Meta is Dropping the Biggest Llama Next Week
Intuit's CEO Blames Layoffs on AI
In today’s email:
🤑 How is OpenAI making $3.4B in ARR?
📈 Does Generative AI Facilitate Investor Trading? Evidence from ChatGPT Outages
📚 A beginner's guide to quantization
🧰 8 new AI-powered tools and resources. Make sure to check the online version for the full list of tools.
Meta is Dropping the Biggest Llama Next Week
Meta Platforms is set to release its largest open-source language model, Llama 3, on July 23, according to an insider report. The new model boasts an impressive 405 billion parameters and is multimodal, capable of processing both images and text, allowing for the generation of new images from combined inputs. This development marks a significant advancement from previous Llama models, which were limited to text generation.
Despite initial rumors that the model’s weights might not be made available, reports suggest that Meta has decided to release the model, including its weights, as open source. This decision follows alleged internal debates, particularly objections from Facebook co-founder Dustin Moskovitz. The release of the weights has sparked discussions around financial and safety concerns, as model training is costly and providing weights makes the model more accessible, raising potential safety issues.
Weights are crucial in AI models as they optimize predictions, making them essential for reproducibility, practical application, transparency, and comparability. Without weights, developers only have the model architecture, which is essentially an "empty shell." Training a model to obtain these weights is resource-intensive, but access to pre-trained weights allows developers to utilize and enhance advanced AI models without the need for substantial training capabilities, highlighting their importance in the open-source AI community.
AI-Powered Productivity: Transform Your Day
Revolutionize your workday with the power of ChatGPT! Dive into HubSpot’s highly anticipated guide to discover how AI can elevate your productivity and creativity.
Learn to automate tasks, enhance decision-making, and foster innovation, all through the capabilities of ChatGPT. Download the offer for actionable insights and tips on harnessing AI for a more efficient, impactful professional life.
Download Free GuideIntuit's CEO Just Said AI Is the Reason He's Laying Off 1,800 Employees
Intuit's CEO, Sasan Goodarzi, recently announced that the company will be laying off 1,800 employees, or 10 percent of its workforce. In a memo, Goodarzi explained that the layoffs are not about cutting costs but about reallocating resources to support critical areas and drive growth, particularly through investments in AI. He emphasized that AI is transforming industries at an unprecedented pace and that companies unprepared for this shift will fall behind.
Goodarzi stated that, despite the layoffs, Intuit plans to hire the same number of employees in roles related to engineering, product development, and customer-facing positions. This move aligns with the company's focus on leveraging AI to enhance customer experiences, increase monetization, and improve efficiency. However, the announcement has drawn criticism for its approach, as the layoffs affect employees who were performing their jobs well, highlighting the harsh impact on those who depend on their jobs for financial stability.
Critics argue that Goodarzi's memo lacks empathy and fails to acknowledge the real-life consequences for those losing their jobs. The decision to lay off employees while shifting to AI-driven roles is seen as a reflection of poor leadership. The memo underscores the broader trend of tech companies prioritizing AI integration, often at the expense of their workforce, raising questions about the ethical implications of such strategic shifts.
Microsoft’s new AI system ‘SpreadsheetLLM’ unlocks insights from spreadsheets, boosting enterprise productivity
Microsoft has unveiled "SpreadsheetLLM," an innovative AI model designed to work with spreadsheets, significantly enhancing enterprise productivity. This model addresses the complexities of spreadsheet formats by combining large language models (LLMs) with structured data, enabling efficient understanding and analysis of complex spreadsheet information. The key component, SheetCompressor, optimizes the AI's performance, achieving state-of-the-art accuracy. By encoding spreadsheet data to be comprehensible for LLMs, SpreadsheetLLM bridges the gap in current AI capabilities, allowing for better data analysis and decision-making.
The potential applications of SpreadsheetLLM are extensive, ranging from automating routine data analysis tasks to providing intelligent insights based on spreadsheet data. It enables users to query and manipulate spreadsheet data using plain English, which could democratize access to data insights and empower more individuals to make data-driven decisions. Additionally, SpreadsheetLLM could automate time-consuming tasks like data cleaning and aggregation, saving businesses time and resources and allowing employees to focus on more valuable activities.
Microsoft's release of SpreadsheetLLM aligns with its broader investment in enterprise AI technologies, following the introduction of Microsoft 365 Copilot and Copilot for Finance. These tools demonstrate Microsoft's commitment to transforming data handling in the enterprise sector. While the rise of AI presents challenges, including workforce adaptation and job displacement, the potential benefits of enhanced data analysis and decision-making are substantial. As SpreadsheetLLM transitions from research to practical use, it promises to revolutionize how businesses work with spreadsheets, driving innovation and creating value.
Other stuff
How is OpenAI making $3.4B in ARR?
Tammy Lovin · Sora Showcase
Does Generative AI Facilitate Investor Trading? Evidence from ChatGPT Outages
Honey, I shrunk the LLM! A beginner's guide to quantization – and testing it
Did you ever feel that ChatGPT is done generating your response and then suddenly a burst of tokens show up?
Can AI be Meaningfully Regulated, or is Regulation a Deceitful Fudge?
All your ChatGPT images in one place 🎉
You can now search for images, see their prompts, and download all images in one place.
Get the extension for FREEEditor's Pick
Garandor is developing invisible watermarking technology to disrupt deepfakes
Bird Buddy’s new AI feature lets people name and identify individual birds
Shadow - Bot-Free AI Autopilot for Post-Meeting Tasks
Sugar-Free: Food Scanner - Sugar detox app
Brick Photos - Turn your pics into bricks
Project Atlas Desktop - Generate no-code agents in natural language
Clean Inbox for Gmail™ - A privacy-oriented Gmail add on against marketing email
Mutahunter - Advanced LLM-Based Open Source Language-Agnostic Mutation Testing Tool
Unclassified 🌀
WFH Team - Work from anywhere in the world
Subscribe to The Intelligent Friend, the newsletter that gives you insights on AI's psychological and social aspects, based only on scientific papers.
