Sunday May 4, 2025

AI's time-saving potential is counterbalanced by new tasks, CMU's TheAgentCompany benchmark shows AI agents handle only 24% of professional tasks autonomously, and JetBrains AI Assistant now supports third-party LLMs with ProxyAsLocalModel.

News

Time saved by AI offset by new work created, study suggests

A new study analyzing the Danish labor market in 2023 and 2024 found that the adoption of generative AI models like ChatGPT has had almost no significant impact on overall wages or employment, despite rapid adoption in some workplaces. The study revealed that while AI chatbots saved time for many workers, they also created new job tasks for 8.4 percent of workers, offsetting potential time savings and raising questions about who benefits from the efficiency.

N8n – Flexible AI workflow automation for technical teams

n8n is a workflow automation platform that allows technical teams to build and integrate multi-step AI agents with over 400 app integrations, providing flexibility and customization options. The platform offers a range of features, including a visual interface, coding capabilities, and collaboration tools, making it suitable for various use cases, from IT operations to sales and marketing automation.

Google Gemini has the worst LLM API

Google's Gemini models are at the forefront of AI technology, offering competitive coding and reasoning models, long-context models, and multimodal models, but they are hindered by a poor developer experience. The Gemini API is available in two places, Vertex AI and Google AI Studio, with differing functionality and documentation, making it difficult for developers to navigate and use the models effectively.

Google can train search AI with web content even with opt-out

Unusual activity has been detected from your computer network, and to continue, you must verify you're not a robot by clicking a box, ensuring your browser supports JavaScript and cookies. If you need help or have inquiries related to this message, you can contact the support team and provide the given reference ID for assistance.

'The Worst Internet-Research Ethics Violation I Have Ever Seen'

A secret AI experiment on Reddit sparked outrage after it was discovered that researchers had been posting AI-written comments on a popular subreddit, passing them off as human thoughts, in an effort to study persuasion. The experiment has been widely condemned as a serious ethics violation, with many internet researchers and Reddit users expressing concern and calling for accountability.

Research

Stop treating `AGI' as the north-star goal of AI research

The AI research community's focus on artificial general intelligence (AGI) hinders its ability to set effective goals, leading to six key obstacles that hinder productive goal setting. To avoid these traps, the community should prioritize specific engineering and societal goals, embrace pluralism, and foster innovation through greater inclusion of diverse disciplines and communities, ultimately moving away from AGI as its primary objective.

A Survey of AI Agent Protocols

The rapid deployment of large language models across various industries has been hindered by the lack of a standardized communication protocol, making it difficult for agents to work together and scale effectively. A unified protocol could enable smoother interactions, encourage collaboration, and facilitate collective intelligence, with key characteristics of next-generation protocols including adaptability, privacy preservation, and group-based interaction.

CMU TheAgentCompany: Benchmarking LLM Agents on Consequential Real World Tasks

Researchers have developed a benchmark called TheAgentCompany to evaluate the performance of AI agents in completing real-world professional tasks, such as browsing the web and writing code, in a simulated work environment. The results show that the most competitive AI agent can autonomously complete 24% of tasks, indicating that while AI can handle simpler tasks, more complex tasks remain beyond its capabilities.

The Algebra of Patterns (Extended Version)

Pattern matching in programming languages often uses a first-match semantics, which can limit the declarative nature of the programming model due to the importance of clause order. This paper proposes a solution to make order-independent pattern matching practical by introducing a boolean algebra of patterns and default clauses, allowing for more expressive and flexible pattern matching without sacrificing order-independence.

Topology of Musical Data

This paper explores the application of topological structures to musical data through various metrics, using both classical topology and modern data-driven techniques like persistent homology. The analyses successfully recover known musical structures, such as the circle of notes and the circle of fifths, and reveal interesting features in a variety of musical works, including folk music.

Code

Show HN: Use Third Party LLM API in JetBrains AI Assistant

The ProxyAsLocalModel project is a proxy server that allows users to access custom large language models (LLMs) from third-party APIs, such as OpenAI and Gemini, in JetBrains AI Assistant, which only supports local models from LM Studio and Ollama. The project supports multiple LLM APIs and can be run as a fat runnable jar or a GraalVM native image, making it cross-platform and fast to start.

Zero dolars vibe debugging menace

Cloi is a local, context-aware debugging agent that runs in your terminal, streamlining the debugging process by analyzing errors and applying fixes directly to your codebase while keeping your code and data private and secure. It is currently in beta development, offering features such as automated fixes, safe changes, and extensibility, with system requirements including 8GB RAM, 10GB+ free space, and Node.js 14+ or Python 3.6+.

You can now directly sync Postgres with Redis

Redfly.ai has developed a schema-agnostic caching system that allows users to synchronize their database with Redis, generating a data access layer that integrates data access code with caching, and is now making this technology available to the public through an open-source repository. The system aims to improve database performance, scalability, and cost by bypassing traditional disk-based database issues, and currently supports SQL Server, Postgres, Redis, Azure Search, and Azure Cloud, with plans to expand to other relational databases and public clouds.

Show HN: Mimir-AIP a Modular Open-Source AI Platform

The MIMIR Artificial Intelligence Platform is an open-source framework that connects inputs and outputs to AI systems with integrated storage capabilities, featuring a modular plugin system for easy customization and extension. The platform supports large language models, provides a flexible architecture, and aims to democratize access to AI technology as an open alternative to proprietary solutions like Palantir AIP.

Intel AI Playground: PC starter app for doing image creation, image and chatbot

There is no text to summarize. The input appears to be an error message indicating that a README file could not be retrieved.