US House Select Committee on Chinese Communist Party (CCP) release new investigative report exposing DeepSeek as a serious national security threat to US; Demand answers from Nvidia
- Titled, “DeepSeek Unmasked: Exposing the CCP’s Latest Tool For Spying, Stealing, and Subverting U.S. Export Control Restrictions,” the report reveals that DeepSeek covertly funnels American user data to the Chinese Communist Party, manipulates information to align with CCP propaganda, and was trained using material unlawfully obtained from U.S. AI models. - The report also highlights reporting that DeepSeek operates on tens of thousands of Nvidia chips—some of which are subject to U.S. export controls.
- Key findings include:
- Censorship by Design: More than 85% of DeepSeek’s responses are manipulated to suppress content related to democracy, Taiwan, Hong Kong, and human rights—without disclosure to users.
- Foreign Control: DeepSeek is owned and operated by a CCP-linked company led by Lian Wenfang and ideologically aligned with Xi Jinping Thought.
- U.S. User Data at Risk: The platform funnels American user data through unsecured networks to China, serving as a high-value open-source intelligence asset for the CCP.
- Surveillance Network Ties: DeepSeek’s infrastructure is linked to Chinese state-affiliated firms including ByteDance, Baidu, Tencent, and China Mobile—entities known for censorship, surveillance, and data harvesting.
- Illicit Chip Procurement: DeepSeek was reportedly developed using over 60,000 Nvidia chips, which may have been obtained in circumvention of U.S. export controls.
- Corporate Complicity: Public records show Nvidia CEO Jensen Huang directed the company to design a modified chip specifically to exploit regulatory loopholes after October 2023 restrictions. The Trump Administration is working to close this loophole.
- In response to the report’s findings, the Select Committee has sent a formal letter to Nvidia demanding answers about sales to China and Southeast Asia to examine whether and how its chips ended up powering DeepSeek’s AI models—despite U.S. export restrictions.
- The Committee will continue investigating how American innovation is being exploited by the Chinese Communist Party and will work to ensure that U.S. companies are not enabling the CCP’s efforts to undermine our national security.
New report backs iPhone Fold punch-hole camera with Touch ID power button
While the end of the wait for a foldable iPhone appears to be in sight, there’s uncertainty over Apple’s plans for the outer and inner displays of what’s being colloquially referred to as the iPhone Fold.
A new report adds weight the the suggestion that there will be no Dynamic Island on the outer display, Apple instead opting for a single punch-hole for the front-facing camera, with Touch ID used to unlock the device …
The iPhone Fold
For a long time, two things have been unclear. First, will Apple take the Flip or a Fold approach? It now seems clear the company is taking the Fold path, with a normal-sized iPhone unfolding into something roughly the size of an iPad mini.
Second, will Apple embed Face ID beneath the display – a problem the company has reportedly solved – or would it instead put a Touch ID sensor into the power button, iPad Air style?
There’s now growing evidence of the latter approach.
Punch-hole camera, with Touch ID power button
Apple analyst Ming-Chi Kuo indicated last month that Apple would not use Face ID on the iPhone Fold, as there simply isn’t room in the thin display.
Kuo says thickness and internal space constraints appear to prevent the Face ID module from being included in the first-generation device. Touch ID biometric authentication will be available instead, through a capacitive side button. This is similar to the Touch ID button in the iPad Air.
If accurate, this would mean Apple could forgo the Dynamic Island in favor of a simple punch-hole camera, and a new report backs this. The yeux1122 blog, which rounds up reports from Chinese social media site Weibo, says that the device will have a punch-hole design.
The content about the internal and external display of Apple’s first foldable device came out. (@ weibo)Internal 7.76 inches 2713 x 1920 resolutionExternal 5.49 inches 20881422p resolutionPunch hole solution that is almost the same as the existing foldables.
The screen size and resolution almost certainly comes direct from yesterday’s Digital chat station report, while the reference to “existing foldables” means existing Android folding phones like the Samsung Galaxy Z Fold 6 and the Google Pixel 9 Pro Fold.
Why VC Investors and Startups Should Be Worried About the Bond Market
Rising interest rates keep investors away from risky assets, and recent volatility could stoke those fears
Last week’s surge in interest rates, sparked by President Donald Trump’s tariff moves, might end a recent uptick in venture capital investing.
Global VC deal value hit $126.3 billion in the first quarter, its highest level in nearly three years, according to data provider PitchBook. The rebound came as the Federal Reserve started cutting interest rates last year, putting the VC market on more stable ground. High interest rates tend to choke off VC investing, while lower rates can have the opposite effect.
The 10-year Treasury bond yield’s surge from about 4% 10 days ago to over 4.5% last week raises the specter of another VC drought, similar to the one that occurred in 2022, when rates jumped from 3% to nearly 4%.
For a longer-term view of the connection between interest rates and VC investing, look at the past decade, when falling 10-year Treasury bond yields set the stage for a more active tech dealmaking landscape. The yield on the 10-year Treasury bond reached a recent peak in the third quarter of 2018 and tumbled when Covid-19 hit, the economy stalled and the Fed slashed rates. The 10-year yield hit a trough of 0.65% in the second quarter of 2020, the average for the period.
A surge in VC dealmaking followed: Quarterly VC deal value swelled from roughly $79 billion in the second quarter of 2020 to a high of $211 billion in the last quarter of 2021, according to data from PitchBook.
That flipped when surging inflation prompted the Federal Reserve to jack up interest rates, cutting off funding for startups.
While volatility has died down since Trump retreated on tariffs, the market action last week was worrisome. Not only did the Treasury yields pop, meaning investors dumped the bonds, but the stock market fell too. Typically, when stocks tumble, investors turn to the safety of Treasurys. This time they also sold Treasury bonds, a sign they had lost confidence in the U.S. government’s policies.
Rates have fallen a bit this week, which should help money flow into risky assets. But other economic indicators are flashing yellow or in some cases red. Consumers are bracing for a weaker economy and surging prices, and economists surveyed by The Wall Street Journal have cut their expectations for growth while predicting higher unemployment and inflation.
None of those factors give investors confidence to make big bets on startups.
OpenAI launches a pair of AI reasoning models, o3 and o4-mini
OpenAI announced on Wednesday the launch of o3 and o4-mini, new AI reasoning models designed to pause and work through questions before responding.
The company calls o3 its most advanced reasoning model ever, outperforming the company’s previous models on tests measuring math, coding, reasoning, science, and visual understanding capabilities. Meanwhile, o4-mini offers what OpenAI says is a competitive trade-off between price, speed, and performance — three factors developers often consider when choosing an AI model to power their applications.
Unlike previous reasoning models, o3 and o4-mini can generate responses using tools in ChatGPT such as web browsing, Python code execution, image processing, and image generation. Starting today, the models, plus a variant of o4-mini called “o4-mini-high” that spends more time crafting answers to improve its reliability, are available for subscribers to OpenAI’s Pro, Plus, and Team plans.
The new models are part of OpenAI’s effort to beat out Google, Meta, xAI, Anthropic, and DeepSeek in the cutthroat global AI race. While OpenAI was first to release an AI reasoning model, o1, competitors quickly followed with versions of their own that match or exceed the performance of OpenAI’s lineup. In fact, reasoning models have begun to dominate the field as AI labs look to eke more performance out of their systems.
O3 nearly wasn’t released in ChatGPT. OpenAI CEO Sam Altman signaled in February that the company intended to devote more resources to a sophisticated alternative that incorporated o3’s technology. But competitive pressure seemingly spurred OpenAI to reverse course in the end.
OpenAI says that o3 achieves state-of-the-art performance on SWE-bench verified (without custom scaffolding), a test measuring coding abilities, scoring 69.1%. The o4-mini model achieves similar performance, scoring 68.1%. OpenAI’s next best model, o3-mini, scored 49.3% on the test, while Claude 3.7 Sonnet scored 62.3%.
OpenAI claims that o3 and o4-mini are its first models that can “think with images.” In practice, users can upload images to ChatGPT, such as whiteboard sketches or diagrams from PDFs, and the models will analyze the images during their “chain-of-thought” phase before answering. Thanks to this newfound ability, o3 and o4-mini can understand blurry and low-quality images and can perform tasks such as zooming or rotating images as they reason.
Beyond image-processing capabilities, o3 and o4-mini can run and execute Python code directly in your browser via ChatGPT’s Canvas feature, and search the web when asked about current events.
In addition to ChatGPT, all three models — o3, o4-mini, and o4-mini-high — will be available via OpenAI’s developer-facing endpoints, the Chat Completions API and Responses API, allowing engineers to build applications with the company’s models at usage-based rates.
OpenAI is charging developers a relatively low price for o3, given its improved performance, at $10 per million input tokens (roughly 750,000 words, longer than the Lord of the Rings series) and $40 per million output tokens. For o4-mini, OpenAI is charging the same as o3-mini, $1.10 per million input tokens and $4.40 per million output tokens.
In the coming weeks, OpenAI says it plans to release o3-pro, a version of o3 that uses more computing resources to produce its answers, exclusively for ChatGPT Pro subscribers.
OpenAI CEO Sam Altman has indicated o3 and o4-mini may be its last standalone AI reasoning models in ChatGPT before GPT-5, a model that the company has said will unify traditional models like GPT-4.1 with its reasoning models.