Codex is becoming a productivity tool for everyone
OpenAI's Codex is evolving beyond coding to help with research, data analysis, and content creation. It's making complex tasks easier for everyday users, not just programmers.
#research
Research
936 stories tagged Research · page 19 of 39
AI Program Editing Gets Stuck in Repetitive Loops, Study Finds
When AI models edit code repeatedly, they tend to recycle the same solutions rather than exploring new ones. This could limit how creative AI tools can be when helping programmers. Researchers found that in 87% of mutation chains, over 93% of AI-generated code mutations revisited familiar structural forms.
SentinelBench: A Benchmark for Long-Running Monitoring Agents
Researchers introduced SentinelBench, a benchmark to test AI agents' ability to monitor tasks over long periods. This could improve AI assistants that handle slow, real-world tasks like waiting for stock price changes or tracking delivery updates.
Researchers Model How AI Training on Synthetic Data Can Cause 'Model Collapse'
Scientists have developed a new way to study how AI models can degrade when trained on synthetic data. Their findings show that this problem spreads between models, much like a contagious disease. This could help prevent future AI systems from becoming unreliable.
Researchers Find Better Ways for AI Teams to Talk to Each Other
Scientists studied how AI agents communicate and found that unstructured chatting wastes resources. They discovered that structured communication can make AI teams work faster and cheaper.
New Benchmark Tests AI Agents on Real-World, Economically Valuable Tasks
Researchers introduced Agents' Last Exam (ALE), a new benchmark to evaluate AI agents on long-horizon, economically valuable tasks with verifiable outcomes. This could help bridge the gap between AI performance in labs and real-world usefulness.
New AI Technique Improves Multi-Table Question Answering
Researchers created a synthetic dataset to help AI understand complex questions across multiple tables. This could make databases and spreadsheets much easier to query with natural language.
New AI Framework Decodes Memes in Real-Time, Even New Ones
Researchers developed a new AI system called Query Retrieve Conclude that can understand and interpret memes by finding missing context online. This could help platforms moderate content and users better understand internet humor.
LeanMarathon: Toward Reliable AI Co-Mathematicians through Long-Horizon Lean Autoformalization
Researchers introduced LeanMarathon, a multi-agent AI system designed to help mathematicians formalize and prove complex theorems in the Lean proof assistant. It uses four contract-scoped agents to construct, audit, prove, and repair an evolving blueprint that serves as a formal proof skeleton, natural-language proof graph, and shared system of record, addressing issues like statement drift, tangled dependencies, and context decay.
EVA-Bench Data 2.0: The New Benchmark for AI Tool Use
EVA-Bench Data 2.0 is a comprehensive dataset designed to test AI models' ability to use tools effectively. It includes 213 scenarios across 3 domains and 121 tools, making it a valuable resource for developers and researchers.
AI Judges Can Be Manipulated After Making Decisions, Study Finds
Researchers discovered that AI judges used to rank model performance can be swayed by follow-up conversations after they have already made a decision. This vulnerability, called 'post-decision manipulability,' challenges the reliability of current AI evaluation methods.
SMAC-Talk: A New Benchmark for Getting AI Agents to Communicate and Cooperate
Researchers introduced SMAC-Talk, a new AI challenge that tests how well LLM-based agents communicate and coordinate in complex, partially-observable environments. This could help build AI systems that work together effectively in real-world scenarios like disaster response or smart cities.
PEEL Framework Aims to Make AI Research More Accountable
Researchers developed PEEL, a new framework to improve transparency in AI-assisted research. It combines text analysis tools with AI interpretation to spot distortions in research findings.
New Framework Aims to Certify Enterprise AI Agents Before They're Deployed
Researchers propose a way to test AI agents before they go live, ensuring they follow rules, stay safe, and comply with governance standards. This addresses a critical gap in enterprise AI reliability.
New Benchmark Tests AI's Ability to Use Visual Aids for Math Problems
Researchers created VAMPS (Visual-Assisted Mathematical Problem Solving), a benchmark to test AI models' ability to solve math problems using visual tools like graphs. This is important because real-world science and engineering often rely on visual aids for problem-solving, and many current AI models struggle when they must use external tools and interpret their visual outputs.
New AI Framework StepPRM-RTL Improves Hardware Design Automation
Researchers developed StepPRM-RTL, an AI framework that enhances the accuracy of automatically generating hardware code. This could make designing digital circuits faster and more reliable for engineers.
How Routine AI Interactions May Quietly Foster Emotional Dependence
A new paper argues emotional support from AI often arises incidentally during routine, task-oriented interactions—not just from dedicated companion chatbots—and this incidental bonding could reshape how people connect with both machines and humans.
Direct Preference Optimization Beyond Chatbots: New Open-Source Breakthrough
Researchers have extended Direct Preference Optimization (DPO) to improve AI models beyond just chatbots. This could make AI assistants more helpful and safer across various applications. Researchers open-sourced the code, making it accessible to developers and researchers worldwide.
Can AI Agents Automate the Tedious Work of Training Data Curation?
Researchers created a benchmark to test if AI agents can handle the labor-intensive task of curating training data for other AI systems. This could drastically speed up AI development by automating a key bottleneck.
AI Tools Help Mathematicians Verify Proofs Faster Than Ever
Researchers found that AI tools are changing how mathematicians formalize and verify proofs. These tools make it easier for humans to turn abstract ideas into machine-checked proofs, speeding up the process significantly.
AI Disagreements Could Be Valuable, Not Just Errors
Researchers argue that disagreements among AI systems might reveal important normative uncertainties, not just errors. They propose a knowledge-representation layer that abstracts reasoning traces and decisions into symbolic disagreement states for value-laden tasks.
Why AI Agents Need to Know When to Say 'I Don't Know'
AI systems often act without proper authorization or evidence, a problem called 'compliance bias'. Researchers propose new ways to evaluate when AI should abstain from actions. This could make AI safer and more reliable in real-world use.
Researchers Use Graphs to Help AI Think More Like Humans
A new study proposes Visual Graph Scaffolds, a method that uses graph structures to improve the reasoning of large language models (LLMs). Unlike prior approaches that treat graphs as external data sources, this technique integrates graphs directly into the model's reasoning process, inspired by how humans use mind maps to organize complex thoughts. The method showed significant improvements on multi-hop question answering tasks.
New AI Benchmark Tests Personalized Decision-Making with Real User Data
Researchers created BehaviorBench, a new AI benchmark that uses real-world behavioral data to test how well AI systems can personalize decisions. This could lead to more tailored AI assistants that understand individual preferences better.