Yazılar

Russian TV Airs Fake Report on DeepSeek’s ‘Soviet Code

/in /tarafından

A fabricated Russian news story claiming that China’s DeepSeek AI app is based on secret Soviet code has made its way onto state TV, illustrating a wave of nostalgia in Russia for a bygone era of technological might. The hoax originated from Panorama, a satirical fake news outlet that openly publishes fictional content. The fake report featured an interview with DeepSeek’s founder, Liang Wenfeng, who was quoted as praising Soviet-era programmers and their alleged role in developing the code for the AI startup.

According to the spoof, the DeepSeek code was allegedly created in 1985 by a team led by Viktor Glushkov, a renowned Soviet scientist credited with developing the first personal computer in the Soviet Union in the 1960s. Glushkov was also behind the creation of a data-processing network designed to manage the Soviet planned economy, which some argue contained early features of artificial intelligence.

Despite its fictional nature, the story gained traction, appearing on Rossiya One, a national state television channel, as if it were genuine news. It was further amplified on social media, with prominent figures like Communist Party leader Gennady Zyuganov sharing the report, calling the Soviet Union “the most educated and advanced country”—a post that was later deleted.

Russia’s domestic AI landscape, however, lags behind its global competitors, ranking 31st out of 83 nations for AI implementation, investment, and innovation, according to the Global AI Index by UK-based Tortoise Media. Russia not only trails technological giants like the United States and China but also faces stiff competition from other BRICS members like India and Brazil. Despite boasting two significant domestic AI models, Russia closely monitors China’s AI advancements, particularly the success of DeepSeek’s recent models, which have shaken up the global tech scene.

Lyft Partners with Anthropic for AI-Powered Customer Care

/in /tarafından

Lyft (LYFT.O) announced on Thursday that it has partnered with Amazon (AMZN.O) and Alphabet-backed startup Anthropic to introduce artificial intelligence tools to enhance its customer care operations. The company has already been using Anthropic’s Claude AI model, which is integrated with Amazon’s Bedrock generative AI platform. This collaboration has reportedly reduced average customer service resolution times by 87%, allowing the platform to address thousands of customer inquiries daily.

Despite concerns about AI-driven job losses, Lyft emphasized that the goal is not to replace human workers but to enhance the quality and efficiency of its customer support services. Lyft’s approach involves initially addressing customer issues with the AI assistant, directing users to human agents only if further assistance is required.

“We see AI as an opportunity to improve the quality and effectiveness of our operations, not to reduce headcount,” said Jason Vogrinec, Lyft’s executive vice president of platforms. However, industry experts have pointed out that AI models can sometimes produce incorrect or fabricated information, limiting their ability to completely replace human agents. Lyft also noted that complex issues such as safety concerns, account deactivations, and fraud will still be handled by human representatives.

Through this collaboration, Lyft and Anthropic plan to explore additional AI-driven tools for both riders and drivers. Anthropic will also provide training for Lyft’s engineers on the technology, further enhancing the platform’s AI capabilities.

Lyft is scheduled to report its quarterly earnings after market close on Tuesday.

 

Qualcomm Shares Fall on Downbeat Forecast for Licensing Business

/in /tarafından

Qualcomm’s (QCOM.O) shares dropped by around 5% in early trading on Thursday following a disappointing forecast for its patent licensing business, despite strong expectations for quarterly sales and profits. The chipmaker revealed that its licensing business, which contributed 14.8% to its total revenue in the reported quarter, would experience no sales growth this year due to the expiration of its agreement with Huawei Technologies (HWT.UL).

TD Cowen analysts had initially expected the removal of Huawei’s royalty payments to have a mild impact, but they noted that the development adds to the “wall of worry” surrounding Qualcomm’s stock. However, analysts pointed out that Qualcomm has secured licensing agreements with two other Chinese smartphone manufacturers, which may help mitigate some of the losses.

The company’s first-quarter performance exceeded expectations, driven by strong demand for AI features in mobile devices, and is often seen as a barometer for broader smartphone industry trends. Qualcomm’s second-quarter sales forecast of $10.75 billion, with adjusted profits of $2.80 per share, surpassed analysts’ estimates of $10.34 billion and $2.69 per share, respectively, as reported by LSEG data.

While Qualcomm credited growth in its smartphone division to strong sales from China, powered by government subsidies and flagship smartphone launches, it also highlighted positive performance across other business segments, including handsets, autos, and IoT.

Despite gains in 2024, Qualcomm’s stock has underperformed AI chip leader Nvidia (NVDA.O), whose shares surged by 171%. Qualcomm’s stock has increased by 6% this year, far surpassing the losses seen by competitors like Intel (INTC.O), which saw a 60% decline, and Advanced Micro Devices (AMD.O), which dropped by 18%.

As a result of the company’s outlook, Qualcomm’s median price target decreased slightly to $192, down from $199 prior to the report, according to LSEG data. The company’s forward price-to-earnings ratio stands at 15.02, significantly lower than Nvidia’s 27.64 and Intel’s 32.21.