 |
|
|
|||||
Anthropic launched its newest model, Claude Opus 4.5, putting the company back atop the benchmark rankings for AI software coding. Opus 4.5 scores over 80% on the widely-used SWE-bench, which tests models for software engineering skill. Googles impressive Gemini 3 Pro, launched last week, briefly held the top score with 76.2%. Anthropics Claude product lead Scott White tells Fast Company that the model has also scored higher than any human on the engineering take-home assignment the company gives to engineering job candidates. Of course Opus 4.5 does a lot more than coding. Anthropic says Opus 4.5 is also the best model in the world for powering AI agents and for operating a computer, and that it’s meaningfully better than other models at tasks like deep research and working with slides and spreadsheets. Opus 4.5 also notched state-of-the-art (best) scores in several other key benchmarks, including Agentic coding SWE-bench Verified, Agentic tool use T-2 bench, and Novel problem solving ARC-AGI-2. A major challenge with applying AI in real-world work settings is the models ability to deal with complexity and ambiguity. White says Anthropic customers feel that Opus 4.5 is better than earlier models at dealing with uncertainty and handling trade-offs without a lot of hand-holding from human workers. Enterprise customers are increasingly using Anthropic models for office task automation, financial modeling, and document creation, White says. Customer Fundamental Research Labs reported 20% accuracy improvements and 15% efficiency gains on Excel automation tasks using the new model, he adds. Anthropic has been on a sprint for the past couple of months, releasing Claude Sonnet 4.5, Haiku 4.5, and new products including Claude Skills, Claude Code on the web, and industry-specific versions for financial services and life sciences. Opus 4.5 will become the new default model for subscribers of higher-end plans, and available as a drop-down menu option for Pro, Standard, Team, and Enterprise users. Its also available to developer customers via the company’s API, as well as via the Amazon Bedrock, Google Vertex, and Microsoft Azure clouds. Anthropic says its also extending access to a beta version of the Claude plugin for Chrome, which has been in limited preview, to all Mac users. The company is also making Claude for Excel available to Mac Team and Enterprise users in beta, expanding beyond its previous invite-only research preview.
Category:
E-Commerce
Its official: DOGE, the Trump administrations Department of Government Efficiency, has met an early end. Even before Trump took office, DOGE was conceived of as an outside advisory board that would recommend government reforms and find $500 billion in annual spending to cut. The day after Trump was inaugurated, the department was officially founded, with Elon MuskCEO of Tesla and the worlds richest manat its helm. Within the first 100 days of Trumps second term, DOGE played a central role in cutting hundreds of thousands of jobs, nixing funds to foreign food aid and medical research, overhauling longtime government cybersecurity systems, targeting federal DEI programs for elimination, and more. But it seems like DOGEs fate was ultimately to burn bright and fizzle out. After Musk departed from his leadership position amidst a public feud with Trump, DOGE slowly faded from news headlines and the public consciousness. Finally, earlier this month, a Reuters reporter asked Scott Kupor, director of the U.S. Office of Personnel Management, about DOGEs status. “That doesn’t exist,” Kupor responded. He added that DOGE is no longer a centralized entity, and Reuters found that several of the department’s former employees have moved on to other roles, including two workers who are now involved with the new National Design Studio. The agencys lackluster shutdown comes months ahead of its official conclusion, which was meant to be July 2026, according to an executive order signed earlier this year by Trump. Here, we take a look back at some of DOGEs most egregious moves as the agency meets its untimely end. DOGE shuts down USAID One of DOGEs first high-profile moves was to dismantle the U.S. Agency for International Development (USAID), a branch of the government dedicated to administering foreign aid. At the time, Musk took to X to share that he had spent the weekend feeding USAID into the wood chipper. In 2023, USAID managed over $40 billion of appropriations provided to around 130 countries, according to the Congressional Research Service. Of that amount, $16.8 billion went toward governance, while $10.5 billion went to humanitarian aid and $7 billion to health efforts. The USAID shutdown has had devastating ripple effects, including wasting massive food stores amidst a global hunger crisis and threatening Agent Orange cleanup efforts in Vietnam. According to one recent analysis from Atul Gawande, a Harvard T.H. Chan School of Public Health professor and former assistant administrator for global health at USAID, the shutdown has already resulted in hundreds of thousands of deaths from infectious diseases and malnutrition. Thousands of federal workers laid off As part of its cost-cutting initiative, one of DOGEs main focuses became eliminating federal jobs that it deemed to be redundant. This included firings at the U.S. Department of Health and Human Services, the Food and Drug Administration, the U.S. Postal Service (USPS), and more. Kupor told the New York Times in an August interview that the DOGE cuts accounted for almost 300,000 jobs eliminated in total. More recently, some agencieslike the General Services Administrationhave begun asking former employees to return to work ahead of the end of the fiscal year. DOGE reportedly gains unprecedented access to federal databases After DOGEs founding, the agency reportedly received unprecedented access to a number of government databases and computer systems, including the Department of Treasurys payment systems, sensitive IRS data, Social Security records, and data held by the Department of the Interior. In April, David Evan Harris, a chancellors public scholar with the University of California, Berkeley, told Fast Company, Its very unclear what kinds of security protocols the DOGE team is using, and if they are taking any steps to make sure that private data of government employees and U.S. citizens, and even confidential data about U.S. government programs is not being turned into training data or retained improperly by any of these AI companies that theyre working with. According to a September report from Senate Democrats, Harris had reason to be concerned. The report alleges that DOGE copied Americans sensitive Social Security and employment data into a cloud database without any verified security controls, putting people’s personal data at risk of foreign hacks.
Category:
E-Commerce
The U.S. stock market is rising again on Monday, ahead of a week with shortened trading because of the Thanksgiving holiday. The S&P 500 climbed 1.4% and added to its jump from Friday. The Dow Jones Industrial Average was up 221 points, or 0.5%, as of 12:30 p.m. Eastern time, and the Nasdaq composite was 2.4% higher. Stocks got a lift from rising hopes that the Federal Reserve will cut its main interest rate again at its next meeting in December, a move that could boost the economy and investment prices. The market also benefited from strength for stocks caught up in the artificial-intelligence frenzy. Alphabet, which has been getting praise for its newest Gemini AI model, rallied 5.2% and was one of the strongest forces lifting the S&P 500. Nvidia rose 2.1%. But Monday’s gains were hesitant, and the S&P 500 rallied to a gain of 1% only to halve it within the first 15 minutes of trading, before picking up momentum again. Stocks have been swinging sharply, not just day to day but also hour to hour, in recent weeks as worries weigh about what the Fed will do with interest rates and whether too much money is pouring into AI and creating a bubble. All the uncertainty is creating the biggest test for investors since an April sell-off, when President Donald Trump shocked the world with his Liberation Day tariffs. Still, despite all the recent fear, the S&P 500 remains within 2.8% of its record set last month. Several more tests lie ahead this week for the market, though none loom quite as large as last weeks profit report from Nvidia or the delayed jobs report from the U.S. government for September. One of the biggest tests will arrive Tuesday, when the U.S. government will deliver data showing how bad inflation was at the wholesale level in September. Economists expect it to show a 2.6% rise from a year earlier, the same inflation rate as August. A higher-than-expected reading could deter the Fed from cutting its main interest rate in December for a third time this year, because lower rates can worsen inflation. Some Fed officials have already argued against a December cut in part because inflation has stubbornly remained above their 2% target. Traders are nevertheless betting on a 77% probability that the Fed will cut rates next month, up from 71% on Friday and from less than a coin flips chance a week ago, according to data from CME Group. U.S. markets will be closed on Thursday for the Thanksgiving holiday. A day later, its on to the rush of Black Friday and Cyber Monday. On Wall Street, U.S.-listed shares of Danish drugmaker Novo Nordisk fell 5.8% Monday after it reported that its Alzheimers drug failed to slow progression of the disease in a trial. Grindr dropped 9.9% after saying it’s breaking off talks with a couple of investors who had offered to buy the company, which helps its gay users connect with each other. A special committee of the company’s board of directors said it had questions about the financing for the deal by the investors, who collectively own more than 60% of Grindr’s stock. Bitcoin, meanwhile, continued it sharp swings. It was sitting near $87,600 after bouncing between $82,000 and $94,000 over the last week. It was near $125,000 last month. In stock markets abroad, indexes were mixed in Europe following a mixed finish in Asia. Hong Kongs Hang Seng jumped 2% for one of the worlds biggest moves. It got a boost from a 4.7% leap for Alibaba, which has reported strong demand for its updated Qwen AI app. Alibaba is due to report earnings on Tuesday. In the bond market, Treasury yields held relatively steady. The yield on the 10-year Treasury eased to 4.04% from 4.06% late Friday. Elaine Kurtenbach and Matt Ott AP Business Writers Matt Ott and Elaine Kurtenbach contributed.
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||