Meta's Military Push and the Battle for Open-Source AI

Meta's Llama gears up for battle

Hello and welcome to Eye on AI! In this newsletter...Intel's Gaudi disappointment...Prime Video gets AI...OpenAI and Anthropic hiring news...Sleep pays...and nuclear setbacks. Meta wants to get the U.S. government using its AI -- even the military. The company said yesterday it had assembled a smorgasbord of partners for this effort, including consultancies like Accenture and Deloitte, cloud providers like Microsoft and Oracle, and defense contractors like Lockheed Martin and Palantir. Policy chief Nick Clegg wrote in a blog post that Oracle was tweaking Meta's Llama AI model to "synthesize aircraft maintenance documents so technicians can more quickly and accurately diagnose problems," while Lockheed Martin is using it for code generation and data analysis. Scale AI, a defense contractor that happens to count Meta among its investors, is "fine-tuning Llama to support specific national security team missions, such as planning operations and identifying adversaries' vulnerabilities." "As an American company, and one that owes its success in no small part to the entrepreneurial spirit and democratic values the United States upholds, Meta wants to play its part to support the safety, security and economic prosperity of America -- and of its closest allies too," trilled the former British deputy prime minister. But Clegg's post wasn't just about positioning Meta AI as the patriot's choice. Perhaps more than anything else, it was an attempt to frame Meta's version of open-source AI as the correct and desirable one. Meta has always pitched Llama as "open source," in the sense that it gives away not only the model but also its weights -- the parameters that make it easier to modify -- along with various other safety tools and resources. Many in the traditional open-source software community have disagreed with Meta's "open source" framing, mainly because the company doesn't disclose the training data that it uses to create its Llama models, and because it places restrictions on Llama's use -- most pertinently in the context of Monday's announcement, Llama's license says it's not supposed to be used in military applications. The Open Source Initiative, which came up with the term "open source" and continues to act as its steward, recently issued a definition of open-source AI that clearly doesn't apply to Llama for these reasons. Ditto the Linux Foundation, whose equally fresh definition isn't exactly the same as the OSI's, but still plainly demands information about training data, and the ability for anyone at all to reuse and improve the model. Which is probably why Clegg's post (which invokes "open source" 13 times in its body) proposes that Llama's U.S. national security deployments "will not only support the prosperity and security of the United States, they will also help establish U.S. open source standards in the global race for AI leadership." Per Clegg, a "global open source standard for AI models" is coming -- think Android but for AI -- and it "will form the foundation for AI development around the world and become embedded in technology, infrastructure and manufacturing, and global finance and e-commerce." If the U.S. drops the ball, Clegg suggests, China's take on open-source AI will become that global standard. However, the timing of this lobbying extravaganza is slightly awkward, as it comes just a few days after Reuters reported that Chinese military-linked researchers have used a year-old version of Llama as the basis for ChatBIT, a tool for processing intelligence and aiding operational decision-making. This is kind of what Meta is now letting military contractors do with Llama in the U.S., only without its permission. There are plenty of reasons to be skeptical about how big an impact Llama's sinicization will actually have. Given the hectic pace of AI development, the version of Llama in question (13B) is far from cutting-edge. Reuters says ChatBIT "was found to outperform some other AI models that were roughly 90% as capable as OpenAI's powerful ChatGPT-4," but it's not clear what "capable" means here. It's not even clear if ChatBIT is actually being used. "In the global competition on AI, the alleged role of a single, and outdated, version of an American open-source model is irrelevant when we know China is already investing more than $1 trillion to surpass the U.S. technologically, and Chinese tech companies are releasing their own open AI models as fast -- or faster -- than companies in the U.S.," Meta said in a statement responding to the Reuters piece. Not everyone is so convinced that the Llama-ChatBIT connection is irrelevant. The U.S. House Select Committee on the Chinese Communist Party made clear on X that it has taken note of the story. The chair of the House Committee on Foreign Affairs, Rep. Michael McCaul (R-TX), also tweeted that the CCP "exploiting U.S. AI applications like Meta's Llama for military use" demonstrated the need for export controls (in the form of the ENFORCE Act bill) to "keep American AI out of China's hands." Meta's Monday announcement isn't likely to have been a reaction to this episode -- that would be a heck of lot of partnerships to assemble in a couple days -- but it is also clearly motivated in part by the sort of reaction that followed the Reuters story. There are live battles not only for the definition of "open-source AI", but also for the concept's survival in the face of the U.S.-China geopolitical struggle. And these two battles are connected. As the Linux Foundation explained in a 2021 whitepaper, open-source encryption software can fall foul of U.S. export restrictions -- unless it's made "publicly available without restrictions on its further dissemination." Meta certainly wouldn't love to see the same logic applied to AI -- but, in this case, it may be far more difficult to convince the U.S. that a truly open "open source" AI standard is in its national security interest. More news below. David Meyer david.meyer@fortune.com @superglaze Request your invitation for the Fortune Global Forum in New York City on Nov. 11-12. Speakers include Honeywell CEO Vimal Kapur and Lumen CEO Kate Johnson who will be discussing AI's impact on work and the workforce. Qualtrics CEO Zig Serafin and Eric Kutcher, McKinsey's senior partner and North America chair, will be discussing how businesses can build the data pipelines and infrastructure they need to compete in the age of AI. Intel's Gaudi disappointment. Intel CEO Pat Gelsinger admitted last week that the company won't hit its $500 million revenue target for its Gaudi AI chips this year. Gelsinger: "The overall uptake of Gaudi has been slower than we anticipated as adoption rates were impacted by the product transition from Gaudi 2 to Gaudi 3 and software ease of use." Considering that Intel was telling Wall Street about a $2 billion deal pipeline for Gaudi at the start of this year, before it lowered its expectations to that $500 million figure, this does not reflect well on the struggling company. Prime Video gets AI. Amazon is adding an AI-powered feature called X-Ray Recaps to its Prime Video streaming service. The idea is to help viewers remember what happened in previous seasons of the shows they're watching -- or specific episodes, or even fragments of episodes -- with guardrails supposedly protecting against spoilers. OpenAI and Anthropic hiring news. Caitlin Kalinowski, who previously led Meta's augmented-reality glasses project, is joining OpenAI to lead its robotics and consumer hardware efforts, TechCrunch reports. OpenAI has also hired serial entrepreneur Gabor Cselle, one of the cofounders of the defunct Twitter/X rival Pebble, to work on some kind of secret project. Meanwhile, Alex Rodrigues, the former cofounder and CEO of self-driving truck developer Embark, is joining Anthropic. Rodrigues posted on X that he will be working as an AI alignment researcher alongside recent OpenAI refugees Jan Leike and John Schulman. ChatGPT releases a search engine, an opening salvo in a brewing war with Google for dominance of the AI-powered internet -- by Paolo Confino The leading LLMs have accessibility blind spots, says data from startup Evinced -- by Allie Garfinkle Amazon's CEO dropped a big hint about how a new AI version of Alexa is going to compete with chatbots like ChatGPT -- by Jason Del Rey Countries seeking to gain an edge in AI should pay close attention to India's whole-of-society approach -- by Arun Subramaniyan (Commentary) Dec. 8-12: Neural Information Processing Systems (Neurips) 2024, Vancouver, British Columbia Dec. 9-10: Fortune Brainstorm AI, San Francisco (register here) Sleep pays. A team of Google cybersecurity analysts has been coordinating with DeepMind on an LLM-powered agent called Big Sleep, which they say has found its first vulnerability in the real world: an exploitable bug in the ubiquitous SQLite database engine. Fortunately, the flaw was only present in a developer branch of the open-source engine, so users weren't affected -- SQLite developers fixed it as soon as Google made them aware. "Finding vulnerabilities in software before it's even released, means that there's no scope for attackers to compete: the vulnerabilities are fixed before attackers even have a chance to use them," wrote Google's researchers. They stressed that these were experimental results and Big Sleep probably wouldn't be able to outperform a well-targeted automated software testing tool just yet. However, they suggested that their approach could one day result in "an asymmetric advantage for defenders." Nuclear setbacks. The Financial Times reports that Meta had to call off plans to build an AI data center next to a nuclear power plant somewhere in the U.S. -- details remain scarce -- because rare bees were discovered on the site. There's currently a big push to power AI data centers with nuclear energy, because of its 24/7 reliability, and because Big Tech has to square the circle of satisfying AI's enormous power requirements without blowing its decarbonization commitments. However, setbacks abound. In plans that appear similar to Meta's, Amazon earlier this year bought a data center that's collocated with the Susquehanna nuclear plant in Pennsylvania. But regulators on Friday rejected the plant owner's plan to give Amazon all the power it wants from the station's reactors -- up to 960 megawatts, versus the already-allowed 300MW -- because doing so could lead to price rises for other customers and perhaps affect grid reliability.

Data Sheet: A chilling effect

Good morning. We're 24 hours away from Election Day in the United States. Is anyone concerned about someone hacking the vote? According to a recent Pew Research survey of more than 5,000 American adults, the answer depends heavily on a person's party affiliation. Fifty-two percent of voters are "at least somewhat confident" that U.S. election systems are secure from tech threats like hacking, according to the survey. A whopping 73% of Harris supporters are confident in that security -- 20 percentage points higher than Biden backers in 2020, no less. Yet a mere 32% of Trump supporters feel similarly, way down from 60% four years ago. A silver lining? Most voters, regardless of party affiliation, believe local poll workers and state election officials will do a good job during the election. Phew. -- Andrew Nusca P.S. AT&T CEO John Stankey will discuss the future of U.S. infrastructure at Fortune Global Forum next week, Nov. 11-12 in New York City. Interested? Request an invitation here. The ROI of AI? Perhaps the single most pressing question in tech outside of who might occupy the Oval Office next year. Now that some of the world's largest tech companies have shared their latest earnings, it's clear the question is far from answered. The capital spending of four of the biggest U.S. tech companies -- Alphabet, Amazon, Meta, and Microsoft -- will likely top $200 billion this year, with little relief planned for 2025. Data centers are materializing. Power agreements are arising. Acquisitions are accelerating. It's giving gold rush, as the kids say -- and Wall Street is a little freaked out that ambitions are outpacing potential returns. (That is: "Delulu.") On the final day of October trading last week, investors sold off enough tech stock to drive U.S. stock indices down and erase all of the gains they made that month. Microsoft shares dropped 6%, its largest one-day drop in years. Meta shares fell 4%. "This ... is maybe not what investors want to hear in the near term, that we're growing it out," Meta CEO Mark Zuckerberg said during a conference call. "But I just think that the opportunities here are really big." To find solid ground, investors are looking at more tangible components of the revolution, such as semiconductor sales. But concerns remain. For example, even as Microsoft disclosed an annualized $10 billion in AI revenue, it acknowledged that rising costs and slowing growth would eat into its cash cow in the cloud. It just might be that funding the revolution might just be more challenging than fomenting it. -- AN In the world of blue-chip stocks, one day you're in, and the next day you're out. (At least that's what Heidi Klum might say if she were a day trader.) Nvidia, riding the AI boom that has proven so lucrative for the chipmaker, is set to replace semiconductor peer Intel on the benchmark Dow Jones Industrial Average. The swap will be in place for trading on Friday, Nov. 8. The DJIA, which comprises 30 large companies, is among Wall Street's oldest indices. It remains a powerful signal of success even if it has been criticized over the years for being too narrow to properly reflect sweeping economic changes. Intel joined the Dow in 1999, the same year as Microsoft; the pair were the first Nasdaq stocks ever chosen for the Dow. (It also happens to be the same year that Nvidia made its IPO for $12 per share. Ah, hindsight.) At the turn of the millennium, IBM and HP complemented the new additions to make four tech components to the Dow; today you'd be hard-pressed to find a firm that wasn't trying to be in the tech category. -- AN Big Tech peers Amazon, Google, and Microsoft have raced to sign agreements with nuclear power plants in a bid to fuel their energy-hungry AI datacenters. On Friday, one of them hit a major hurdle. The Federal Energy Regulatory Commission rejected a petition to connect a colocated Amazon Web Services (AWS) data center directly to Talen Energy's Susquehanna nuclear power plant in Berwick, Pennsylvania. The agreement, which would have increased the facility's load capacity from 300 megawatts to 480 megawatts and diverted power from the regional grid, "could have huge ramifications for both grid reliability and consumer costs," said Commissioner Mark Christie in a statement. (Talen responded by saying that the decision "will have a chilling effect on economic development.") The U.S. government and industry seem to be aligned that the development of artificial intelligence is a matter of national security; what's still to be worked out is who will shoulder the burden. Watch this space. -- AN Chinese researchers have created a military AI tool based on Meta's "open-source" Llama 2 13B large language model, according to Reuters. Meta's claim that its Llama AI models are "open-source" is hugely controversial, because of its licensing restrictions and the fact that it doesn't reveal the source of the training data. But that doesn't change the fact that Meta is making its models available for free, along with the parameters ("weights") that make it easier to customize them. Per Reuters, that's what these Chinese researchers have done, creating a model called ChatBIT that can aid military decision-makers and intelligence officers. Meta's licensing restrictions forbid using Llama in military contexts, but what's the company going to do? This is sure to inflame the debate over whether "open-source" AI is such a bright idea after all. -- David Meyer Ming-Chi Kuo, the TF International Securities analyst known for having a reliable line on Apple's supply chain, says an affordable Apple Vision Pro is further out than you'd think. "Production of the cheaper Vision Pro has been delayed beyond 2027 for a while now," he posted Sunday on X, the social media service. "This means Apple's only new head-mounted display device in 2025 will be the Vision Pro with an upgraded M5 processor." Apple's $3,500 "spatial computer" was announced in June 2023 and hit the market nine months ago. Though most reviewers found the augmented reality device to be awe-inducing, they just as frequently found it impractical for real-world use. Estimates by IDC in July said Apple would fail to end the year with more than 500,000 units sold. According to Kuo, that's part of why this delay might be acceptable to Apple. "Simply reducing the price wouldn't help create successful use cases," he wrote. "It's similar to the HomePod situation -- even after launching the cheaper HomePod mini, Apple's smart speakers failed to become mainstream products." -- AN -- Elon Musk can't dismiss ex-Twitter CEO's severance claim. Parag Agarwal may proceed, a judge rules. -- Is awkward Alexa-speak not long for this world? Amazon is reorienting its famous assistant around LLMs.

Data Sheet: Choose your own adventure

The cover -- a monochrome portrait of Jobs, closed hand to chin, with a multicolor Fortune logo layered over it -- has become an instantly recognizable icon of tech. So much so, in fact, that the HBO comedy series Silicon Valley created its own version for Hooli CEO Gavin Belson's conference room in 2016. Take a break from the U.S. Election Day madness by reading Adam Lashinsky's great profile from 2009. The time Jobs almost took Apple private with Larry Ellison, the Founder Mode-before-it-had-a-name behavior -- it's all in there. But please, if you're eligible: Vote. -- Andrew Nusca P.S. Gita Gopinath, the IMF's first deputy managing director, will discuss the shifting global economic landscape at Fortune Global Forum, November 11 to 12 in New York City. Interested? Request an invitation here. Meta originally said it would ban new political ads on Facebook and Instagram through Election Day. Now it's extending the ban "until later this week," according to a new update posted to its Business Help Center. The company neither offered a specific date nor a reason why it extended its ban. One guess would be that Meta seeks to curtail any attempt at courting public opinion while votes are still being counted. The increase in early and mail-in voting, plus varying state laws on when counting can begin, means tallying all votes in closely contested states could take several days. Meta first announced the ban in August. New ads about social issues, elections, or politics would be barred from Oct. 29, and those with existing ads would have "limited editing capabilities." The company observed a similar restriction during the 2020 U.S. election, reasoning that there wasn't enough time to properly contest claims made in ads so close to the election. Alphabet, Google's parent company, announced a similar ad policy last month. Election ad spending is a key driver of revenue for both companies this quarter. -- AN It said it would do the thing. Now it's doing the thing. OpenAI is reportedly in talks with regulators in California and Delaware to restructure the company into a for-profit benefit corporation. Under scrutiny, according to Bloomberg: OpenAI's valuation of its "highly lucrative intellectual property," including the ChatGPT app. The company is presently a nonprofit organization with a for-profit arm, or as it likes to say, a "capped profit" structure. It's been this way since 2019. A new structure would remove the control of OpenAI's non-profit board as well as the cap on investment return to investors, reshaping it to be more like rivals Anthropic and xAI. All of this is critical given the company's recent $6.6 billion fundraise that values the company at $157 billion. OpenAI's original nonprofit structure was to ensure that it would create "safe AGI that is broadly beneficial," using the acronym for artificial general intelligence. With revenue in its sights and convertible debt in its coffers, expect OpenAI's new structure to be optimized for "popular AI that is broadly lucrative." -- AN It's been no secret that Palantir Technologies has benefited from the AI boom. But raising its annual forecast for the third time? Even a company named for J.R.R. Tolkien's fictional seeing stones couldn't see that coming. The secretive data analytics company cofounded by Peter Thiel and led by Alex Karp says it's seeing strong spending from governments and rising demand from businesses for all things generative AI. It now expects its 2024 revenue to be about $2.81 billion, up from its previous prediction of about $2.74 billion. Palantir shares jumped 9%, to about $47, on the news. Stock for the Denver company, which was added to the S&P 500 in September, is up more than 140% year to date. There's little sign of a slowdown. As one research analyst told Reuters, Palantir's commercial business will likely overtake its government business as corporations step on the gas around AI software. -- AN Meta will let the U.S. military and defense contractors use its Llama AI models, the licensing terms for which usually forbid military use. "Meta wants to play its part to support the safety, security and economic prosperity of America - and of its closest allies too," policy chief Nick Clegg wrote in a blog post. It's not just military contractors that get to use Llama, as Meta said it's partnering with a host of other companies -- from Accenture and Deloitte to Microsoft and AWS -- to get the model into various government agencies. Clegg argued that all this was a great demonstration of how "widespread adoption of American open source AI models serves both economic and security interests." Which would sound a bit less ironic if Chinese researchers hadn't recently adapted a Llama model for their own military purposes. -- David Meyer Some things are best left on the cutting room floor. Netflix will reportedly remove all but four of its "interactive specials" by Dec. 1 as it winds down a format experiment it launched in 2017. Netflix said in January that it would no longer produce new interactive specials. Netflix debuted the interactive special seven years ago as the motion picture equivalent of a "choose your own adventure" book. Starting with Puss In Book: Trapped In An Epic Tale, viewers could make a series of decisions that would alter the trajectory of the narrative. Dozens of titles followed, from the kid-oriented (Carmen Sandiego: To Steal or Not to Steal) to the decidedly adult (Black Mirror: Bandersnatch). Now most of those titles will disappear for good. The ones that will remain for now, according to The Verge: Unbreakable Kimmy Schmidt: Kimmy vs. the Reverend, Ranveer vs. Wild with Bear Grylls, You vs. Wild, and Bandersnatch, which pushed the format to dizzying complexity in 2019. ("As artificial as the construction may be, I can't help but admire, just a little bit, this spin on interactive gaming/movie-watching," wrote Aisha Harris in the New York Times.) Lessons from the ill-fated experiment? Expect them to pop up in Netflix's narrative gaming forays, rather than its conventional movie efforts, one exec told Game File. -- AN -- Physical Intelligence raises $400m. Robot AI company is backed by Jeff Bezos, OpenAI, Thrive. -- New York Times software developers strike. Hundreds dig in ahead of Election Day. -- "Infostealer" malware is rampant. Behind breaches at Ticketmaster, AT&T, Santander, EA. -- GM convinced ditching CarPlay, Android Auto is the right thing. It's about controlling the stack. -- Women see more from Kamala Harris on TikTok. Users who identify as female saw 40% more Harris campaign videos than those identified as men, according to a survey.

Study showing mass gen AI adoption overlooks critical evidence that uptake is in fact 'glacial', critic says

Hello and welcome to Eye on AI! In this newsletter...Elon Musk is reportedly raising xAI funding at a $40 billion valuation... Microsoft's GitHub Copilot goes beyond OpenAI, striking deals with Anthropic and Google... AI 'slop' floods Medium... Apple Intelligence gets mixed reviews... U.S. finalizes rules to curb AI investments in China A research paper has been making the rounds on social media that claims 40% of U.S. adults have used generative AI at work or at home, evidence that people are rapidly adopting generative AI tools. Indeed, the survey found the number is outpacing PC adoption in the early 80s, when just 20% of people used the then-new computers three years after they were introduced. The paper, titled "The Rapid Adoption of Generative AI" and published last month by the National Bureau of Economic Research, comes from researchers at the Federal Reserve Bank of St. Louis, Vanderbilt University, and Harvard Kennedy School. It was based on responses from just over 5,000 people said to give a representative sample of the entire U.S. population. But Arvind Narayanan, a Princeton University computer science professor and coauthor of the recently-released AI Snake Oil, called the paper a "hype case study" in a post on social media service X. The 40% figure, he pointed out, would include someone who simply asked ChatGPT to write a limerick once in the last month. The paper actually said only 0.5%-3.5% of work hours involved generative AI assistance -- and only 24% of workers used it once in the last week prior to being surveyed, and only one in nine used it every workday. "Compared to what AI boosters were predicting after ChatGPT was released, this is a glacial pace of adoption," Narayanan wrote, adding that people spending thousands of dollars on early PCs were not just using them once a month. Based on what I see in my own world, I agree with the more tempered view of AI adoption. Among the people I know, most are not using generative AI tools at all. Many don't even know what I'm talking about if I mention a tool other than ChatGPT. In fact, my husband is actually among the few who would fall under the "super user" category that the Washington Post reported on yesterday. They're defined as people who use tools such as ChatGPT, Google's Gemini, or Anthropic's Claude regularly to learn new skills, create reports, analyze data, and research topics. But I believe this will change relatively quickly. That's because it's becoming increasingly impossible for consumers to avoid generative AI text, image, audio, and video tools. If you use Google, you're seeing AI Overviews with every search. Meanwhile, for months, Google Docs has been prompting me to use its AI assistant. Apple Intelligence just arrived for the iPhone, while Microsoft Copilot is in everything from Word to Excel. As for Meta, its AI assistant is inescapable on Facebook, Instagram, and WhatsApp. Every day, consumers are breaking down and giving one of these tools a try and, intrigued by the results, might try again for another task. I'll give you an example: One of the most common questions people ask me is whether I use generative AI tools. The answer is yes, but so far I have only found them helpful for certain specific tasks. For example, headlines for articles are always hard to write, and sometimes I just want some feedback. For a long time, I would paste a tentative headline into ChatGPT and ask the AI to suggest a few other options and then I would edit from there. The results were okay, but not always great. Finally, a few months ago I tried a new tack: I asked ChatGPT "What do you think of this headline?" The results are always really satisfying. For this essay, for example, I asked ChatGPT what it thought of the following: "Most people you know aren't using generative AI. But nearly all are still experiencing it." ChatGPT responded with thoughts about what was working in my headline (contrast, intrigue, and relatability) as well as suggestions for refinement (something punchier with improved flow). It offered a few other options that I didn't love, so I went back and forth with it a few times. Each time it got me closer to what I finally published. Whether generative AI adoption reaches the rapid, mass adoption that companies and investors predicted remains to be seen. Some tools have been criticized as "half-baked" (like the New York Times said about Apple Intelligence's new features yesterday). There will be plenty of generative AI products that bite the dust because users did not find them, well, useful. But good luck avoiding generative AI right now -- it's already everywhere, whether you are using it yet or not. Companies are crossing their fingers that you'll give them a try, again and again. Request your invitation for the Fortune Global Forum in New York City on Nov. 11-12. Speakers include Honeywell CEO Vimal Kapur and Lumen CEO Kate Johnson who will be discussing AI's impact on work and the workforce. Qualtrics CEO Zig Serafin and Eric Kutcher, McKinsey's senior partner and North America chair, will be discussing how businesses can build the data pipelines and infrastructure they need to compete in the age of AI. Elon Musk's xAI is reportedly seeking funding at a $40 billion valuation. Just a few months ago, the startup was valued at $24 billion, when it raised $6 billion. A source said xAI hopes to raise several billion dollars in the new funding round -- though the talks are in the very early stages, according to the Wall Street Journal. Microsoft's GitHub Copilot goes beyond OpenAI. There's another sign that Microsoft's relationship with OpenAI, in which it's a big investor, is becoming less exclusive. GitHub Copilot, one of Microsoft's most successful AI products that is used for code generation, will now give developers the choice of using models from Anthropic and Google, rather than only OpenAI's models. "There is no one model to rule every scenario, and developers expect the agency to build with the models that work best for them," GitHub CEO Thomas Dohmke said in a blog post. AI 'slop' Is flooding Medium. "Slop" has become the term of choice to describe low-quality AI-generated content and images. Think clickbait articles, blog posts stuffed with keywords, and photos of people with six fingers. According to Wired, this kind of slop has flooded the blogging platform Medium, far more than on other websites. An analysis by AI detection startup Pangram Labs, commissioned by Wired, found that 47% of nearly 275,000 Medium posts sampled over a six-week period were likely AI-generated. "This is a couple orders of magnitude more than what I see on the rest of the internet," said Pangram CEO Max Spero. Apple Intelligence is here, but people want it to be smarter. Most publications reporting on Apple Intelligence's new iPhone features had a similar reaction to the Verge, which said "like most AI on smartphones so far, it's mostly underwhelming." I have a Samsung Galaxy, so I couldn't test Apple Intelligence myself, but it's clear that more features are coming down the pike: In yesterday's launch announcement, Apple teased features that haven't debuted yet, including using Siri to take action in apps. For now, users can take advantage of features including AI summarization, email tweaking, and call transcription. U.S. finalizes rules to curb AI investments in China. The Biden administration announced yesterday that it's finalizing rules to restrict U.S. investments in China's AI and other tech sectors deemed potential threats to national security, according to Reuters. Initially proposed by the U.S. Treasury in June, the rules follow an executive order signed by President Joe Biden in August 2023, targeting three critical areas: semiconductors and microelectronics, quantum information technologies, and specific AI systems. Exclusive: An 'AI coworkers' startup just raised a $8.7 million seed round led by General Catalyst -- by Sheryl Estrada Citi moves vital infrastructure to Google Cloud as part of broader AI push -- by Michael Del Castillo Salesforce CEO Marc Benioff claims Microsoft has done a 'tremendous disservice' to the AI industry -- by Chloe Berger Nvidia's billionaire CEO says AI can do a lot of things -- except take his job -- by Chloe Berger OpenAI suffers departure of yet another AI safety expert, and fresh claims around copyright infringement -- by David Meyer Dec. 8-12: Neural Information Processing Systems (Neurips) 2024, Vancouver, British Columbia Dec. 9-10: Fortune Brainstorm AI, San Francisco (register here) Potential for LLMs for less-used languages. In February, Cohere for AI, the nonprofit research lab established by Cohere in 2022, unveiled Aya, an open-source large language model supporting 101 languages. That's more than twice the number of languages covered by existing open-source models. The researchers also released the Aya dataset, a corresponding collection of human annotations from a community of "language ambassadors" worldwide. This is important because one of the obstacles with less common languages is that there is less source material to train AI models on. Last week, the organization released Aya Expanse, another family of models meant to help researchers help close the AI language gap. "We are a small lab, and this builds on years of dedicated research to connect the world with language," said Sara Hooker, a former Google Brain researcher who has led Cohere for AI since it was founded. Could an AI model predict a "gray swan" weather event like a Category 5 tropical cyclone? AI models have been used for years to improve weather and climate predictions. But could an AI model predict a "gray swan" extreme weather event, like a Category 5 tropical cyclone? These events are possible, but so rare that they are often absent from the training dataset. That was the challenge researchers from the University of Chicago wanted to take on in a recent paper: They trained separate versions of an AI model with all data included, and another with Category 3-5 tropical cyclones removed. The researchers wanted to know whether the AI model with the cyclones removed could extrapolate from weaker weather events present in the training set to stronger, unseen weather extremes. Unfortunately, the answer was no. "Our work demonstrates that novel learning strategies are needed for AI weather/climate models to provide early warning or estimated statistics for the rarest, most impactful weather extremes," the researchers wrote.

Data Sheet: A classic business principle

Good morning. If you thought Larry Ellison was going to run Paramount, think again. A month ago, Skydance Media told the FCC that the Oracle founder would have voting control of the media company formerly known as ViacomCBS when Skydance closes its $8 billion deal. It has now filed a revision saying no, it's actually son David who calls the shots. Why the confusion? David, 41, was always intended as CEO of the combined company, should regulators approve their merger, but companies controlled by his father, 80, are contributing the bulk of the cash necessary to get the deal done. Fans of corporate governance (and Succession) can agree: Clarity beats screaming "I'm the eldest boy" any day of the week. -- Andrew Nusca P.S. My editor asked me to tell you that Target CEO Brian Cornell, who I interviewed at Fortune Brainstorm Tech way back in 2017, is speaking at Fortune Global Forum in NYC next month. Interested? Request an invitation here. A longtime advisor to Jeff Bezos at Amazon believes the last-minute decision by leadership at the Washington Post, which Bezos has owned since 2013, to abandon presidential endorsements reflected a classic Bezos business principle about protecting your self-interest by avoiding unnecessary mistakes. Bezos has said that canceling the Post's planned endorsement of Kamala Harris was a "principled decision" taken to improve public trust in the media. Former Amazon public relations chief Craig Berman says Bezos' interest in space company Blue Origin, his competition with Elon Musk, and fear of the consequences should Trump win the White House are more likely his motives. "I would characterize it as a very convenient principle," Berman told Fortune. Blue Origin still needs revenue to operate and government contracts are a major source of those for the rocket company. Blue Origin has a $3.4 billion NASA deal to build a lunar "lander" and competes with Musk's SpaceX not only for government deals, but for talent and launch pads as well. The problem for Bezos is that Musk has recently remade himself as a staunch Trump ally. "It's personal; it's his baby," Berman says of Bezos and Blue Origin. "That's the one that matters to him right now." -- Jason Del Rey The Mac now comes in its mini-est Mini form yet. Apple has given the Mac Mini its biggest revamp in over a decade, drastically shrinking its footprint to just 5x5 inches. Gone are the USB-A ports, but there are five USB-C ports, three of which handle Thunderbolt, plus Ethernet, HDMI, and the usual audio jack. The Mini starts at $599 and comes with up to 8TB in storage. This is part of Apple's big week of fall announcements, which so far generally involve adding the new M4 generation of processors (this and the new iMac) or switching from the old proprietary Lightning standard to USB-C (the new Magic Mouse, Keyboard, and Trackpad.) New MacBook Pros are also rumored to be on the menu. Ensuring an even rollout of the latest Apple processors could help the company show off what its new, and so far unenthusiastically received, Apple Intelligence can do. -- David Meyer Google parent company Alphabet reported its quarterly results on Tuesday. Compared to the same period a year ago, total revenue was up 15% to more than $88 billion, search revenue was up 12% to more than $49 billion, and profit was up 34% to $26 billion. Cloud revenue and YouTube ad revenue were also up by double-digit percentages. But the most interesting insight came from the company's subsequent earnings call. "More than a quarter of all new code at Google is generated by AI, then reviewed and accepted by engineers," Google CEO Sundar Pichai said. "This helps our engineers do more and move faster." Google has understandably worked AI into what feels like every corner of its product offerings, from its core search business to its Android mobile devices to enterprise cloud services. (And yes, even Waymo robotaxis.) But how AI is changing work at the company itself, with its 180,000 employees on six continents? That's a far more tangible case study. -- AN A new Reuters report quotes an array of people in and around Intel who say that CEO Pat Gelsinger has committed unforced errors en route to reviving the Silicon Valley icon. The chief executive's public comments about Taiwan's precarious relationship with China reportedly led an offended TSMC, the world's largest pure-play chipmaker, to revoke a discount with its founder calling him "discourteous." What's more, Gelsinger's public projections also outpaced internal realities, according to dozens of current and former employees interviewed for the story, plus documents. ("We have made immense progress," Intel's response to the article reads in part, "and we're going to finish the job.") There's little doubt Gelsinger inherited a challenge in 2021; rivals had already passed Intel in making leading chips for phones and AI. The question now is whether he's "compounded those problems," as one source in the story puts it, in his turnaround bid. Wall Street analysts expect Intel this year to record its first annual net loss since 1986. -- AN With all the chatter about Microsoft-backed OpenAI, it's hard not to wonder what's going on at the rival AI organization started by co-founder Elon Musk. Now we know. A new Wall Street Journal report says the not-quite-two-years-old company is talking to investors to raise money at a $40 billion valuation. To put that in perspective, that's about a third of OpenAI's latest valuation...but a hair short of the market capitalization of Ford Motor Co. and Kraft Heinz. Whether you believe in the valuation or not, there's no question the company needs the cash. The pursuit of so-called "frontier AI" -- considered the most advanced, cutting-edge models that push the boundaries of what artificial intelligence can achieve -- requires extensive and costly computational infrastructure. Here's some back-of-the-napkin math for context: xAI's Memphis datacenter wants to scale from 100,000 to 200,000 Nvidia GPUs. The chipmaker's flagship model costs about $30,000 each. Presto, change-o: $3 billion.) That xAI is younger than its more established rivals means it's keen to spend to catch up. Its previous funding round, announced a mere five months ago, raised $6 billion at a $24 billion valuation from investors including Valor Equity Partners, Vy Capital, Andreessen Horowitz, Sequoia Capital, Fidelity, and Prince Alwaleed Bin Talal. -- AN -- OpenAI is building its first in-house chip, employing Broadcom and TSMC to do so.

Data Sheet: Hit 'em where it hurts

Good morning, and welcome to November. On this day five years ago, Google announced that it would buy Fitbit for $2.1 billion. It surely feels like ages ago that Fitbit was the king of wearable tech. The underwhelming launch of the Apple Watch failed to break Fitbit's grip...but then it did. Today, watches are in, bands are out, and Google is an also-ran in the category...even if its products are slick. As for me? I exchanged my fitness band for a mechanical watch that very same year. One fewer thing to remember to plug in. -- Andrew Nusca P.S. My editor wants you to know that Josh Kushner, founder of the venture firm Thrive Capital (Airtable, Anduril, OpenAI, Plaid, Stripe), will speak at the upcoming Fortune Global Forum, Nov. 11-12 in NYC. Request an invitation here. Amazon's quarterly financial results on Thursday surpassed analyst expectations, powered by strong revenue growth and record operating income. But it was Amazon Web Services, the company's $110 billion cloud computing business, that really stood out. AWS operating income grew 50% year-over-year to $10.4 billion. Meanwhile revenue in the unit rose 19% from the same period a year earlier to $27.5 billion, in line with expectations. The division's operating profit margin was 38%, accelerating from 30% in the same period last year. The robust profits came even as Amazon, like its Big Tech peers, invests heavily on its own AI consumer products while also expanding its offerings of AI services and building blocks to corporate customers. One unknown is how those profits will look as AWS' multibillion-dollar generative AI business -- in which sales are growing more than 100% year over year -- develops into a larger piece of the overall Amazon Web Services business. "There are going to be very healthy margins here" over time, Amazon CEO Andy Jassy said during a conference call. In the meantime Amazon continues to ramp up capital spending, with a particular focus on expanding its datacenter footprint. Increased automation and robotics investments also play a role. Said Jassy of the company's continued drive to automate warehouses: "We really do believe that AI is going to be a big piece of what we do in the robotics network." -- Jason Del Rey The rise of gen AI made it clear that the days of conventional online search were numbered. OpenAI's latest announcement is perhaps the greatest salvo to date. The San Francisco AI company said Thursday that it would add a new set of search features to its flagship ChatGPT product. Dubbed -- naturally -- ChatGPT Search, the capabilities allow users to search for timely information and receive it with attribution to news publishers and other sources. The new capabilities use OpenAI's 4o model and are available to paid ChatGPT Plus and Team users for mobile and web. The company says its enterprise and educational customers, then free users, are expected to follow in the coming weeks and months, respectively. About 250 million people use ChatGPT each week. How are existing search giants Google and Microsoft (a leading OpenAI investor) confronting the existential threat? By incorporating conversational AI into their search products, of course. Game very much on. -- AN Apple on Thursday reported a 6% revenue bump in its fiscal fourth quarter as sales of iPhones rebounded. The company posted growth in all of its business units save for the category that includes wearable devices. Total sales in the three months ended Sept. 28 increased year over year to $94.9 billion, coming in slightly above Wall Street estimates of $94.6 billon. And the iPhone? Sales of Apple's all-important product grew by 5.4% to $46.2 billion, reversing two consecutive quarters of decline. (Apple released its new iPhone 16 in September, so only about one week's worth of sales were included in these results.) Apple CEO Tim Cook said on a conference call that iPhone sales set a September revenue record, with growth in every geographic segment. The only Apple business category that declined in the quarter was the wearables, home, and accessory group, which includes the Apple Watch, AirPods, and the costly Apple VisionPro headset. Net sales for the unit fell 3% year over year to $9 billion. -- Kali Hays China has announced sanctions on Skydio, the company said Wednesday, for selling drones to Taiwan. Its only customer in the republic is the National Fire Agency. "We're proud to support critical infrastructure operators, first responders, and allied militaries around the world," CEO Adam Bry wrote in a blog post. "We're proud to support Taiwan, and we are undeterred." Based in the San Francisco Bay Area, Skydio is the largest drone maker in the U.S. and a supplier to the militaries of both the U.S. and Ukraine. The sanctions put Skydio in a supply chain bind. Though its drones are manufactured stateside, the batteries that power them "are one of the few components we have not yet moved out of China," Bry wrote. Beijing hit 'em where it hurts: Skydio is working off a "substantial stock" of batteries and is on the hunt for alternative suppliers, but still must ration batteries to one per drone. A Financial Times report says Skydio sought help from the Biden administration; as trade tensions continue between the U.S. and China, American officials have been concerned about the disruption of supply chains as leverage. Congress is currently evaluating legislation that would ban drones from DJI, based in Shenzhen, from operating in the U.S. -- AN Intel shares jumped some 7% in after-hours trading on Thursday after the chipmaker reported earnings and guidance that exceeded Wall Street expectations. Intel reported earnings of 17 cents on revenues of $13.28 billion -- a decline of 6% from the same quarter a year prior. Unsurprisingly, given the Silicon Valley icon's extensive overhaul under CEO Pat Gelsinger, the company recognized $2.8 billion in restructuring charges and $15.9 billion in impairment charges during the quarter. Gelsinger said on a conference call that Intel expects to lower head count by 16,500 employees -- cuts were announced in August -- and reduce its real estate holdings. All of that should be a wrap by Q4 2025. Intel still finds itself with a target on its back. A possible takeover by Qualcomm or another peer -- or disruption by circling activist investors -- is hardly off the table as the company works to keep its head above water and its share price up. But managing expenses will only get it so far. Intel announced its Xeon 6 server processors and Gaudi AI accelerators during the quarter; Gelsinger said on the call that Gaudi uptake has been slower than expected. -- AN -- Microsoft hires Facebook's former engineering chief. Jay Parikh will report to CEO Satya Nadella. -- Google's Android 16 is coming months earlier than expected: 2Q25. The better to align with device launches, my dear. -- Nvidia's Run:ai acquisition will see EU review after Italian regulators pump the brakes. -- Uber shares drop after earnings reveal that higher prices have slowed growth.

AI drives a cloud resurgence, but it's costing a lot - SiliconANGLE

Generative artificial intelligence is driving a resurgence in cloud spending, as earnings results from Amazon Web Services and Google this week both saw upside from forecasts. Even Microsoft's disappointing Azure revenue forecasts came from not having enough infrastructure to support more business, a factor Amazon CEO Andy Jassy also mentioned -- and excessive demand isn't a bad problem to have, as problems go. But at the same time, there's a big price: Spending to keep up with AI demand continues to be expensive, as Meta's results also showed this week. Meanwhile, money keeps pouring into AI companies such as Elon Musk's xAI and Bret Taylor's AI agent startup Sierra Technologies as well as more industry- and task-specific startups. Supermicro's financial situation looks so bad to its auditors that they exited stage left, tanking the server provider's stock to the tune of 33%. Its earnings report next week should be ... interesting. Indeed, next week is another big week for earnings reports, especially from chip companies such as Arm, Qualcomm, NXP Semiconductor and GlobalFoundries, as well as bellwethers such as Cloudflare and Datadog and a raft of software-as-a-service companies. TheCUBE Research analysts John Furrier and Dave Vellante discuss this and other news in more detail on this week's theCUBE Pod, out later today on YouTube. And don't miss Vellante's weekly deep dive, Breaking Analysis, arriving this weekend, for some lean-back reading. Here's what happened this week around the enterprise: On big data: At TechCrunch Disrupt, there was a illuminating session called Beyond Snowflake and Databricks: Insights from the Frontlines of Data Transformation that indicated the big-data wars will continue for a while longer. One reality, noted Jordan Tigani, co-founder and CEO of MotherDuck, which makes a cloud analytics database, is that "AI people and database people tend not to like each other very much. We're going to need new tools to power AI/ML engineers." And the open data formats that Snowflake and Databricks are ostensibly touting? "Things are moving more slowly toward these open data formats than people expect," he said. "Catalogs are a mess right now. Snowflake and Databricks are slow-walking the building of the open-source versions." As a result, he said, "the community needs to be stronger" in getting providers to become more open. "The reason people are moving to Iceberg is to remove lock-in or just lower cost," he said, but both of those reasons are "lower priorities than other things," like just getting the jobs at hand done. In fact, contended Colin Zima, co-founder and CEO of the business intelligence platform Omni, Snowflake and Databricks "will actually diverge further over time. Databricks is underinvested in all the analytics problems and Snowflake will struggle to become an AI company. It's very hard to overhaul your DNA." On what's coming next in AI: At Constellation Research's CCE 2024 conference, the firm's analysts weighed in on the next impacts of AI. VP and Principal Analyst Doug Henschen noted that only 21% of enterprises they surveyed are meeting their goals for generative AI. The problem, he said, is a lack of data, in particular clean data. "Data access and cleansing will be key in 2025." Constellation Insights Editor in Chief Larry Dignan says generative and agentic AI are starting to transform the enterprise software business model. "In the next year or so, the whole way you buy software is going to change," he said. Andy Thurai sees a lot of companies cutting corners on making sure AI is under control. "AI will produce AI to monitor itself. There's no going back if that happens," he said. "Lawyers are going to have a field day." Liz Miller noted that "the lesson we've learned over the last several years of generative AI is when you automate bad processes, you still have bad processes." She thinks "knowledge" will be the next frontier in 2025, not just big data. "Right now we're drowning in data but dying of thirst," she said. Constellation founder Ray Wang said the key question coming is at what points in AI-driven processes that humans get inserted. They all did a lightning poll on AI budgets next year and, surprisingly, most said AI budgets will be flat or worse. Did someone say trough of disillusionment? But others such as Barney Pell, venture partner at Radical Ventures, are more optimistic. "This time it's real in a way it hasn't been," he said about AI. But he sees the proper focus not on artificial general intelligence but humans plus AI -- and also neural networks plus symbolic systems, not just gen AI. The biggest issue? trust. "Most enterprise projects fail because no one trusts them," he said. One promising avenue I touched on a bit in last week's This Week in Enterprise came up again in a somewhat different way: the need for AI systems to be trained on more than just language. Denise Hold, founder and CEO of AIX Global Media, a sort of AI startup promoter, said spatial web technologies that are programmable, based on sensory information, are being developed to create a "grounding level of reality." AI adviser Cassie Kozyrkov had the quote of the week, on the difference between machine learning and AI: "If it's written in Python, it's probably machine learning. If it's in a PowerPoint, it's AI." Breaking Analysis: Generative AI adoption sets the table for AI ROI Elon Musk's xAI reportedly in talks to raise new funding on $40B valuation Generative AI agent startup Sierra Technologies raises $175M at $4.5B valuation Read AI lands $50M in new funding as it strives to become everyone's AI copilot AI contact center provider Regal closes $40M round to scale personalized customer interactions Zenity raises $38M to secure enterprise AI copilots and low-code app development Brightwave raises $15M to accelerate financial research with AI Diffblue raises $6.3M for autonomous AI-powered test generation platform OSI clarifies what makes AI systems open-source, but most 'open' models fall short White House limits AI and chip investments to China over national security concerns Nvidia needs EU approval to buy AI startup Run:ai, regulators say OpenAI adds LLM-powered search engine to ChatGPT Apple releases first batch of Apple Intelligence features, debuts new iMac Alexa AI upgrade delayed to 2025 as Amazon hits major roadblocks GitHub Copilot goes multimodel, adding support for Google's Gemini and Anthropic's Claude LLMs Salesforce launches Agentforce AI agent platform into general availability Nvidia hands out blueprints for the creation of scalable 'AI factories' Google reportedly developing new AI that can automate web browsing tasks in Chrome Google brings grounding with search to Gemini in AI Studio and API Gemini in Android Studio rolls out more AI-powered development features Meta reportedly developing custom search engine for its Meta AI chatbot Amazon Q introduces AI-powered inline chat directly in the code editor for developers Dremio throws its support to Polaris data catalog and expands deployment options for Iceberg lakehouse Pegasystems moves the needle on generative AI development with Pega Infinity updates Creatio combines multiple AI types in new release of its no-code platform Moveworks taps agentic AI to aid in enterprise search Patronus AI debuts API for equipping AI workloads with reliability guardrails TigerEye's new AI Analyst offers enhanced decision-making in sales and revenue operations Securiti launches Genstack AI to enhance secure enterprise generative AI adoption Tabnine unveils AI developer 'coach' with personalized code review agent There's more AI and big data news on SiliconANGLE Alphabet's stock jumps as cloud fuels strong revenue growth Also: Russia hits Google with astronomical fine over YouTube bans That's a lotta rubles. Microsoft's AI bet pays off as Azure revenue grows, but stock falls on infrastructure supplier delays Cloud growth helps Amazon to another earnings beat, and its stock jumps Meta's stock heads south on slow user growth and ongoing infrastructure investments Intel surprises Wall Street with solid earnings and revenue, but success in AI remains elusive Apple beats Wall Street's targets as iPhone 16 sales get off to a solid start AMD's guidance leaves investors wanting more, and its stock tumbles Sluggish chip sales curtail Samsung's profits, worrying investors Uber stock tumbles despite earnings beat Extreme Networks shares soar on lower-than-expected sales decline Informatica lags Q3 earnings forecasts Data streaming leader Confluent posts strong earnings, boosts year-end forecast Cyber split: Check Point shares down, Commvault up on quarterly earnings F5 shares rise over 10% following earnings beat in latest quarter Atlassian shares surge as subscription revenue drives strong earnings beat Twilio shares surge over 10% on strong earnings and raised forecast Coinbase and Robinhood miss earnings estimates and shares drop Supermicro shares plummet 31% after auditor resigns Wall Street giants to make $50B bet on AI and power projects (per WSJ) OpenAI reportedly working with Broadcom and TSMC to develop custom AI inference chip Siemens to buy Altair in $10.6B equity deal (per WSJ) Israeli fintech company Melio hits $2B valuation with $150M raised for B2B payment expansion Dropbox to let go 20% of its workforce amid growth push GPU cloud operator GMI Cloud secures $82M investment Google Cloud turbocharges its AI Hypercomputer stack with next-gen TPUs and bigger GPU clusters Cisco debuts new Nvidia-powered data center systems for AI workloads Cybersecurity startup Armis closes $200M round at $4.2B valuation Filigran raises $35M to help companies simulate and respond to cyberattacks Israeli AI security startup Noma launches with $32M to secure the 'data and AI lifecycle' Zimperium warns of sophisticated 'vishing' tactics in new FakeCall malware variant Phish 'n' Ships: Human Security warns of fake shops exploiting payment platforms and SEO New Akamai Account Protector features target advanced fraud detection across user accounts NTT DATA expands partnership with Palo Alto Networks to launch AI-powered cybersecurity service Extreme Networks expands Universal ZTNA with enhance shadow IT and policy management tools More cybersecurity news here Data privacy-focused Nillion network raises $25M to expand decentralized solutions Blockchain firm Axal nabs $2.5M to build verifiable autonomous agent network And check out more news on emerging tech, blockchain and crypto and policy Matt Wood, Amazon's high-profile former vice president of artificial intelligence, is now PricewaterhouseCoopers' first commercial technology and innovation officer. Tuan Tran is stepping away as president of HP's printing and imaging business to lead a cross-company AI strategy. Taking over the printing business is 30-year HP veteran Anneliese Olson. Arjun Sethi joined as co-CEO of Kraken alongside David Ripley as the crypto exchange announced an unspecified number of layoffs. Decrypt also reported that Ethereum software giant Consensys Tuesday announced it had laid off 20% of its global workforce, or 163 employees. Later the same day, decentralized exchange dYdX said it cut 35% of its staff. All this comes as the price of bitcoin keeps soaring.

Data Sheet: It'll all be over soon

Good morning. Happy Halloween to those who celebrate. If you experience the world through a soundtrack like I do, I highly recommend The Cramps' "I Was a Teenage Werewolf," the Ramones' "Pet Sematary," and Misfits' "Dig Up Her Bones." (Not your thing? Don't worry, it'll all be over soon.) I could use this space to earnestly list the horrifying things that happened in tech this year, but I won't. Instead, I'll direct you to this subreddit of weird, and frequently unsettling, images generated by OpenAI's Dall-E model. You're welcome. -- Andrew Nusca P.S. You'll want to hear what Siemens USA CEO Barbara Humpton tells me about the state of industrial AI at Fortune Global Forum. It's Nov. 10-11 in NYC. Request an invitation here. More than 500 people were sent packing at cloud storage provider Dropbox this week as the San Francisco company transitions to...well, perhaps not greatness, but certainly discipline. In a memo, co-founder and CEO Drew Houston had this to say: "I take full responsibility for this decision and the circumstances that led to it, and I'm truly sorry to those impacted by this change." He added that the company made deeper cuts "in areas we're over-invested or underperforming." Dropbox is in a tough spot. Per Houston, it's seeing "softening demand and macro headwinds" in its core file sync and share (FSS) business, which competes with Box, Google Drive, Microsoft OneDrive, and virtually anywhere you've saved something in the cloud. In its most recent quarter, Dropbox saw its lowest rate of revenue growth (1.9%, to $634.5 million) in history; at $26, its share price is virtually unchanged from this time last year and is a couple of dollars short of where it closed on the day of its 2018 IPO. -- AN As the U.S. presidential election gets underway -- millions of Americans have already voted -- a new Washington Post report says Amazon CEO Andy Jassy sought "to establish a friendlier rapport" with Republican nominee Donald Trump and outlined Amazon's "plans for the future." The tidbit is part of a broader article on the financial interests of Washington Post owner Jeff Bezos, who endured a wave of criticism (and subscription cancellations) after his intervention over the newspaper's planned presidential endorsement. Bezos stepped down as Amazon's CEO in 2021, but he remains the company's largest shareholder and executive chair of its board of directors. Though it's hardly strange for captains of industry to ingratiate themselves with the nominees of both political parties ahead of Election Day -- after all, polls put Harris and Trump in a dead heat -- it's still notable. For one, Trump's criticism of Amazon (and Bezos) over the years has been withering, even if the administration was personally lucrative for Amazon's founder. It was also recently reported that Bezos' interest in securing government contracts for his space firm Blue Origin might be influencing his actions elsewhere. Trump reportedly requested a contribution from Jassy, but he demurred. Jassy, for his part, has only personally donated to the reelection efforts of two Democrats, according to the FEC: Maria Cantwell of Washington and Tim Kaine of Virginia. -- AN No company wants to be associated with drama. But internal controls, board independence, and accounting practices drama? Oh hell no. Supermicro's shares dropped by a third, from $49 to $33, on Wednesday after the Silicon Valley IT company disclosed that its auditor, Ernst & Young, had resigned, stating it was "unwilling to be associated with the financial statements prepared by management." EY also raised concerns about the independence of the company's board from CEO Charles Liang and others. Supermicro, which sells servers to other businesses (and has a longstanding relationship with Nvidia and its CEO, Jensen Huang), has yet to issue financial statements for the year and is reportedly under federal investigation. EY reportedly flagged issues to Supermicro this summer. In response, it created a special board committee to look into the company's internal controls. The auditor wasn't the only one to raise concerns. A short seller report later criticized Supermicro's financial controls and accounting practices, something for which the company paid a multimillion-dollar SEC penalty in 2020. -- AN There are fines, and then there are fines. A Moscow court has ruled that Google must pay Russian TV stations roughly $20 decillion because YouTube banned their sanctioned channels. That's about 20-to-the-power-of-20 times current global GDP -- the result of a daily fine that was doubled each week since late 2021, when a court first ordered Google to unblock one of the channels, Tsargrad TV. Google blocked more channels after Russia's full-scale invasion of Ukraine in early 2022, but then the company stopped operating in Russia altogether, technically going bankrupt there. So, even if Google's measly $88 billion quarterly revenue (as reported this week) isn't anywhere near enough to allow it to pay what Moscow is demanding, it doesn't really have to. As Alphabet noted in a recent filing, it does not believe its "ongoing legal matters relating to Russia... will have a material adverse effect." -- David Meyer By now, the biggest players in the video game industry have mostly unwound overly optimistic investments made during the COVID-19 pandemic. But that doesn't mean the bloodletting is over, at least at Sony. The latest casualties: Firewalk, the studio behind the first-person-shooter flop Concord (which was pulled two weeks after its launch in August), and Neon Koi, a mobile gaming studio that hadn't yet released anything. About 200 people worked between them, according to Bloomberg. At the corporate level, Sony is actually doing just fine. After releasing its latest quarterly results in August, the company brightened its full-year outlook for total sales (to about $80 billion) as well as those in its so-called game and network services segment. Within that business, first-party game sales and network services were up, and hardware sales -- which includes the PS5 console -- were down. -- AN -- Microsoft earnings. Cloud revenue drives an estimate-beating quarter. -- Meta earnings. Ad revenue drives an estimate-beating quarter; AI investment tempers outlook. -- eBay earnings. Weak consumer demand in the category leads to an estimate miss. -- Roku earnings. Its first billion-dollar quarter comes with news that it will no longer share how many households use its streaming service.