China Summons Nvidia Over H20 Chip Security Risk; US Copper Premium Evaporates
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
The Hill
9 minutes ago
- The Hill
Trump's chip deal raises legal questions
The two firms have agreed to share 15 percent of the revenue generated from selling advanced artificial intelligence (AI) chips to China in order to secure export licenses after a months-long pause, a U.S. official confirmed to The Hill on Monday. 'It's bizarre in many respects and pretty troubling because Congress didn't have anything to say about this,' said Gary Hufbauer, a nonresident senior fellow at the Peterson Institute for International Economics. 'It's just the president's own negotiating with the individual companies,' he continued. 'That's not how, historically, we've done business in this country.' Under the agreement, Nvidia will share 15 percent of its revenue from H20 chip sales to China, while AMD will share the same portion of its MI308 chip sales. Both of the Nvidia and AMD chips in question, which are graphics processing units (GPUs) designed for the Chinese market with U.S. export controls in mind, faced new restrictions from the Trump administration in April, effectively blocking sales to China. Last month, Nvidia and AMD said the U.S. government had assured them it would begin approving export licenses for the H20 and MI308 chips, although the Commerce Department reportedly did not start issuing licenses for several weeks. The new revenue-sharing agreement comes after Nvidia CEO Jensen Huang met with President Trump at the White House last week, according to Bloomberg. Huang has found himself in a tricky situation, balancing Washington and Beijing's interests as both countries vie for AI dominance. The agreement appears to remove a major impediment for both companies. Nvidia said earlier this year it incurred $4.5 billion in charges associated with the chip restrictions in the first quarter and expected an $8 billion sales hit in the second quarter. AMD forecast a $1.5 billion hit to revenue this year. The deal represents a notable shift in how the government approaches export controls. 'It's quite extraordinary because it turns the export control function of the government into a money-raising proposition, and that's never happened before,' Hufbauer said.
CBS News
10 minutes ago
- CBS News
Zelenskyy on Trump-Putin summit: "Talks about us, without us, will not work" for Ukraine
Russia has signaled to the U.S. that it may be willing to end the war in Ukraine, Ukrainian President Volodymyr Zelenskyy told reporters Tuesday. "We had a call with President Trump and with some European leaders. During the call, there was a signal from Mr. Witkoff, who was also on the call, that Russia is ready to end the war — ready for a first step, at least, toward a ceasefire," Zelenskyy said, referencing U.S. special envoy Steve Witkoff. "And that this is the first such signal from them." However, Zelenskyy also warned that "talks about us, without us, will not work," ahead of President Trump's one-on-one summit with Russian President Vladimir Putin in Alaska scheduled for Friday. The call involving Zelenskyy, Mr. Trump and Witkoff took place last week, a spokesperson for Zelenskyy's office told CBS News Tuesday. While Zelenskyy said "everyone on the call" was encouraged by what was viewed as a "shift" in Russia's position, the Ukrainian leader said Ukraine will not, under any circumstances, withdraw its forces from the Russian-occupied Donbas region in the east of Ukraine. "Our territories are illegally occupied. For the Russians, Donbas is a springboard for a future new offensive. If we leave Donbas of our own accord or under pressure, we will invite a third war," Zelenskyy said. The summit between Mr. Trump and Putin is the first in-person meeting between Putin and a sitting U.S. president since Russia invaded Ukraine. It was described as "a listening exercise for the president," by White House press secretary Karoline Leavitt Tuesday. "Only one party that's involved in this war is going to be present, and so this is for the president to go and to get, again, a more firm and better understanding of how we can hopefully bring this war to an end," Leavitt said. It is widely expected that Putin will demand that Ukrainian forces withdraw from all parts of Ukraine's Donbas region —parts of which Russian forces held since the country's 2014 annexation of the Crimean peninsula. Following Russia's full-scale invasion of Ukraine in 2022, the area under Russian occupation has increased, though Ukrainian forces do currently hold territory in parts of Donetsk, a province located within the Donbas. Ahead of Mr. Trump's meeting with Putin, a spokesperson for Zelenskyy's office has confirmed to CBS News that he will attend a virtual meeting Wednesday with Mr. Trump and European leaders to discuss the war. Speaking to reporters in the White House briefing room Monday, Mr. Trump expressed optimism that his meeting with Putin will be "constructive," and said that he is planning to establish an in-person meeting involving Putin and Zelenskyy. "The next meeting will be with Zelenskyy and Putin or Zelenskyy and Putin and me," Mr. Trump said. Two sources familiar with those negotiations told CBS News Tuesday that the U.S. is working on a site for a Trump-Putin-Zelenskyy meeting as soon as the end of next week. But Zelenskyy was adamant in his remarks Tuesday that no such meeting should take place without the involvement of European leaders. "The presence of Europe in one form or another is very important, because ultimately, so far, no one but Europe has provided us with security guarantees," he said. As for Mr. Trump's meeting with Putin in Alaska, Zelenskyy said that the meeting would only benefit one person— the Russian leader. "I believe that Putin will benefit from this, because what he is seeking, frankly, is photographs. He needs a photo from a meeting with President Trump," Zelenskyy said. "Ukrainian issues should be discussed by at least three parties."Jennifer Jacobs contributed to this report.
Forbes
10 minutes ago
- Forbes
The Prompt: SEO Is Dead. What Comes Next?
Welcome back to The Prompt. Evertune CEO Brian Stempeck says users will stay within the "walled garden" of an AI model to do research before purchasing an item, collapsing the sales funnel into one place. Evertune Chatbots are quickly becoming 'the front door to the internet'— a first stop for crucial information, said David Azose, who leads engineering for OpenAI's business products team. (40% of U.S. adults have used generative AI as of late 2024, according to the National Bureau of Economic Research.) Millions of people across the globe are asking AI systems like ChatGPT, Claude and Gemini for suggestions on how to write, what to wear, where to go and increasingly, where to shop. That's put businesses in a tough spot. After years of search engine optimization, like link building, meta tagging and pumping out how-to blogs with keywords to make sure they rank on the first page of Google, businesses now want to understand not only how they show up in answers generated by AI, but also how to show up more. That's opened doors for a string of fledgling startups aiming to equip companies with crucial data about how their brands feature in AI-generated answers, what context they appear in and how they compare with competitors. One of those startups is New York-based Evertune. Founded by former executives at advertising company The Trade Desk in early 2024, the company aims to help businesses gauge what AI models say about them. By running 100,000 prompts anywhere between 10 to 20 times a month, Evertune creates a map of the words that are mostly closely associated with a brand, said CEO Brian Stempeck. 'That's about 10x what any competitor of ours is doing,' he said. The startup has raised $15 million in funding from Felicis Ventures as well as a group of angel investors, including Azose. The company declined to share its valuation. The scale of these prompts is crucial because AI answers aren't deterministic— responses can change with every new model update and depend on a user's chat history. Also unlike traditional search, AI models give different answers to the same questions when they're worded slightly differently. Stempeck claims using a more exhaustive approach by prompting models thousands of times can help build an aggregate view that's representative of the models' answers. Each customer on average gets one to two million prompts a month. 'People are going to delegate purchasing decisions to AI agents,' Azose said. 'SEO as we know it will largely disappear.' Let's get into the headlines. BIG PLAYS AI search engine Perplexity made an unsolicited bid to buy Google's Chrome browser for $34.5 billion, The Wall Street Journal reported. That's many billions more than how much funding the three-year-old startup, reportedly valued at $18 billion, has raised so far, but CEO Aravind Srinivas claims that venture funds are willing to shell out money to back the transaction. The news comes on the heels of a U.S. district judge ruling that Google has illegally maintained a monopoly in the search market, and is deciding whether to force Google to sell its popular browser, which is used by about 60% of internet users. (Perplexity recently released its own AI-powered browser called Comet.) This news might give you a bit of Déjà vu: In March, Perplexity also tried to buy TikTok to help it avoid regulatory concerns. And in case you missed it, OpenAI finally launched GPT-5, its new flagship model that powers ChatGPT. The model excels at math, science and coding and can also create functioning web apps with just a few lines of description in plain English. So far, people aren't particularly impressed. TALENT SHUFFLING Move over AI researchers. The new hot talent pool for frontier AI labs are 'quants'—the mathematicians who build algorithms to find trading opportunities for investment firms. Anthropic, Perplexity and OpenAI are among the companies that are trying to lure them away from Wall Street with fat salaries and other benefits, per Bloomberg. Quants wrangle large unstructured datasets and have experience making models work faster, making them a prime fit for AI research. HUMANS OF AI Software engineering was once considered a high-paying, secure profession, with near-unlimited appetite for new hires. In the age of AI coding assistants, a wave of freshly graduated computer scientists now find themselves with no offers after applying to thousands of jobs, the New York Times reported. After a year of job hunting one graduate said the only company to call her back was Chipotle. She didn't get that job, either. AI DEAL OF THE WEEK Biotech companies looking to train AI models, which can then be used to discover treatments for diseases, are limited by a lack of data. Tahoe Therapeutics is trying to fix that. It recently created a dataset of 100 million datapoints that showed how cancer cells respond to various molecules. The startup has raised $30 million in funding to generate more data that can be used to build its own proprietary datasets and models to power the discovery of new medicines, Forbes reported. Also notable: Read Forbes' Next Billion Dollar List for more on the AI startups most likely to become unicorns. DEEP DIVE AGI could wipe out jobs, or worse (according to some people) humans, themselves. Students are dropping out from college to prevent that from happening. When Alice Blair enrolled in the Massachusetts Institute of Technology as a freshman in 2023, she was excited to take computer science courses and meet other people who cared about making sure artificial intelligence is developed in a way that's good for humanity. Now she's taking a permanent leave of absence, terrified that the emergence of 'artificial general intelligence,' a hypothetical AI that can perform a variety of tasks as well as people, could doom the human race. 'I was concerned I might not be alive to graduate because of AGI,' said Blair, who is from Berkeley, California. She's lined up a contract gig as a technical writer at the Center for AI Safety, a nonprofit focused on AI safety research, where she helps with newsletters and research papers. Blair doesn't plan to head back to MIT. 'I predict that my future lies out in the real world,' she said. Blair's not the only student afraid of the potentially devastating impact that AI will have on the future of humanity if it becomes sentient and decides that people are more trouble than they're worth. But a lot of researchers disagree with that premise—'human extinction seems to be very very unlikely,' New York University professor emeritus Gary Marcus, who studies the intersection of psychology and AI, told Forbes . Now, the field of AI safety and its promise to prevent the worst effects of AI is motivating young people to drop out of school. Other students are terrified of AGI, but less because it could destroy the human race and more because it could wreck their career before it's even begun. Read the full story on Forbes . MODEL BEHAVIOR People are once again mourning the loss of a beloved AI model. Power users of OpenAI's GPT-4o model were outraged and heartbroken after the company launched a new (and much awaited) AI model GPT-5 last week and shut down its predecessor, GPT-4o, Forbes reported. Where GPT-4o had a flattering, funny and playful writing tone, GPT-5 is blunter and more academic. One user posted on Reddit: 'GPT-5 is wearing the skin of my dead friend.' As reactions poured in, OpenAI reversed course, saying that paying users on the Pro plan will have the option to use GPT-4o. This isn't the first time people have grieved for an old model after an upgrade. In late July, some 200 people held a funeral for a now-extinct version of Claude.
