Plus: Are Chokepoints the only way to control AI? - ChatGPT has been Updated and is allegedly “less Lazy.”  ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌

	Forwarded this email? Subscribe here for more 💼 Replacing Human Workers with AI is more Expensive. Plus: Are Chokepoints the only way to control AI? - ChatGPT has been Updated and is allegedly “less Lazy.” Yaro Celis Jan 26   READ IN APP   Sign Up - Advertise Let’s dive in… ChatGPT has been Updated and is allegedly “less Lazy.” Replacing Human Workers with AI is more Expensive. Are Chokepoints the only way to control AI? Our Sections: 📰 News and Trends 🤔 What Does that even Mean? Term of the Day (Overfitting) 💰 Follow the Money 🧰 AI Tools of the Day (Recruiting and Hiring) 📰 AI News and Trends AI Program Finds Thousands of Possible Psychedelics. Will They Lead to New Drugs? (SciAM) Meet the Mind Reading AI hat (DesignBoom) AI-Generated Taylor Swift Porn Went Viral on Twitter (404) Hugging Face has partnered with Google Cloud to give easy access to models and computing for customers of both platforms (HF) 🌐 Other Tech news I knew I could 3d print crocs like shoes (BOF) Microsoft has announced “Mesh” a virtual meeting platform that lets employees' avatars meet in the same place, even if they are spread out. (Engadget) Ultraviolet light can kill almost all the viruses in a room. Why isn’t it everywhere? (Vox) Elon Musk says Chinese electric car brands could ‘demolish’ rivals as Tesla earnings fall (CNN) MIT has announced a new method of 3D printing metal that prioritizes printing speeds and scale over resolution (TC) Apple is bringing sideloading and alternate app stores to the iPhone (TheVerge) ChatGPT has been Updated and is allegedly “less Lazy.” OpenAI's GPT-4 Turbo addresses 'laziness' in task completion with an update trained on data up to April 2023. Over 70% of GPT-4 API users have migrated to this version for its improved capabilities. Upcoming features include vision capabilities for multimodal interactions. Additionally, OpenAI introduced 'embeddings' models—'text-embedding-3-small' and 'text-embedding-3-large'—to enhance AI's understanding of content relationships, marking significant progress in AI technology. New Embedding Models: OpenAI introduces two new models, text-embedding-3-small and text-embedding-3-large, with improved performance and lower pricing. GPT-3.5 and GPT-4 Turbo Updates: A new GPT-3.5 Turbo version is coming with significantly reduced prices. GPT-4 Turbo has been updated for better task completion. Future Vision Capabilities: Plans to launch GPT-4 Turbo with vision features. Updated Moderation Model: The launch of text-moderation-007 for better text safety analysis. API Management Enhancements: New tools for better control and insights into API usage. Leave a comment The US Data Privacy Checklist Scale your consumer data privacy program with ease. ‍Vanta’s exclusive US Data Privacy (USDP) framework is all you need to comply with state-level consumer privacy laws. By covering all state privacy laws with one comprehensive set of controls, this unlocks a fast, scalable way to keep you ahead of ever changing privacy laws across the US. Do the work once and ensure your business simultaneously adheres to CCPA/CPRA, CPA, CTDPA, UCPA, and VCDPA. The framework is constantly being updated to keep up with new and evolving regulations. Download the US Data Privacy checklist to learn more. Replacing Human Workers with AI is more Expensive MIT researchers found that replacing human workers with AI in visual inspection tasks is often more expensive than keeping human labor. Their study across 800 occupations showed only 23% of tasks are cost-effective to automate due to high AI system costs. This finding suggests that AI-driven job displacement will be significant but gradual, allowing time for policy development and workforce retraining. Despite the rapid growth of AI technologies like ChatGPT and Bard, and recent tech sector layoffs, experts believe this shift could lead to new types of jobs. The study, supported by IBM's Watson AI Lab, highlights that while automation may become more economically feasible over time, it currently remains limited and requires thoughtful regulatory preparation. Are Chokepoints the only way to control AI? In the realm of AI development, researchers are exploring the concept of embedding restrictions within crucial components like GPUs to regulate AI's power and prevent dangerous applications. This approach, which ties AI's capabilities to its hardware, aims to offer a more enforceable solution than conventional laws or treaties. Modern devices already employ similar security measures, like iPhones' secure enclaves for biometric data. The proposal, while theoretically sound, faces significant challenges in technical implementation and political acceptance. It draws parallels to nuclear nonproliferation methods, suggesting a new frontier in AI governance. However, the tech industry has shown resistance to such hardware-based controls, citing ethical and complexity issues. The U.S. government has expressed interest, indicating a potential shift in how AI capabilities might be regulated in the future. 💰 Follow The Money $44M raised by Bastille, a CA-based supplier of wireless threat intelligence technology to the high-tech, banking, and intelligence community. $28M obtained by Vector database Qdrant $14M amassed by Cultivo, a CA-based public benefit corporation that provides an AI-driven science and technology platform. $6.5M raised by bluesheets, a Singapore-based AI automation software company. $3.3M collected by Krepling, an e-commerce channel management platform. $2.5M obtained by ViralMoment, a CA-based provider of AI-powered social video insights and analytics solutions. 🤔 What Does that even Mean? Term of the Day Overfitting: A modeling error in machine learning that occurs when a function is too closely fit to a limited set of data points. Example: A weather prediction model is trained with a very specific set of data from a short period. It performs exceptionally well on this training data but fails to accurately predict weather when given data from a different time frame or location. This is because it has learned the peculiarities of the training data too well, at the expense of its ability to generalize to new data. 📰 Publications I am currently reading and recommending: Tim Ferris - Get the 5 Things I've Been Loving, Using, and Reading: Books, Gadgets, Hacks, and More. 1-Minute Question - Based on Neuroscience Science Practices to rewire your brain for success. How Brands Win - Distills the strategies behind great brands. And gives you actionable insights for your business. 🧰 AI Tools of the Day Recruiting and Hiring Swiftly AI - Specify role requirements to rank candidates by criteria, and filter through hundreds of applicants in minutes. Mesh - Screens candidates for any type of role through text messages, then provides insights and recommendations. Guide Track - Asks smart follow-up questions to applicants, scores submissions from 1–10, summarises them, and zips everything to an Airtable sheet. Skill Pool - State-of-the-art AI hiring technology to identify your perfect candidates. Download over 300+ Tools free here. 🚀 Showcase Your Innovation in the Premier Tech and AI Newsletter (link) As a vanguard in the realm of technology and artificial intelligence, we pride ourselves in delivering cutting-edge insights, AI tools, and in-depth coverage of emerging technologies to over 55,000+ tech CEOs, managers, programmers, entrepreneurs, and enthusiasts. Our readers represent the brightest minds from industry giants such as Tesla, OpenAI, Samsung, IBM, NVIDIA, and countless others. Reserve a Spot Explore sponsorship possibilities and elevate your brand's presence in the world of tech and AI. Learn more about partnering with us. You’re a free subscriber to Yaro’s Newsletter. For the full experience, become a paying subscriber. Subscribe for $1/Month Disclaimer: We do not give financial advice. Everything we share is the result of our research and our opinions. Please do your own research and make conscious decisions.   Like Comment Restack   © 2024 Yaro Celis 548 Market Street PMB 72296, San Francisco, CA 94104 Unsubscribe