Yazılar

Anthropic launches low-cost Haiku 4.5 model to make AI more accessible for businesses

/in /tarafından

AI startup Anthropic has unveiled a major update to its smallest model, Haiku, as it seeks to make artificial intelligence more affordable and practical for companies outside Silicon Valley. The new version, Haiku 4.5, costs about one-third as much as Anthropic’s Sonnet 4 and just one-fifteenth the price of its flagship Opus model, while matching or outperforming mid-tier models on tasks like coding and data synthesis.

Chief Product Officer Mike Krieger said the upgrade reflects a growing demand among traditional businesses for cost-effective AI tools that still deliver high performance. “Small models really help because they can be a more economical way of deploying at scale,” Krieger told Reuters, noting that cheaper AI makes it easier for firms to integrate intelligent assistants into systems used by thousands of employees.

Anthropic’s enterprise business now accounts for about 80% of its revenue, with over 300,000 corporate customers using its AI tools internally or within their products. The company’s annual revenue run rate has reached nearly $7 billion, underscoring its rapid ascent in the AI sector.

Founded in 2021 by former OpenAI employees, the San Francisco-based company has become one of the strongest challengers to OpenAI, backed by a recent valuation of $183 billion.

Anthropic’s smaller models, such as Haiku, aim to balance power and affordability at a time when companies are pushing back against the massive computational costs of training and running large-scale AI systems. The firm says businesses can even combine models — using advanced ones for strategic planning and smaller ones for everyday tasks like information synthesis and web searches.

Zalando Harnesses Generative AI to Slash Campaign Costs and Speed Up Fashion Marketing

/in /tarafından

Zalando, the European online fashion retailer, announced it is using generative artificial intelligence (AI) to accelerate content production for its app and website, allowing it to respond quickly to viral fashion trends while cutting marketing costs by up to 90%.

According to Matthias Haase, VP of Content Solutions at Zalando, generative AI has reduced image production times from six to eight weeks down to just three to four days. The tech enables the brand to stay agile in an industry where fast reaction to fleeting trends—like “brat summer” or “mob wife”—can determine visibility and sales.

More than 70% of Zalando’s editorial campaign imagery in Q4 2024 was AI-generated, including content used in its trend recaps and seasonal promotions. The company is also developing AI-generated “digital twins” of human models, creating three-dimensional replicas for consistent visual use across product pages and campaigns without the need for repeated photo shoots.

“It’s not that AI content is better than human-created content,” Haase noted, “but it’s more timely and relevant to what customers care about now.”

Zalando joins a growing list of retailers integrating AI into their workflows. In March, H&M announced a similar initiative using digital twin technology developed with a modeling agency. The technology appeals especially to mid-tier retailers, offering an alternative to expensive, logistically intensive photo shoots favored by high-end fashion houses.

Asked about the impact of AI on creative jobs, Haase said that photographers and creatives will remain essential, but must adapt. “Creative minds now have, instead of two hands, six hands,” he said, emphasizing AI as a tool, not a replacement.

Four Ways DeepSeek Could Change Everything

/in /tarafından

The release of DeepSeek’s highly effective and cost-efficient large language model has made waves in the AI industry, promising far-reaching implications for technology, trade, and U.S.-China economic relations. While the immediate market impact may have been brief, the long-term effects could be profound. Here are four predictions on how DeepSeek might shape the future:

  1. Artificial Intelligence Costs Will Continue to Plummet
    Innovations typically aim to achieve more with less, and AI is no different. Before DeepSeek’s release, the costs of leading AI models had already fallen by about 80% annually over the past two years. DeepSeek has accelerated this trend by making AI models 30 times cheaper compared to market leader OpenAI, through algorithmic advancements and aggressive pricing strategies. This deflationary trend is expected to persist as more research and competition in the AI field drive costs lower.

  2. The AI Economic Pie Will Get Bigger and Be Sliced Differently
    As AI becomes more affordable and accessible, demand is expected to grow, following the concept of Jevons paradox, which suggests that more efficient technology leads to greater consumption of resources. As foundational models become commoditized, the focus will shift to applications, pushing more resources toward the deployment of AI in specific tasks, or “inference,” rather than training models. This shift could spark increased demand for custom-designed chips like XPUs, optimized for specific AI applications, as opposed to traditional GPUs. Nvidia has already observed that demand for inference chips is growing faster than for training chips, signaling a broader industry shift.

  3. U.S. Chip Export Controls Will Deserve Careful Reassessment
    DeepSeek’s success came from utilizing less advanced and fewer chips than its U.S. counterparts, illustrating how innovation can thrive even under constraints. Despite ongoing U.S. export controls that may limit DeepSeek and other Chinese companies in the short term, these restrictions are unlikely to halt their progress. The U.S. risks isolating its chip technologies from China’s market, potentially on a permanent basis. Additionally, the export controls may undermine U.S. efforts to address trade imbalances with China, as the country may opt to focus on developing its own capabilities rather than relying on U.S. imports.

  4. U.S. and Chinese Tech Leaders’ Interests May Align
    While initially concerning to U.S. investors, DeepSeek’s breakthrough and its open-source model have been embraced by many major U.S. tech companies. Cloud platforms like Microsoft, AWS, and Hugging Face are already incorporating models based on DeepSeek’s R-1, noting that cheaper large language models should increase demand for their cloud services, boosting their revenue streams. In the long run, businesses across both countries could benefit from the productivity gains and cost savings that AI applications offer. This could foster potential collaboration between U.S. and Chinese tech leaders, despite existing tensions. The evolution of AI presents a tremendous opportunity for both superpowers to collaborate, especially as they pursue artificial general intelligence, though ongoing geopolitical conflicts could limit this cooperation.