🚀Meta Unleashes Stealthy Web Crawler to Fuel AI Ambitions

Tech Insights AI 0007, Monday 26th August 2024

📰AI News Rundown

1. 🚀Meta Unleashes Stealthy Web Crawler to Fuel AI Ambitions 

2. PHINXT Robotics Secures £2.6M in AI-Driven Warehouse Automation Funding: 🤖💰📈

3. Silicon Valley Showdown: Influencers Engage in Ethical Debates Over AI's Future: 🗣️⚖️

4. Geekbench AI: A New Benchmarking Tool for AI Workloads: 📊🧠🛠️💡

5. Authors File Lawsuit Against Anthropic Over Alleged Piracy: ✍️⚖️📚🛑

1.🚀Meta Unleashes Stealthy Web Crawler to Fuel AI Ambitions

Meta has quietly deployed a new web crawler called "Meta External Agent" to collect vast amounts of online data for training its AI models. Launched last month, this automated bot scrapes publicly available content from websites, including text from news articles and online discussions.

The crawler's existence was confirmed by three firms specializing in tracking web scrapers and bots, including Dark Visitors, a service that helps website owners block known scraping bots. While nearly 25% of popular websites block OpenAI's GPTBot, only 2% are currently blocking Meta's new crawler, according to data from Dark Visitors.

Meta updated its developer website in late July to disclose the crawler's existence but has not made any public announcements. A Meta spokesperson stated that the company has used crawlers under different names for years, and that they train their AI models on publicly available online content.

This development comes as Meta continues to expand its AI capabilities, including updating its Llama language model and Meta AI chatbot. The company is projected to spend up to $40 billion this year, primarily on AI infrastructure and related expenses

Source : click here and here

This table now includes a broader range of firms that help website owners manage and protect their content from web scrapers.

2.PHINXT Robotics Secures £2.6M in AI-Driven Warehouse Automation Funding💰📈🤖

PHINXT Robotics, an AI-led robotics software company based in London, has raised a total of £2.6 million in funding, positioning itself as a leader in warehouse automation.

The company initially secured £608,000 in a pre-seed round led by Fuel Ventures, with participation from Dhyan Ventures and others. This was followed by an oversubscribed £2 million seed round led by Sure Valley Ventures, with additional investments from Ada Ventures and Heartfelt.

Founded in April 2022 by Dr. Yanwen Chen and Dr. Quirino Zagarese, PHINXT is addressing the challenges of warehouse automation with its innovative cloud-based platform. The platform simplifies robotic deployments, reducing setup costs by tenfold and offering a flexible subscription model. This approach aims to make robotics accessible to the 90% of warehouses that remain fully manual due to high costs.

PHINXT's AI technology coordinates robots at the edge, enhancing reliability and management ease. Early adopters have reported doubled profit margins and up to four times increased productivity. The company has already secured contracts, including a pilot program with a major UK grocery retailer.

With the new funding, PHINXT plans to expand its engineering and sales teams and enter mainland European markets. Dr. Chen, CEO and co-founder, brings significant expertise with two PhDs in Computer Science focused on robotics.

As the global mobile robot market is projected to reach $16 billion by 2027, PHINXT Robotics is poised to revolutionize warehouse automation and beyond with its AI-driven solutions.

Source : click here

3.Silicon Valley Showdown: Influencers Engage in Ethical Debates Over AI's Future🗣️⚖️

 In a surprising turn of events, a significant rift has emerged between tech titans Meta and Microsoft, sending ripples across Silicon Valley. The two companies, both leaders in the artificial intelligence (AI) space, are reportedly at odds over differing visions for the future of AI technology. This clash highlights the competitive and rapidly evolving landscape of AI development, where collaboration and rivalry often go hand in hand.

Microsoft's AI Ethical Strategy

Microsoft's approach to AI ethics is comprehensive and structured around six key principles: accountability, inclusiveness, reliability and safety, fairness, transparency, and privacy and security.These principles guide the development and deployment of AI systems to ensure they are responsible and trustworthy. Microsoft has established frameworks like the Responsible AI Standard, which provides actionable guidance for building AI systems that uphold these values.

• Accountability: Microsoft emphasizes accountability by ensuring that those who design and deploy AI systems are responsible for their actions and decisions. This includes establishing internal review bodies to provide oversight and guidance.

• Inclusiveness: The company focuses on inclusive design practices to address potential barriers and empower people with impairments through assistive technologies.

• Fairness and Transparency: Microsoft actively works to prevent discrimination in AI systems and provides tools like Fairlearn to assess and improve model fairness. Transparency is achieved by making AI systems understandable and providing documentation like Transparency Notes.

• Privacy and Security: Microsoft prioritizes the protection of personal data and employs techniques like differential privacy to preserve user privacy while safeguarding AI systems against cybersecurity threats

4.Geekbench AI: A New Benchmarking Tool for AI Workloads🛠️💡

Geekbench AI is a cross-platform benchmarking tool designed to evaluate the performance of AI workloads on various devices. It measures the capabilities of CPUs, GPUs, and NPUs using real-world machine learning tasks, providing insights into whether a device is equipped to handle current and future AI applications.

Geekbench AI runs ten AI workloads with three different data types, offering a comprehensive view of on-device AI performance. Users can select specific hardware components for testing and choose from available AI frameworks like Core ML or QNN to determine the best combination for specific workloads

Source : click here

Competitors in the AI Benchmarking Space

Several other tools and platforms compete with Geekbench AI in the AI benchmarking space:

• MLPerf mlcommons.org: A widely recognized benchmarking suite that evaluates machine learning performance across various workloads and hardware configurations. It is known for its comprehensive and standardized approach to benchmarking AI models.

• AI Benchmark ai-benchmark.com: This tool focuses on evaluating the performance of AI models on mobile devices. It provides detailed insights into the capabilities of smartphones and tablets in handling AI tasks.

• SPEC AIspec.org : Developed by the Standard Performance Evaluation Corporation, SPEC AI benchmarks are designed to evaluate the performance of AI workloads on server and data center hardware.

These competitors, along with Geekbench AI, provide valuable insights into the performance of AI hardware and software, helping developers and consumers make informed decisions about their technology choices.

5.Authors File Lawsuit Against Anthropic Over Alleged Piracy⚖️📚

A group of renowned authors has filed a lawsuit against AI company Anthropic, accusing it of using their copyrighted works without permission to train its AI models.

The lawsuit, filed in federal court, seeks damages and an injunction to stop Anthropic from further using the disputed materials.The authors claim that Anthropic's actions infringe on their intellectual property rights and devalue their creative work.

What does this mean to AI development?
This case highlights the ongoing tension between AI developers and content creators over data usage and copyright laws.Anthropic has not yet publicly responded to the allegations but is expected to argue that its data practices are protected under fair use provisions. The outcome of this lawsuit could set a significant precedent for how AI companies source and utilize data, influencing the broader AI industry.

Source click here

🛠️AI Tools

💰AI Investments