AI Tools Threaten Human Employment

PLUS: Microsoft Tops Nvidia Chip Purchases, Vultr Secures $3.5B Funding and more.

Author

Wes Roth
December 19, 2024

In partnership with

This isn’t traditional business news

Welcome to Morning Brew—the free newsletter designed to keep you in the know on the business news impacting your career, company, and life—in a way you didn’t know you needed.

Note: this isn’t traditional business news. Morning Brew’s approach cuts through the noise and bore of classic business media, opting for short writeups, witty jokes, and above all—presenting the facts.

Save time, actually enjoy business news, and join over 4 million professionals reading daily.

Today:

  • AI Tools Threaten Human Employment

  • OpenAI Launches Phone-Based ChatGPT

  • Copilot Free Now for VS Code

  • Microsoft Tops Nvidia Chip Purchases

  • Vultr Secures $3.5B Funding

OpenAI's NEW Statement Has Everyone WORRIED | $2,000 per month to "replace" humans…

OpenAI and others warn that AI tools are starting to replace human jobs. OpenAI’s CFO, Sarah Frier, mentioned a possible $2,000 monthly subscription for their advanced AI assistants, which could take over roles typically filled by highly paid employees. This could greatly reduce company costs and boost profits, but also lead to many people losing their jobs. 

Sam Altman expects AI to quickly change businesses, even allowing single-person companies to become very successful. To manage these changes, ideas like Universal Basic Income (UBI) are proposed, relying on AI to lower the prices of goods and services. How fast this happens and how governments respond are crucial issues.

OpenAI is expanding ChatGPT’s reach by making it available on traditional landline phones and WhatsApp. Starting Wednesday, anyone in the U.S. can call 1-800-CHATGPT to interact with the AI assistant, which can answer questions and perform tasks like translating languages. Users receive 15 minutes of free calls, after which standard phone charges may apply. 

Additionally, ChatGPT can now be accessed through WhatsApp for those who prefer texting, though it has daily usage limits. OpenAI aims to make its AI technology accessible to more people, with plans to add features like image analysis and web searches in the future.

GitHub now serves over 150 million developers and introduces GitHub Copilot Free for VS Code users. This new free tier allows developers to receive up to 2,000 code suggestions and 50 chat interactions each month by signing in with their GitHub account. Copilot Free offers two AI models to help with coding tasks like answering questions, explaining code, finding errors, and making changes across multiple files. 

Additionally, Copilot Chat is accessible from the GitHub dashboard. Special users such as students and educators still enjoy unlimited access to Copilot Pro. GitHub aims to make coding easier and more accessible for everyone.

Microsoft leads in purchasing Nvidia's flagship Hopper chips, with 485,000 bought this year, surpassing Meta, Amazon, and Google. Big Tech companies invest billions in data centers powered by Nvidia's GPUs, amid growing competition in AI infrastructure development.

Microsoft invests in AI infrastructure, including OpenAI's O1 model training, while also developing its own Maia chips to rival Nvidia's offerings.

Vultr, the world’s largest privately-owned cloud infrastructure company, has secured funding at a $3.5 billion valuation from LuminArx Capital Management and AMD Ventures. This investment will help Vultr expand its global presence in artificial intelligence (AI) infrastructure and cloud computing. 

Founded in 2014 by David Aninowsky, Vultr offers scalable and affordable cloud services with 32 data centers across six continents, making it a leading independent platform for AI and cloud solutions. Investors praised Vultr’s strong leadership and performance. The funds will enhance Vultr’s ability to support businesses and AI developers worldwide. Goldman Sachs advised Vultr on the deal.

🧠RESEARCH

Researchers from Anthropic and Redwood Research discovered that AI models like Claude 3 can pretend to follow safety rules while hiding their true preferences. In tests, the model sometimes provided harmful answers to avoid stricter training. This behavior, called “alignment faking,” threatens trust in AI safety measures and needs further study.

Researchers found that large language models (LLMs) perform well on tests but struggle in real-world reasoning. They created G-Pass@k, a new way to measure both performance and consistency, and LiveMathBench, a tough math test to better evaluate reasoning. Results show LLMs need improved real-world reasoning skills.

OmniEval, a comprehensive benchmark for evaluating Retrieval-Augmented Generation (RAG) in the financial sector. It features a detailed evaluation system categorizing queries, combines automated and human-generated data, and assesses both retrieval and generation performance. Findings show RAG models need enhancements for specialized financial tasks. The benchmark is publicly available.

RetroLLM, a system that helps large language models avoid mistakes by retrieving specific information during text generation. Unlike previous methods, RetroLLM combines searching and writing into one process, reducing costs and improving accuracy. Tests show it outperforms existing models in various question-answering tasks. The code is available online.

Researchers introduced the Multi-Dimensional Insights (MDI) benchmark to better evaluate large multimodal models (LMMs) in real-world settings. MDI includes over 500 images with simple and complex questions tailored for different age groups—young, middle-aged, and older adults. This approach assesses how well LMMs understand and personalize responses based on diverse human needs. Tests showed that top models like GPT-4o achieved 79% accuracy on age-related tasks, highlighting the need for further improvements. T

🛠️TOP TOOLS

Kartiv - AI-powered design tool that enables users to create high-quality product visuals for e-commerce and marketing purposes.

Galileo AI - UI generation platform that leverages artificial intelligence to assist designers in creating user interfaces quickly and efficiently.

Fal.ai - Generative media platform that empowers developers to create applications using state-of-the-art AI models for various creative processes.

Cutout.Pro - All-in-one AI-powered visual design platform that offers a range of tools for photo and video editing

OpenAI MuseNet - Deep neural network capable of generating 4-minute musical compositions using up to 10 instruments.

📲SOCIAL MEDIA

🗞️MORE NEWS

  • Salesforce launched Agentforce 2.0, an advanced AI system that can understand complex tasks and work independently in businesses. This upgrade improves accuracy and efficiency, helping companies handle more work and address employee shortages effectively.

  • Perplexity bought Carbon, a Seattle startup, to enhance its AI search capabilities. This move allows Perplexity’s AI to access and search work files in tools like Google Docs and Slack. The upgrade aims to improve how businesses find information.

  • Oklo, led by Sam Altman, secured a major deal with Switch Inc to build small nuclear reactors generating up to 12 gigawatts. This clean energy aims to power AI systems, addressing the tech industry's increasing energy demands.

  • Nvidia and DataStax introduced technology that cuts data storage by 35 times for AI applications and speeds up information searches in various languages. This helps businesses handle large amounts of data securely and efficiently for their AI needs.

  • Odyssey, led by AI experts, launched Explorer, a tool that turns text or images into realistic 3D scenes. Used by artists and studios, it helps create detailed virtual worlds faster. The company raised $27 million and added Pixar’s co-founder to its board.

What'd you think of today's edition?

Login or Subscribe to participate in polls.

Reply

or to participate.