem is waking up

American AI companies and the federal government have taken notice of the recent rise of Chinese models. Experts have described America’s lack of powerful open-source models as an “existential” threat to democracy.

Although Meta’s high-profile Llama series has historically led American open-source efforts, CEO Mark Zuckerberg has signaled that Meta does not intend to make all of its “superintelligence” AI models open-source.

The stagnation in the performance of Llama models in recent years is also seen as one of the reasons open-source users have shifted to better-performing Chinese models.

But the US open-source ecosystem may be gradually awakening, with efforts by American innovators to enhance their competitiveness.

In July, the White House released an AI Action Plan that called on the federal government to “Promote Open-Source and Open-Weight AI.”

In August, OpenAI, the creator of ChatGPT, released its first open-source model in five years. Announcing the model’s launch, OpenAI referenced the importance of American open-source models, stating, “Broad access to these capable open-weight models created in the US helps expand democratic AI.”

The Seattle-based Allen Institute also released its latest open-source model, Olmo 3, at the end of November, designed to help users “quickly build reliable features for research, education, or applications,” according to the launch announcement.

Lambert from the Allen Institute also launched the “ATOM Project” (American Truly Open Models).

The ATOM Project’s manifesto states: “America has lost its lead in both performance and adoption in open models and is on track to fall further behind.”

“If we want to be the leading nation in the age of AI, we cannot cede such a critical piece of the ecosystem to any one nation,” Lambert said in a statement to NBC News.

America

Twenty-five US states sue Trump administration over Medicaid work-requirement exemptions

Published

on

A coalition of 25 US states and the District of Columbia has filed a joint lawsuit against the Trump administration, challenging a new regulation that restricts work-requirement exemptions for medically frail individuals under Medicaid, the government health insurance program for low-income populations.

The lawsuit, filed on Monday in the US District Court for the District of Massachusetts, alleges that the Centers for Medicare & Medicaid Services (CMS) violated statutory protections established by Congress through its issuance of an interim final rule governing who qualifies for exemptions from the new work mandates.

In their joint complaint, the states argue that the newly adopted rule “dramatically narrows the work exemption boundaries legally secured by Congress for some of the most vulnerable members of the Medicaid program.”

The states contend that the regulation will cause a significant number of individuals who are currently working or who legitimately qualify for exemptions to lose their health coverage or be denied access to these vital services.

“This regulation introduces new rules that restrict who should be exempted due to their medically frail status, forcing these vulnerable individuals who require healthcare services to navigate unnecessary bureaucratic hurdles to obtain and maintain their vital health coverage,” the lawsuit states.

The rule, published earlier this month, serves as implementation guidance for how the work requirements enacted under the “One Big Beautiful Bill Act” will be applied across 42 states and the District of Columbia.

Republican lawmakers and administration officials have defended the policy, characterizing it as a mechanism to combat waste, fraud, and abuse within the Medicaid program.

Under the new rules, which are scheduled to take effect in January, beneficiaries enrolled in expanded Medicaid programs must work, participate in volunteer activities, attend an educational institution at least part-time, or take part in job training programs for at least 80 hours per month to maintain their insurance coverage.

While the statutory text of the legislation outlines various exceptions for specific vulnerable groups—explicitly exempting “medically frail” individuals from the mandate—the statute did not provide a precise definition for the term.

The administration’s new rules narrow the definition of medical frailty by tying it directly to an individual’s capacity to work. Under the new regulation, to qualify for an exemption, a beneficiary must prove that their medical condition completely prevents them from working.

State governments state that they had spent months negotiating implementation plans with CMS prior to the publication of the regulation, but were caught unprepared by this highly restrictive definition, which they argue was not present in the legislative text.

State officials emphasize that individuals who are legally entitled to protection risk losing their health coverage because they will be unable to overcome the bureaucratic barriers imposed to prove their exempt status.

“These changes flagrantly ignore the concrete evidence that the agency was required to consider, or which was already before it,” the complaint states regarding the agency’s decision-making process. “Reasonable alternatives and potential major adverse consequences were not adequately evaluated, nor was it clarified what exactly is being demanded of the plaintiff states.”

The plaintiff states point out that Congress deliberately kept the scope of the exemptions broad when drafting the legislation.

Stressing that the broad exemptions in the law are well-founded, the complaint states: “Individuals with disabilities, patients undergoing cancer treatment, or people battling serious and complex health conditions must not be placed at risk of losing this vital care that helps them maintain their health.”

Continue Reading

America

Bipartisan majority of US voters back strict government reviews for advanced AI, poll shows

Published

on

US demand for strict artificial intelligence regulation is rising, driven by a powerful bipartisan consensus that favors tighter rules for the technology, a new public opinion survey shows.

The poll, conducted by the Artificial Intelligence Policy Institute (AIPI), revealed that 68% of respondents support a requirement for the government to subject “the most advanced AI models to a formal review process before they are widely released.”

In contrast, 20% of participants expressed the view that the government should “rely primarily on companies to test their own AI models and step in mainly after problems arise.” The remaining 12% of respondents were undecided on which approach to support.

Broken down by political affiliation, the formal review process is supported by 76% of Democrats, 64% of Republicans, and 63% of independent voters.

Conversely, the survey showed that 15% of Democrats, 24% of Republicans, and 23% of independents believe the government should rely chiefly on self-testing by companies.

Backlash against data centers grows

Efforts to block or restrict the construction of data centers have gained momentum at both the state and local government levels in recent months.

Public opposition to the construction of massive AI infrastructure projects in local communities is becoming increasingly pronounced across the US.

Two months ago, the Maine State Legislature passed a bill halting the development of large-scale data centers, making it the first state in the country to take such action.

Resistance to data centers has been on the rise for the past several years. Local communities have expressed growing concern over the high energy consumption of these infrastructure facilities and their subsequent impact on the cost of living and the environment.

In March, Democratic Representative Alexandria Ocasio-Cortez and Independent Senator Bernie Sanders announced the Artificial Intelligence Data Center Moratorium Act.

The proposed legislation seeks to halt the construction of AI infrastructure until lawmakers enact measures that mandate state reviews of AI products, prevent mass job losses, and limit increases in consumer electricity prices.

The public opinion poll was conducted between June 10 and June 11 among 1,007 likely voters. The survey has a margin of error of 4.2%.

Continue Reading

America

US Supreme Court bolsters Trump’s executive power over agencies but blocks swift removal of Fed governor

Published

on

The US Supreme Court on Monday delivered a mixed set of rulings for President Donald Trump, simultaneously bolstering his executive control over independent federal agencies while rejecting a key pillar of his political agenda aimed at restricting mail-in voting.

The justices ruled that the heads of independent agencies can be removed by the president, significantly strengthening executive authority over the federal bureaucracy. However, the court also ruled that Federal Reserve Governor Lisa Cook must be afforded due process rights before she can be removed from her post, and rejected the president’s appeal in a separate civil lawsuit.

Major expansion of presidential authority

In a 6-3 decision, the court cleared the way for Trump to remove Federal Trade Commission (FTC) Commissioner Rebecca Slaughter, triggering a major expansion of the president’s removal powers.

The ruling sweeps aside 91 years of judicial precedent that had guaranteed a degree of independence from the White House for certain regulatory agencies, achieving a long-sought goal of conservative legal scholars.

Trump characterized the decision as the “largest increase” in presidential powers seen in a century.

The conservative majority framed the ruling as restoring the presidency to its constitutionally intended form.

Chief Justice John Roberts wrote that Trump’s expanded removal power is inherent to the constitutional system:

“When authority is exercised well, the public knows whom to thank; when it is exercised poorly, they know whom to blame and whom to remove. This is the very foundation of our system of government.”

The ruling enables the president to dismiss officials across numerous agencies beyond the FTC.

More than a dozen other agencies across the executive branch enjoy similar protections. These bodies oversee critical sectors, including nuclear energy, aircraft accident investigations, product safety recalls, and credit unions.

The court’s liberal justices dissented, arguing that the decision grants Trump a level of authority “unknown even to the British Crown.”

“Today, the Court casts aside this democratic regime in favor of one that distorts the structure of government to fit the majority’s theory of unitary and absolute executive control,” Justice Sonia Sotomayor wrote. “The result is a president who emerges with greater power than ever before.”

Justice Neil Gorsuch suggested that Congress retains the power to remedy the balance:

“The power to define new regulatory infractions remains, but the pen is now ultimately in the president’s hand. The ability to adjudicate disputes within the house continues, but that house is now white.”

Acknowledging that the decision concentrates presidential power, Gorsuch argued that the remedy lies with the legislative branch, which could make the agencies less powerful by stripping away their broad authority to regulate American life.

Roberts rules both for and against Trump in a single day

While Roberts led his conservative colleagues in greenlighting the president’s authority to remove certain independent agency heads without cause, the court stopped short of extending that same immediate authority to the Federal Reserve.

In a 5-4 decision also authored by the Chief Justice, the majority found that Trump failed to provide adequate due process to Federal Reserve Board Governor Lisa Cook before attempting to remove her over allegations of mortgage fraud.

Roberts argued that bypassing this step would allow a president to dismiss a Federal Reserve Board member “at any time, for any reason, without prior notice, and without subsequent judicial review.”

The decision capped a contradictory day for Trump’s executive powers, with Roberts positioned at the center of the rulings.

Roberts is widely known for attempting to keep the court out of partisan political battles. However, Trump’s agenda has continued to dominate the docket at a time when public approval of the court has fallen to record lows.

In the Federal Reserve case, Roberts went out of his way to emphasize that the ruling against Trump was narrow in scope.

He stressed that the decision did not resolve the ultimate question of whether Cook could eventually be removed.

This qualification allowed Trump to maintain a confident posture. Pressing for Cook’s removal shortly after the ruling was announced, Trump took to Truth Social to emphasize that the decision was merely “purely procedural.”

Trump suffers defeat on mail-in ballots

In another significant setback for Trump, the Supreme Court ruled 5-4 to uphold a Mississippi law that allows mail-in ballots arriving up to five days after Election Day to be counted.

Writing for the majority, Justice Amy Coney Barrett ruled that the Mississippi statute does not conflict with federal election law, defeating an effort by the Republican National Committee to halt the post-Election Day ballot-counting practice.

Quoting from the Federalist Papers, Barrett wrote:

“The Framers recognized the difficulty of drafting election laws ‘applicable to every possible change in the state of the country.’ For that reason, rather than constitutionalizing election law, they decided that ‘a discretionary power’ over elections ‘must exist somewhere.’ Suffice it to say, that power was not given to this court.”

Justice Samuel Alito led the dissent alongside the court’s three other conservative justices.

“In this day and age, not all voting occurs in person on Election Day. Both mail-in voting and early voting have proliferated, and the respondents do not contest the legality of these modern practices. Neither do I. But the adoption of these practices cannot alter the fact that, under federal law, the collective choice of the electorate must still be authoritatively expressed on Election Day,” Alito wrote.

Reacting to the defeat, Trump renewed his push for Congress to pass the “American Voter Eligibility Protection Act.”

Hours later, however, he conceded that the bill was unlikely to pass, stating: “Because we have four, maybe five Republican senators who absolutely will not vote for it. It’s crazy.”

Trump appointees cast decisive swing votes

Two of the conservative justices appointed by Trump cast decisive votes against him in the separate cases, developments that could leave Justices Brett Kavanaugh and Barrett open to criticism from the president.

During his tenure and post-presidency, Trump has frequently expressed dissatisfaction with Supreme Court rulings that go against him, particularly when those decisions involve his own appointees.

Barrett, the most junior conservative on the bench, authored the majority opinion on the mail-in voting case. She and Roberts joined the court’s three liberal justices to form the majority, leaving the other four conservative justices in dissent.

The ruling deals a blow to the Trump administration’s systematic campaign against mail-in voting, a practice the president has repeatedly claimed contributes to widespread fraud, despite a lack of supporting evidence.

Meanwhile, Kavanaugh joined the 5-4 majority that ruled against Trump in the Federal Reserve case, voting alongside Roberts and the three liberal justices.

In a brief concurring opinion, Kavanaugh highlighted the unique position of the central bank and urged the court to formally protect its independence.

“Even temporary uncertainty regarding the status of the Federal Reserve—including confusion over whether the president can summarily dismiss multiple board members at will—could trigger political turmoil and cause upheaval in the US and global economies. I would not embark on that road,” Kavanaugh wrote.

Continue Reading

MOST READ

Turkey