📰 最新动态
1. 12 Graphs That Explain the State of AI in 2026 - IEEE Spectrum
The capabilities of leading AI models continue to accelerate and the largest AI companies, including OpenAI and Anthropic, are hurtling toward IPOs later this year. ## US companies lead in AI models. The United States has led the charge in AI model releases over the past decade, and that remains as true in 2025 as in any year prior. While U.S. companies released the largest number of notable AI models, China has an equally clear lead in the deployment of robotics. This graph, which uses the compute power of Nvidia’s H100e as a yardstick, shows that the world’s AI compute capacity has increased more than three-fold every year since 2022. Emissions from AI inference also continue to increase, though results again vary by model. The 2025 Stanford AI Index reported the top-ranking model, OpenAI’s o1, correctly answered just 8.8 percent of questions. As the graph above shows, the number of publications on the topic of AI use for drug discovery has more than double over the past two years.
📰 来源：Tavily Search | 🕒 Mon, 13 Apr 2026 13:02:31 GMT
2. Anthropic’s rise is giving some OpenAI investors second thoughts - TechCrunch
# Anthropic’s rise is giving some OpenAI investors second thoughts. OpenAI’s $852 billion valuation is facing skepticism from some of its own investors as the company scrambles to reorient itself around enterprise customers and fend off Anthropic, according to the Financial Times. One investor who has backed both companies told the FT that justifying OpenAI’s round required assuming an IPO valuation of $1.2 trillion or more — making Anthropic’s current $380 billion valuation look like the relative bargain. Iconiq Capital partner Roy Luo — whose firm has invested over $1 billion in Anthropic while holding a smaller stake in OpenAI — told the FT where he stood. “We picked.” OpenAI CFO Sarah Friar pushed back, telling the FT that the company’s $122 billion raise — the largest private fundraising in history — was evidence of continued investor confidence. AI, Anthropic, iconiq capital, OpenAI. ### Anthropic’s rise is giving some OpenAI investors second thoughts.
📰 来源：Tavily Search | 🕒 Wed, 15 Apr 2026 02:31:04 GMT
3. AI & Tech Brief: Radical activists and AI safety - The Washington Post
# AI & Tech Brief: Radical activists and AI safety. Plus, Anthropic hires Ballard Partners. Make us preferred on Google. Sign up here to get this newsletter in your inbox. * AI safety advocates and industry groups are casting blame on each other as the nation reckons with bouts of violence related to AI and data centers. * While Congress remains preoccupied, the fight over AI regulation is spilling out into states including Colorado and Illinois. (**Plus,** Anthropic hires Ballard Partners to lobby.). * Should the outputs of AI chatbots be treated like free speech? Court decisions will determine if AI companies are shielded from liability associated with the outputs of their models.
📰 来源：Tavily Search | 🕒 Mon, 13 Apr 2026 18:53:22 GMT
4. Anthropic caused panic that Mythos will expose cybersecurity weak spots, but one industry veteran says the real problem is fixing, not finding, them - Fortune
# Anthropic caused panic that Mythos will expose cybersecurity weak spots, but one industry veteran says the real problem is fixing, not finding, them. Anthropic caused an industrywide panic last week when it announced Claude Mythos Preview, an AI model with a knack for uncovering high-level cybersecurity vulnerabilities. Anthropic has said Mythos is so powerful it won’t be publicly released, and it is being made available only to a group of 40 organizations including tech companies such as Microsoft, Apple, and Google, as well as others like cybersecurity company CrowdStrike and bank JPMorgan Chase so they can use the technology to improve their own security infrastructure through an effort it called Project Glasswing. Incidentally, *Fortune* was the first to report on the development of Mythos, thanks to a security lapse in which the company left details about the large language model in a publicly accessible database. Anthropic caused panic that Mythos will expose cybersecurity weak spots, but one industry veteran says the real problem is fixing, not finding, them.
📰 来源：Tavily Search | 🕒 Mon, 13 Apr 2026 19:04:00 GMT
5. OpenAI Releases Cyber Model to Limited Group in Race With Mythos - Bloomberg.com
# OpenAI Releases Cyber Model to Limited Group in Race With Mythos. Provide news feedback or report an error. Send a tip to our reporters. OpenAI is letting a select group of users access a new artificial intelligence model that’s meant to be more adept at spotting software security vulnerabilities, one week after rival Anthropic PBC announced a limited release of an AI tool called Mythos. The ChatGPT maker said Tuesday that it’s beginning to roll out GPT-5.4-Cyber, which is aimed at finding issues in software so organizations can fix them. GPT-5.4-Cyber also places fewer constraints on the ways users can probe the model for that task, OpenAI said. The model will be offered to some participants of OpenAI’s Trusted Access for Cyber program, which the company rolled out in February to let certain customers and cybersecurity professionals try its most capable offerings. Before it's here, it's on the Bloomberg Terminal. NewslettersExplainersPointed News QuizAlphadots GameThe Big TakeGraphicsSubmit a TipAbout Us.
📰 来源：Tavily Search | 🕒 Tue, 14 Apr 2026 20:00:22 GMT
6. OpenAI embraces Amazon, says Microsoft partnership was problematic - Axios
# OpenAI rips Anthropic, distances itself from Microsoft. OpenAI is tightening its ties with Amazon, saying its early investor and partner Microsoft was holding it back. **Driving the news:** OpenAI's recently established cloud partnership with Amazon Web Services has generated "frankly staggering" demand from enterprise customers, OpenAI's new revenue chief, Denise Dresser, said in an internal memo. **Friction point:** Dresser wrote that OpenAI's Microsoft partnership has been "foundational to our success" but "has also limited our ability to meet enterprises where they are," which, "for many," is Amazon's less-restrictive Bedrock platform. * The move comes after Amazon recently announced a $50 billion investment in OpenAI and a deal to become the exclusive third-party cloud distribution provider for OpenAI Frontier, its new agent-management platform. * Microsoft, which declined to comment for this story, started describing OpenAI as a competitor in 2024. **In the Sunday memo,** Dresser also ripped OpenAI rival Anthropic, saying the Claude creator is constructing a narrative "on fear, restriction, and the idea that a small group of elites should control AI.".
📰 来源：Tavily Search | 🕒 Mon, 13 Apr 2026 19:24:33 GMT
7. Anthropic Is Speaking to Government About Mythos, Co-Founder Says - Forbes
Our best stories, exclusive reporting and Forbes perspectives on the day’s top news. Create your free account to receive personalized offers to your inbox, save articles and follow authors to get their latest stories in a weekly email roundup just for you! Anthropic co-founder Jack Clark said the company was speaking to the U.S. government about its new AI model Claude Mythos, which it has warned is too dangerous to release to the public—despite a dispute with the Pentagon that led to the company being labeled a supply-chain risk earlier this year. Clark did not elaborate on the nature of the conversations his company was having with the government, or if they were discussing removing the supply-chain risk designation, and Anthropic did not immediately return a request for comment from Forbes. The Pentagon took the unusual step of declaring Anthropic a supply-chain risk after a dispute in February, when Defense Secretary Pete Hegseth demanded the AI company allow it to have “full, unrestricted access” to its models.
📰 来源：Tavily Search | 🕒 Mon, 13 Apr 2026 21:13:28 GMT
8. Meta is building a creepy AI clone of Mark Zuckerberg to chat with staff: report - New York Post
# Meta is building a creepy AI clone of Mark Zuckerberg to chat with staff: report. Meta staffers will soon have the option of chatting with a creepy-sounding virtual clone of CEO Mark Zuckerberg, according to a report published Monday. The Zuckerberg clone is one of multiple AI-powered characters currently in development at Meta, according to the FT. Meanwhile, Zuck has gotten more personally involved with Meta’s AI development efforts and has been spending five to 10 hours per week coding, according to the FT. Meta has said it plans to spend up to $135 billion this year alone as it ramps up development of AI projects and scrambles to compete with rivals like Sam Altman’s OpenAI, Dario Amodei’s Anthropic and Sundar Pichai’s Google. In January, Meta announced that it would no longer allow teens to interact with its AI characters “until the updated experience is ready.”.
📰 来源：Tavily Search | 🕒 Mon, 13 Apr 2026 16:43:06 GMT
9. AI Business Boom Includes Performance Woes, Stanford Index Shows - Bloomberg Law News
# AI Business Boom Includes Performance Woes, Stanford Index Shows. April 13, 2026, 1:00 PM UTC. Businesses are increasingly using artificial intelligence as models improve, yet performance can falter at even basic tasks, according to an annual index. About 88% of organizations used AI for at least one business use last year, Stanford University’s HAI’s 2026 Ai Index, which is being released Monday, found. That’s 10 percentage points more than 2024. “Of those that used AI, 79% of respondents reported that their organizations regularly use generative AI in at least one business function, compared to 71% in 2024,” according to the study. Stanford tracks and collates data from various surveys and studies for its index. ## Learn more about Bloomberg Law or Log In to keep reading:. ### See Breaking News in Context. Bloomberg Law provides trusted coverage of current events enhanced with legal analysis. Log in to keep reading or access research tools and resources. © 2026 Bloomberg Industry Group, Inc.
📰 来源：Tavily Search | 🕒 Mon, 13 Apr 2026 13:00:01 GMT
10. OpenAI snaps up AI personal finance start-up Hiro - FinTech Futures
# OpenAI snaps up AI personal finance start-up Hiro. Hiro, which debuted its AI-powered personal assistant five months ago, marks OpenAI's 15th acquisition within the past year. OpenAI has expanded its acquisition portfolio with the purchase of US-based AI personal finance fintech Hiro. Founded by co-CEOs Ethan Bloch and Rushabh Doshi in 2024, Hiro introduced its AI-powered tool around five months ago. Reflecting on the firm's journey, Bloch writes: "We started Hiro with the vision of building an AI personal CFO, and we worked relentlessly to make it real. ## Join FinTech Futures and a panel of industry experts to uncover how agentic AI is revolutionising banking. #### OpenAI snaps up AI personal finance start-up Hiro. Webinar | 15 April 2026 | Agentic AI in banking: Exploring key applications and best practices for implementation. Webinar | 15 April 2026 | Agentic AI in banking: Exploring key applications and best practices for implementation.
📰 来源：Tavily Search | 🕒 Tue, 14 Apr 2026 13:34:13 GMT
11. Tesla leader believes Shanghai factory operations will play a role in robot mass production - AP News
Workers check on the Tesla Model Y at the production lines at the Tesla Gigafactory assembly plant during a media organized tour, in Shanghai, China, Tuesday, April 14, 2026. Workers check on the Tesla Model Y at the production lines at the Tesla Gigafactory assembly plant during a media organized tour, in Shanghai, China, Tuesday, April 14, 2026. Workers check on the Tesla Model Y at the production lines at the Tesla Gigafactory assembly plant during a media organized tour, in Shanghai, China, Tuesday, April 14, 2026. SHANGHAI (AP) — A Tesla Inc. leader said Tuesday he believes its Shanghai factory operations will help resolve the challenges in achieving mass production of the company’s humanoid robots as the U.S. electric vehicle giant pivots to robotics. Wang, who also serves as president of Tesla China, told reporters on a government-organized tour of one of its Shanghai factories that CEO Elon Musk once noted having production at scale is a critical challenge in manufacturing humanoid robots.
📰 来源：Tavily Search | 🕒 Tue, 14 Apr 2026 11:12:48 GMT
12. PIA Automation launches embodied AI and humanoid robotics division targeting industrial applications - Robotics & Automation News
**PIA Automation, a provider of advanced automation solutions, today announced the launch of a new business segment, embodied AI and humanoid robotics, combining the company’s expertise in artificial intelligence, robotics and advanced manufacturing.**. Thomas Ernst, chief sales officer and chief technology officer at PIA Automation, says: “Integrating embodied AI and humanoid robotics into industrial applications opens up entirely new production possibilities for our customers. | cookielawinfo-checkbox-functional | 11 months | The cookie is set by GDPR cookie consent to record the user consent for the cookies in the category "Functional". | cookielawinfo-checkbox-necessary | 11 months | This cookie is set by GDPR Cookie Consent plugin. The cookies is used to store the user consent for the cookies in the category "Necessary". | viewed\_cookie\_policy | 11 months | The cookie is set by the GDPR Cookie Consent plugin and is used to store whether or not user has consented to the use of cookies.
📰 来源：Tavily Search | 🕒 Wed, 15 Apr 2026 05:04:49 GMT
13. Humanoid robots show off their language and boxing skills in Hong Kong - AP News
A humanoid robot manufactured by Novautek showcased at an exhibition in Hong Kong on Monday, April 13, 2026. A humanoid robot showcased at an exhibition in Hong Kong on Monday, April 13, 2026. A humanoid robot showcased at an exhibition in Hong Kong on Monday, April 13, 2026. A humanoid robot manufactured by Novautek showcased at an exhibition in Hong Kong on Monday, April 13, 2026. A humanoid robot manufactured by Novautek showcased at an exhibition in Hong Kong on Monday, April 13, 2026. A humanoid robot showcased at an exhibition in Hong Kong on Monday, April 13, 2026. HONG KONG (AP) — A humanoid robot about the size of a primary school student had something to share in Hong Kong — it sang songs and spoke to people in Mandarin and English, answering whatever questions they posed and delighting the audience around it. Beijing’s latest five-year plan vows to “target the frontiers of science and technology.” Speeding up the development of products like humanoid robots and their applications is part of the 2026-2030 plan for the world’s second-largest economy.
📰 来源：Tavily Search | 🕒 Mon, 13 Apr 2026 12:40:04 GMT
14. Humanoid Robots’ 88% Fail Rate: Completing Home Tasks - Forbes
Image 10: Humanoid robots are shockingly bad at safely completing common household tasks, says a new report from Stanford. Humanoid robots are shockingly bad at safely completing common household tasks, says a new report from Stanford. Humanoid robots are shockingly bad at completing common household tasks safely, according to the 2026 AI Index Report from Stanford University’s Institute for Human-Centered Artificial Intelligence. In fact, the robots we have today, with the AI we have today, fully succeed safely at only about 12% of real household tasks. One of the toughest tests for humanoid robots that are going to function in our homes is the Behavior-1K, built on 1,000 real-world tasks sourced from actual humans reporting what they want robots to do in their homes. The good news is that leading robot companies like Figure AI are training their robots on home environments and showing them complete real-world tasks like emptying a dishwasher, or putting away the groceries.
📰 来源：Tavily Search | 🕒 Tue, 14 Apr 2026 05:06:26 GMT
15. Chinese carmaker Chery’s AI-powered robot up for sale on major online store - New York Post
# Chinese carmaker Chery’s AI-powered robot up for sale on major online store. Chinese carmaker Chery has officially taken a bold step into the future by becoming one of the first robot makers to sell its humanoid product directly to the public. Chery is believed to be the first car company worldwide to offer a humanoid robot for sale to the public. The company says the robot has been designed to perform dual-hand tasks, open car doors autonomously, and be controlled remotely using VR – features that would come in handy in potential roles such as sales, assistant, or receptionist. Chery is believed to be the first car company worldwide to offer a humanoid robot for sale to the public. The company says the robot has been designed to perform dual-hand tasks, open car doors autonomously, and be controlled remotely using VR - features that would come in handy in potential roles such as sales, assistant, or receptionist.
📰 来源：Tavily Search | 🕒 Tue, 14 Apr 2026 17:55:24 GMT