For years, recruiters used machine learning to find potential hires by searching for keywords in resumes and LinkedIn profiles. Although this method helps to narrow the candidate pool, recruiters still have to manually review each profile to determine the best fit for the job. David Paffenholz (pictured left) and Ishan Gupta, then just 22 and 19, realized that LLMs could find talent faster and more efficiently. They built Juicebox, an AI-powered search engine that uses natural language to analyze profess ...

Key Points The popularity of ChatGPT and similar models has analysts questioning Alphabet's dominance in online search. But based on Alphabet's financial profile, ad revenue from Google appears to be keeping pace in the AI era. Alphabet has reinvested these profits into other moneymaking opportunities that remain overlooked. 10 stocks we like better than Alphabet › Ever since OpenAI introduced ChatGPT to the public a few years ago, some Wall Street analysts have sounded the alarm for Alphabet (NA ...

Good morning. Analysts and investors increasingly are using generative AI to review earnings calls, but new research suggests large language models (LLMs) may soon become powerful tools for stock selections. For many years, financial sentiment analysis relied on simple word lists. Take for instance, on an earnings call, that would mean counting a CEO or CFO’s positive phrases like “strong growth” and negative ones like “unexpected losses”—to assign a sentiment score. This rules-based system was transparen ...

Business Performance - EngageLab 客户数量同比增长 171%[9] - EngageLab 累计签约合同价值同比增长 265%[10] - EngageLab 季度收入同比增长 67%，环比增长 24%[11] - 公司与中国联通合作，通过 EngageLab 平台推出智能综合验证（国际版）[14] - 公司董事会批准一项战略计划，将高达 20% 的现金和现金等价物投资于加密货币和数字资产[26] Financial Highlights - 公司首次实现 GAAP 净利润[40] - 总收入为 8990 万人民币，同比增长 13%，环比增长 1%[43] - 开发者服务收入同比增长 14%，环比增长 3%[43] - 垂直应用收入同比增长 10%，环比下降 4%[43] - 开发者服务（订阅）收入同比增长 12%[47] - 增值服务收入同比增长 30%，环比增长 21%[47] - 财务风险管理收入同比增长 27%，客户数量同比增长 48%[49] - 毛利润同比增长 13% 至 5960 万人民币[52] - 递延收入余额为 1.561 亿人民币[58] - 截至 2025 年 6 月 30 日，现金和现金等价物及受限现金为 1.198 亿人民币[61] - 开发者服务（订阅）的净美元留存率 (NDR) 达到 99%[61]

Voy a suponer que la mayoría de ustedes han usado una de las herramientas de inteligencia artificial generativa, chat, GBT, Google, Gemini, Deepsek, ese tipo de cosas. Estas herramientas aparecieron para uso público hace un poco menos de 3 años y pocas tecnologías han sido tan transformadoras y con un con un impacto tan grande en un tiempo tan corto. Cada vez que yo las uso, me tengo que recordar de un montón de cosas.Tengo que recordar, por ejemplo, para empezar, que la inteligencia artificial no es una co ...

Core Insights - Innodata Inc. (INOD) is transitioning from scale data to smart data to enhance the potential of large language models (LLMs) and is focusing on providing Agentic AI services to clients, capitalizing on the strong prospects of agent-based AI [1][2] Group 1: Business Strategy and Market Positioning - The company is adopting a smart data approach to improve factuality, safety, coherence, and reasoning in AI applications, which is expected to boost demand for simulation data and evaluation services [2] - Innodata plans to invest in growth opportunities through short-cycle, high-return initiatives, including custom annotation pipelines, verticalized agent development, and global delivery expansion [3] - The company aims to provide advisory and integration services for AI-native systems and expand into new domains such as multi-agent systems and robotics [3] Group 2: Financial Performance - In the first half of 2025, Innodata reported a 97.7% year-over-year revenue growth to $116.7 million, driven by increased demand from existing clients and higher subscription volumes in its Agility AI-enabled platform [4][9] - The stock has gained 20.8% over the past three months, outperforming the Zacks Computer - Services industry and the broader S&P 500 index [8][9] - Innodata's stock is currently trading at a premium compared to industry peers, with a forward 12-month price-to-sales (P/S) ratio of 4.91, indicating strong market potential [10] Group 3: Earnings Estimates - Earnings estimates for Innodata have increased for 2025 and 2026, with projected earnings of 71 cents and $1.05 per share, respectively [11] - The revised estimate for 2025 reflects a 20.2% year-over-year decline, while the estimate for 2026 indicates a growth of 48.2% [11]

Q3 FY25 Performance - Total revenue decreased to $62.2 million, compared to $70.5 million in Q3 FY24[5] - Gross margin increased to 73.7% from 71.5% in Q3 FY24[5] - Net loss improved to $(2.7) million from $(313.5) million in Q3 FY24[5] - Adjusted EBITDA decreased to $9.0 million from $12.5 million in Q3 FY24[5] - Cash provided by operating activities significantly increased to $48.4 million from $11.1 million in Q3 FY24[5] - Cash balance & marketable securities decreased to $73.7 million from $121.5 million in Q3 FY24[5] Revenue Details - Variable license revenue increased to $34.2 million in Q3 FY25[7] - Pro forma royalties increased to $43.2 million in Q3 FY25[9] - Adjusted Total Billings TTM increased by 3.5% to $226 million[12] Key Performance Indicators - Cerence technology is present in 52% of worldwide auto production (TTM)[12] - Approximately 12 million units shipped with Cerence technology in Q3, a 2.5% YoY increase[12] - Connected attach rate increased to 31% from 27% a year ago[12] - Average PPU on a TTM basis increased to $4.91 from $4.47 a year ago[12] Fiscal Q4 and FY25 Guidance - Q4FY25 revenue is projected to be between $53 million and $58 million[13] - FY25 revenue is projected to be between $244 million and $249 million[13]

Group 1 - The emergence of large language models (LLMs) is reshaping the competitive landscape, diminishing Amazon.com, Inc.'s (NASDAQ: AMZN) AWS dominance [1] - New competitive dynamics are creating strategic headwinds for Amazon, which may hinder its growth prospects [1]

Core Insights - GSI Technology, Inc. has successfully completed the evaluation of its Gemini-II chip, confirming it is production-ready and optimized for Edge AI applications, particularly in GPS-denied environments and next-generation satellite applications [3] - The company reported net revenues of $6.3 million for the first quarter of fiscal 2026, a significant increase from $4.7 million in the same period last year, and gross margin improved to 58.1% from 46.3% year-over-year [4][9] - The outlook for the second quarter of fiscal 2026 anticipates net revenues between $5.9 million and $6.7 million, with a gross margin of approximately 56% to 58% [3] Financial Performance - First quarter fiscal 2026 net revenues were $6.3 million, up 34% from $4.7 million in the first quarter of fiscal 2025 and up 7% from $5.9 million in the fourth quarter of fiscal 2025 [4] - Gross margin for the first quarter of fiscal 2026 was 58.1%, an increase of 200 basis points from the prior quarter and over 1,100 basis points compared to the prior year [7] - Total operating expenses for the first quarter of fiscal 2026 were $5.8 million, a decrease from $6.8 million in the same period a year ago [6] Customer and Sales Insights - Sales to Cadence Design Systems increased significantly to $1.5 million, representing 23.9% of net revenues, compared to $0 in the same period last year [5] - Sales to KYEC and Nokia decreased significantly, with KYEC contributing only $267,000 (4.3% of net revenues) and Nokia contributing $536,000 (8.5% of net revenues) in the first quarter of fiscal 2026 [5] Research and Development - The company is developing a multi-modal large language model (LLM) optimized for edge applications, with benchmark results expected by fall 2025 [3][7] - Research and development expenses for the first quarter of fiscal 2026 were $3.1 million, down from $4.2 million in the prior-year period [6] Cash and Equity Position - The quarter-end cash balance was $22.7 million, an increase from $13.4 million at the end of the previous quarter, reflecting strong cash flow management [7][10] - Stockholders' equity as of June 30, 2025, was $37.4 million, up from $28.2 million at the end of the previous fiscal year [10]

Search & Discovery Challenges in Grocery E-commerce - Instacart faces challenges with overly broad queries (e.g., "snacks") and very specific, infrequent queries (e.g., "unsweetened plant-based yogurt") due to limited engagement data [6][7] - Instacart aims to improve new item discovery, similar to the experience of browsing a grocery store aisle, but struggles due to lack of engagement data [8][9][10] - Existing models improve recall, but maintaining precision, especially in the long tail of queries, remains a challenge [8] LLM-Powered Query Understanding - Instacart utilizes LLMs to enhance query understanding, specifically focusing on query to category classification and query rewrites [10][11][12] - For query to category classification, LLMs, when augmented with top converting categories as context, significantly improved precision by 18 percentage points and recall by 70 percentage points for tail queries [13][21] - For query rewrites, LLMs generate precise rewrites (substitute, broader, synonymous), leading to a large drop in queries with no results [23][24][25][26] - Instacart pre-computes outputs for head and torso queries and caches them to minimize latency, while using existing or distilled models for the long tail [27][28] LLM-Driven Discovery-Oriented Content - Instacart uses LLMs to generate complementary and substitute items in search results, enhancing product discovery and user engagement [31][34] - Augmenting LLM prompts with Instacart's domain knowledge (e.g., top converting categories, query annotations, subsequent user queries) significantly improves the relevance and effectiveness of generated content [39][40][41] - Instacart serves discovery-oriented content by pre-computing and storing content metadata and product recommendations, enabling fast retrieval [42][43] Key Takeaways & Future Directions - Combining LLMs with Instacart's domain knowledge is crucial for achieving topline wins [47] - Evaluating content and query predictions is more important and difficult than initially anticipated [47][48] - Consolidating multiple query understanding models into a single LLM or SLM can improve consistency and simplify system management [28]