 |
|
|
|||||
Welcome to AI Decoded, Fast Companys weekly newsletter that breaks down the most important news in the world of AI. You can sign up to receive this newsletter every week via email here. Anthropics stance on autonomous weapons may not survive the future Much of the AI world is watching closely as Anthropic tangles with the Pentagon over how the government can use the Claude models. Anthropic has a $200 million contract with the Pentagon, but the contract says the military cant use the AI companys models as the brains for autonomous weapons or for mass surveillance of Americans. Defense Secretary Pete Hegseth insists, after the fact, that the military should be able to use the Anthropic models for all lawful purposes. Hegseth summoned Anthropic CEO Dario Amodei to the Pentagon for a Tuesday morning meeting, in which he reportedly gave Anthropic until 5:01 p.m. Friday to comply with the Pentagons demand. If Anthropic fails to do so, Hegseth threatened to invoke the Defense Production Act to compel the AI company to supply its models with no guardrails. Hegseth also said the government would declare Anthropic models to be a supply chain risk, meaning that all government suppliers would be directed to avoid or discontinue use of Anthropic models. Amodei said in an interview after the Hegseth meeting that his company has no intention of complying with Hegseths demands. (Hes got a strong case: After all, government officials agreed to the terms.) Amodei explained that the military relies on human judgement to avoid violating peoples constitutional rights. If AI is making the decisions, there will be no human being to object. Amodei is right, and his companys willingness to stand up for its values is laudable. The trouble is, were rapidly heading for a future where autonomous systems become the norm in warfare. For years, the defense establishment talked about keeping the human in the loop in AI weapons systems. Often that human is a government lawyer who can make calls on rules-of-engagement issues on the battlefield. Today the Pentagon is talking more about fully autonomous weapons that can manage more of the kill chain, or the series of communications and decisions around the destruction of a target. Military leaders often say that whoever can use technology to shorten the kill-chain will win wars. Things like electronic warfare (cyberwar), hypersonic missiles, and drone swarms are making war faster and response times shorter. This may eventually preclude the opportunity for human review and decision-making. Increasingly, the U.S. military may be forced to take humans out of the loop in order to stay competitive with its adversaries. So the result of Anthropics standoff with the Pentagon may be that a safety-conscious AI lab is forced out, and a generally less scrupulous company like xAI is chosen as the alternative. Trump rips off Mark Kellys idea for powering new data centers In his State of the Union address, Donald Trump spent a few minutes on the subject of new data centers for AI, which has over the past few months become a hot button issue for voters. While the tech industry says it needs hundreds of new data centers to support all the AI it’s building, a growing number of voters now understands that the power grid improvements needed to power the data centers may increase their energy bills. I have negotiated the new Ratepayer Protection Pledge, Trump crowed. We’re telling the major tech companies that they have the obligation to provide for their own power needs. Politicos might recognize that message, as it closely echoes what Arizona Senator Mark Kelly, a Democrat, has been saying for months now. Kellys AI for America plan would create an industry-financed AI Horizon Fund to pay for energy-grid upgrades and workforce reskilling. According to Kellys plan, Congress could require data center developers to buy or lease enough land to contain both their facilities and the renewable energy infrastructure to power and cool them. The data center operators could also be required to pay to connect the renewable sources to the local grid, should the power they generate go unutilized. Trumps idea is more of a suggestion. As of now its non-binding, just words. And there was no mention of how the tech companies would generate their own power. Elon Musks xAI, for example, brought its own power to its massive Colossus data center in Memphis. Unfortunately, they were dirty methane-powered turbines, and the facility quickly became one of the areas biggest polluters. High numbers of young tech job seekers AI-cheated on skills tests Cheating on technical hiring assessments went through the roof in 2025, with fraud attempts more than doubling, according to new research from CodeSignal, which runs a developer-skills evaluation platform used in hiring software engineers. The research found that 35% of proctored assessments showed signs of cheating or fraud last year, up from just 16% in 2024. The biggest culprits? Plagiarism, having someone else take the test for you, and sneaking in AI tools that aren’t allowed. The jump was especially noticeable among entry-level candidates. Fraud rates for junior roles nearly tripled year over yeargoing from 15% to 40%making early-career hiring a particularly vulnerable spot in the recruiting pipeline. In a press release accompanying the report, CodeSignal CEO and cofounder Tigran Sloyan partly blamed the normalization of AI tools, noting that 80% of Gen Z reportedly uses AI in daily life, which has made the line between acceptable help and outright cheating much blurrier. Accessibility to AI also makes unauthorized assistance harder to detect and raises the stakes for maintaining fair and reliable skill evaluation, he noted. CodeSignal’s detection systemswhich combine AI analysis, human review, and digital monitoringidentified a few common patterns across flagged assessments. About 35% of candidates frequently looked off-screen, suggesting they were consulting outside resources during the test. Another 23% showed unusually linear typing patterns, where complex solutions just appeared with barely any pauses or debugging. And 15% had answers that looked a lot like known solutions or leaked content. (It’s worth noting that these numbers reflect attempts that were actually caught, not cases where someone successfully slipped through.) The data also surfaced some geographic and procedural gaps. Fraud attempt rates hit 48% in the Asia-Pacific region, compared to 27% in North America. Testing conditions made a big difference, too: Candidates in unproctored environments shoed score jumps more than four times larger than those being actively monitored, which pretty clearly shows that proctoring works as a deterrent. As for how CodeSignal catches all this: the company says it’s spent a decade building out its fraud-prevention infrastructure, which it’s now applied across millions of assessments. It uses a proprietary “Suspicion Score” and leak-resistant test design to flag things like plagiarism, proxy test-taking, unauthorized AI use, and identity fraud. More AI coverage from Fast Company: Harvard study shows AI stock trading rivals many picks made by fund managers He built a hit podcast about the Epstein files. Its entirely AI-generated What if the SaaSpocalypse is a myth? This AI note-taking startup thinks its building the steering wheel for chatbots Want exclusive reporting and trend analysis on technology, business innovation, future of work, and design? Sign up for Fast Company Premium.
Category:
E-Commerce
Artificial intelligence chipmaker Nvidia on Wednesday announced another quarter of astounding quarterly growth as investors try to decipher whether technology’s latest craze is overblown hyperbole or a springboard into a new era of prosperity and productivity.The results for the November-January period blew past the analyst projections that shape investors’ perceptions, as has been the case since Nvidia’s high-end chips emerged as AI’s best building blocks three years ago.Nvidia’s fiscal fourth-quarter revenue surged 73% from the previous year to $68.1 billion while its profit nearly doubled to roughly $43 billion, or $1.76 per share.“No quarter has had more riding on it than this one,” said Jake Behan, head of capital markets for the investment firm Direxion. “The AI trade needed some positive news and Nvidia’s earnings report brought plenty of it.”The Santa Clara, California, company also provided a forecast exceeding analyst projections while its CEO Jensen Huang reinforced the demand for the company’s chips is still “skyrocketing.” That description feeds into Huang’s thesis that the AI boom is still in the early stages of a buildout that will reshape society. If Nvidia hits its revenue target for the February-April period, it will translate into a 77% increase from last year a sign that the company’s already phenomenal growth rate is still accelerating.“AI is here, AI is not going to go back,” Huang said during a conference call with analysts. “AI is only going to only get better from here.”Despite the stellar results and still-rosy outlook, many investors still evidently are worried about a jarring comedown after a three-year boom that has seen Nvidia’s market value soar from $400 billion at the end of 2022 to nearly $4.8 trillion now. After initially rising 4% in extended trading after the latest quarterly numbers came out, Nvidia’s stock price backtracked and was slightly down following Huang’s upbeat conference call.Nvidia has regularly cleared the bar set by analysts in the past three years, often by a wide margin, but that hasn’t always been enough to satisfy investors who have become increasingly skeptical about whether AI will justify the trillions of dollars that are being spent to develop the technology.After Nvidia delivered a stellar performance that far exceeded analyst forecasts in its last quarterly report, its stock price still fell by 3% during the next day’s trading.The AI fervor has escalated again during the past month as the four companies leading the AI charge Amazon, Microsoft, Google parent Alphabet and Facebook parent Meta Platforms collectively made commitments to spend about $650 billion this year ramping up their AI computing power.A significant amount of the money is expected to be earmarked to buy more Nvidia chips required to power their AI factories, just as has been the case for much of the past three years as Nvidia’s annual revenue soared from $27 billion to $216 billion. Analysts expect the chipmaker’s revenue to surpass $330 billion during the company’s next fiscal year, a more than 50% increase from the past year.“We want to take the great opportunity that we have as we’re in the beginning of this new computing era, this new computing platform shift, to put everybody on Nvidia,” Huang said. Michael Liedtke, AP Technology Writer
Category:
E-Commerce
There are few things that unite the world like animal videos. There are also few things that are so readily commoditized. Both have occurred in the case of Punch, a baby monkey at the Ichikawa City Zoo in Japan. Punch captured hearts around the world after a viral post showed him hugging a stuffed orangutan toy after being rejected by other monkeys. E-commerce sellers act quickly with monkey merch Now, the young Japanese macaque and his stuffed friend are available as everything from toys on Etsy to adecide for yourself if its AIchildren’s book on Amazon. Theres also an official Punch Monkey store with products like stickers, shirts, and mugs. Some of the merchandise even contains hopeful sayings, like Small, but brave, alongside imagery of the pair. In fact, the original plush orangutan doll is available for $19.99, as its one of the Djungelskog soft toys from Ikea. The Swedish retailer has gone so far as to make an advertisement based on Punch and shared to its social channels. In it, a stuffed monkey holds the orangutan while real monkeys appear in the background. The copy reads, Sometimes, family is who we find along the way. It then refers to the stuffed toy as Punchs comfort orangutan. View this post on Instagram Fast Company has reached out to Ikea for more information on the retailer’s orangutan soft toy sales. We will update this post if we hear back. Meanwhile, a new video appears to show Punch having made some progress with his fellow monkeys. But the young creature has already reached the same status as its fellow infamous animals like Moo Deng, the pygmy hippo.
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||