A naukri.com initiative
Data Science News
VentureBeat
2w
253
Image Credit: VentureBeat
Anthropic just made every Claude user a no-code app developer
- Anthropic turns Claude into app platform, enabling users to create shareable applications easily.
- Claude's intelligence embedded in creations allows for real-time adaptation and independent functionality.
- Move signals shift to dynamic, interactive AI experiences, competing with OpenAI's Canvas feature.
- Users can create smart tutors, data analyzers, and games without coding; moderation safeguards in place.
Read Full Article
15 Likes
Analyticsindiamag
2w
219
Image Credit: Analyticsindiamag
India’s Satellite Internet Scramble Has a Spectrum Problem
- India's satellite internet ambitions face spectrum allocation challenges amid regulatory disagreements.
- Players like Eutelsat OneWeb, Starlink, and Jio Satellite Communications Limited are gearing up.
- Debate heats up over pricing, infrastructure parity, spectrum allotment policy, and fair competition.
Read Full Article
13 Likes
Analyticsindiamag
2w
2.3k
Image Credit: Analyticsindiamag
LTTS Bags Another $50 Million Deal In Sustainability Segment
- L&T Technology Services Limited (LTTS) secured a $50 million deal with a top energy company for its sustainability segment as a global engineering partner for Enterprise Data and Digital Services.
- The partnership is an extension of a longstanding relationship between the two companies spanning over a decade.
- LTTS will offer Enterprise Data and Digital Services to the client's affiliates worldwide as part of the agreement.
- With extensive experience in engineering, digital services, and enterprise data management, LTTS aims to provide valuable insights to energy firms globally using its global delivery model.
- The deal, structured as a long-term framework agreement, entrusts LTTS with greater responsibility for delivering top-quality digital services.
- Amit Chadha, CEO of LTTS, highlighted the company's commitment to supporting the client's digital transformation and operational excellence through a combination of engineering expertise and advanced technologies.
- In a recent development, LTTS also secured an $80 million contract with a US-based industrial products manufacturer for digital transformation purposes within the Sustainability segment.
- This new deal win will lead LTTS to establish a Centre of Excellence in India to drive innovation for the client's digital transformation and product lifecycle management.
Read Full Article
5 Likes
Analyticsindiamag
2w
138
Image Credit: Analyticsindiamag
Murati’s Startup Bets on ‘RL for Business’ After Raising $2 Bn
- Former OpenAI CTO Mira Murati's startup, Thinking Machines Lab, has raised $2 billion from investment firm Andreessen Horowitz at a $10 billion valuation.
- TML focuses on creating custom AI solutions for businesses to boost revenue, described as 'RL for businesses'.
- The startup plans to provide industry-specific AI solutions for customer support, retail, and investment banking, potentially with premium pricing.
- TML is also developing a consumer-facing product like a chatbot and collaborates with Google Cloud for powerful NVIDIA servers, attracting interest from other cloud providers.
Read Full Article
4 Likes
Analyticsindiamag
2w
171
Image Credit: Analyticsindiamag
Coralogix introduces Olly, an advanced AI-Powered Observability Agent in India
- Coralogix has launched Olly, an AI-Powered Observability Agent aimed at enhancing enterprise observability.
- Olly allows users at all management levels to access observable data and receive real-time insights.
- Ariel Assaraf, CEO of Coralogix, emphasizes Olly as an intelligent assistant empowering data-driven decision-making.
- The launch in India is part of Coralogix's global expansion drive following a successful $115 million Series E funding round.
- India's observability market is projected to reach $280.3 million by 2030, with a growing demand for scalable observability solutions.
- Olly addresses challenges like managing data volume and quality, leveraging AI to provide contextual insights.
- Observability has gained strategic importance post major outages, emphasizing the need for robust solutions like Olly.
- Platforms like Olly offer real-time feedback, predictive alerts, and deep system visibility, crucial for preventing incidents like the CrowdStrike outage in 2024.
- By embedding intelligence, Olly positions Coralogix at the forefront of AI-native infrastructure.
- Observability is seen as crucial for operational continuity and future autonomous operations.
- Olly combines advanced AI with telemetry data to support smarter decision-making.
- The launch of Olly in India aligns with the increasing demand for observability solutions in the market.
- Coralogix's valuation has surpassed $1 billion, driven by customer growth and demand for observability solutions.
- The launch of Olly comes amidst rising demand for advanced observability solutions in the face of persistent industry challenges.
- Olly's intelligence aims to help enterprises overcome data management issues and AI expertise shortages.
- The introduction of Olly positions Coralogix as a significant player in the observability and AI-powered infrastructure landscape.
Read Full Article
10 Likes
Analyticsindiamag
2w
400
Image Credit: Analyticsindiamag
Foxconn to Invest $1.49 Billion to Set Up iPhone Manufacturing Facility in India
- Foxconn, also known as Hon Hai, has received approval for investments totaling over $2.2 billion in the US and India.
- The investment in India includes $1.49 billion to set up a new facility for manufacturing smartphone components.
- Foxconn is diversifying its production beyond China, with India seen as a key destination for iPhone manufacturing.
- Nearly 97 percent of iPhones assembled in India between March and May this year were exported to the US.
- India's role in iPhone manufacturing is growing, with exports estimated at $3.2 billion.
- Foxconn is proceeding with investments in India despite pressure to bring production back to the US.
- Experts predict that India-made iPhones could make up 30 percent of global shipments by 2025.
- Hon Hai is expanding its production footprint globally, with plants and offices in various countries.
- The investments reflect Foxconn's efforts to reduce dependency on single markets and de-risk its supply chain.
Read Full Article
8 Likes
Dev
2w
408
Image Credit: Dev
Oreo's Top K Treats (Bucket Sort)
- Oreo, a data-driven Persian cat with a treat addiction, sets out to determine the top treats in its daily feast using bucket sort.
- Oreo encounters the 'Top K Frequent Elements' problem and uses bucket sort for a quick solution.
- By analyzing recent treat data provided by its human, Oreo identifies shrimp and cookies as the top treats.
- Oreo first counts treat frequencies using a Map and then organizes treats into frequency-based buckets.
- The bucket sort approach allows Oreo to efficiently find its top K treats while avoiding traditional sorting complexities.
- The time complexity of the approach is O(n), making it a linear time solution well-suited for this specific problem.
- The space complexity is also O(n) due to the use of a map for storing treat counts and buckets for frequency organization.
- Oreo's top 2 treats, shrimp and cookies, are confirmed to appear 3 times each based on the bucket sort results.
Read Full Article
24 Likes
Medium
2w
341
Image Credit: Medium
No Verifier? No Problem: Reinforcement Learning with Reference Probabilities
- RLPR (Reinforcement Learning with Reference Probability Reward) is proposed to address the limitation of RLVR in general domains without verifiers.
- RLVR is often restricted to math and code tasks due to the reliance on domain-specific verifiers for rewards.
- RLPR uses the model's own probability of generating a reference answer as a reward signal, eliminating the need for external verifiers.
- This self-rewarding setup alleviates manual reward engineering and enhances scalability across different domains.
- The authors present two main technical contributions in the RLPR framework.
- Experiments conducted on 7 benchmarks and 3 model families (Qwen, Llama, Gemma) demonstrate that RLPR improves general reasoning and outperforms verifier-free baselines and even specialized verifier methods like LLMs.
- The code for RLPR has been made publicly available.
Read Full Article
20 Likes
Analyticsindiamag
2w
210
Image Credit: Analyticsindiamag
Google Accidentally Leaks Its Open-Source Gemini CLI Ahead of Launch
- Google has accidentally leaked Gemini CLI, an open-source command-line interface designed to bring Gemini AI models to developers' terminals.
- Gemini CLI is expected to offer high-capacity usage for free with full integration with Google's Gemini Code Assist.
- The tool is released under the Apache 2.0 license to support coding, content creation, research, and task automation.
- It allows up to 1,000 free requests daily and provides a one-million-token context window for users.
- Gemini CLI boasts the industry's largest usage allowance at 60 model requests per minute with 1,000 model requests per day at no charge.
- The interface supports extensibility via the Model Context Protocol (MCP) and supports AI invocation within scripts.
- Prompt grounding with Google Search and structured system prompts via GEMINI.md are also included in Gemini CLI.
- This release aligns Gemini CLI with Gemini Code Assist, facilitating seamless transitions between terminal and IDE workflows.
- Gemini Code Assist is a free AI coding tool supporting all programming languages with features like code completion, generation, chat, and natural language prompts.
- The Gemini CLI appeared unexpectedly following Google's emphasis on open agentic frameworks and high-trust developer tools.
- There has been no official correction or clarification from Google regarding this accidental leak.
Read Full Article
12 Likes
Analyticsindiamag
2w
116
Image Credit: Analyticsindiamag
Anthropic Proves in Court Using Copyright Content for AI Training is Fair
- A federal judge in San Francisco ruled that Anthropic's use of copyrighted books to train its AI system is fair use under US copyright law.
- Judge William Alsup found Anthropic's AI training transformative as it analyzed writing, extracted uncopyrightable data, and developed new technology.
- Training the large language model (LLM) was considered transformative as it aimed to teach statistical relationships between words to generate new text.
- The process was compared to learning how to write by reading books rather than copying them verbatim.
- There was no evidence that Anthropic's model reproduced or directly substituted the original works.
- Anthropic downloaded millions of copyrighted books from various sources, including pirate sites, to train its AI model.
- The court ruled that obtaining books through piracy infringed authors' copyrights and was not protected as fair use.
- A trial is scheduled in December to determine damages related to the infringement.
- The case originated from a class action lawsuit by authors alleging Anthropic used unauthorized copies of their books for training its AI model.
- While training AI on lawfully obtained works is legal, relying on pirated materials was criticized for undermining copyright protections.
- The ruling is considered significant for the AI industry, affirming fair use with legally obtained materials but cautioning against piracy.
- Various companies, including OpenAI, Meta, and Stability AI, are facing legal challenges over alleged unauthorized use of copyrighted material for training AI models.
- The decision is anticipated to impact copyright disputes in the AI industry, with appeals and further litigation likely to follow.
Read Full Article
6 Likes
Medium
2w
426
Image Credit: Medium
Brownian Motion: The Art of Getting Lost on a Chessboard
- Brownian motion is a random walk where each step is unpredictable, resulting in a jagged trail on a chessboard.
- Speeding up the process creates Brownian motion, where tiny steps lead to a continuous, irregular path.
- Brownian motion is essential because randomness is prevalent in various aspects of life like weather, stock markets, and molecular movements.
- It is a mathematical model for uncertainty, describing chaos and finding structure in unpredictable situations.
Read Full Article
25 Likes
Analyticsindiamag
2w
88
Image Credit: Analyticsindiamag
Anthropic’s Claude to Support Memory and App Creation: Reports
- Anthropic is introducing two major updates to its AI assistant Claude: a memory function and expanded app creation tools.
- The memory feature will allow Claude to recall past user interactions, bringing it closer to competitors' persistent memory functions.
- The update is expected to launch on web and mobile platforms, but an official release date is pending confirmation.
- Claude's Artifacts feature is evolving into an app canvas where users can embed Claude's functionality.
- This update could enable users to create interactive, Claude-powered utilities without coding.
- Anthropic aims to transform Claude into a no-code development platform for AI-driven mini-apps.
- An advanced version of Artifacts is anticipated to be released soon, focusing on empowering users to build applications with Claude.
- These updates may reduce technical barriers for creating custom AI solutions, setting Claude apart in the AI assistant landscape.
Read Full Article
5 Likes
Analyticsindiamag
2w
444
Image Credit: Analyticsindiamag
Animal Health Firm Zoetis Partners with Infosys to Integrate Advanced AI Solutions
- Infosys and Zoetis have entered a strategic collaboration to enhance Zoetis’ IT operations and digital capabilities using advanced AI solutions.
- The partnership aims to provide greater agility for Zoetis to adapt to changing business landscapes and customer needs.
- The collaboration will help Zoetis deliver value to veterinarians, livestock producers, and pet owners globally by improving operational efficiency.
- Zoetis plans to lead the animal health industry by leveraging digital innovation through this collaboration.
- Infosys aims to reduce operational complexity for Zoetis and support their long-term business objectives through this partnership.
- Zoetis expanded its global capability center in Hyderabad last year to drive technological innovation in its portfolio.
- The expansion in Hyderabad is expected to create high-end technology roles focused on AI, machine learning, data science, ERP, and more.
- Zoetis emphasizes the importance of technology, data, and analytics in the rapidly evolving animal healthcare industry.
- The company is seeking talent with expertise in technologies like SAP S4 HANA, Salesforce, and others to meet future needs.
Read Full Article
24 Likes
Medium
2w
103
Image Credit: Medium
Leading Alone: How To Manage When No One’s Coming To Help
- Transitioning from an engineer to a manager means needing to make decisions independently.
- The challenge lies in handling issues without immediate guidance or support.
- A scenario is outlined where an engineering manager faces resistance in addressing technical debt.
- Dealing with resource constraints can lead to a cycle of problems that are hard to fix at the root.
- Managers often face challenges dealing with technical debt and people issues due to competing priorities.
- Strategies are discussed for building resilience as a manager in the face of challenges.
- Tips include using technology to review meetings for insights and seeking external perspectives.
- Adopting methods like rubber duck debugging and talking out problems can aid in decision-making.
- Writing down thoughts after tough conversations can help clarify one's perspective.
- Being resourceful in seeking solutions, including leveraging online resources and communities, is vital for managers.
- Developing self-reassurance and confident decision-making skills are keys to successful management in a potentially solitary role.
Read Full Article
6 Likes
Analyticsindiamag
2w
163
Image Credit: Analyticsindiamag
Microsoft Launches Mu, a Small Language Model That Runs Locally on Copilot+ PCs
- Microsoft has introduced Mu, a compact on-device language model designed to run on Copilot+ PCs' neural processing units.
- Mu is built on a Transformer encoder-decoder architecture for efficient inference by separating input and output tokens, achieving over 100 tokens per second.
- The 330 million parameter model was trained using Azure's A100 GPUs and incorporates features like dual LayerNorm and rotary positional embeddings for speed and accuracy.
- Microsoft applied post-training quantisation, collaborating with hardware partners like AMD, Intel, and Qualcomm to optimize the model for lower-precision formats on edge devices.
- Mu, integrated into Windows Settings, facilitates system adjustments via natural language queries, mapping them to system actions for user convenience.
- The model performed best on multi-word queries with clear intent, implying the need for sufficient context for accurate interpretation.
- Taskspecific fine-tuning enabled Mu to meet performance and latency goals, scaling the training dataset and support for various system settings.
- Mu's development builds on prior models Phi and Phi Silica, expected to underpin future AI agents on Windows devices.
Read Full Article
9 Likes
For uninterrupted reading, download the app