Machine Learning Research
When Agents Train Algorithms: OpenAI’s MLE-bench tests AI coding agents
Coding agents are improving, but can they tackle machine learning tasks?
Latest
Tech & Society
Does Your Model Comply With the AI Act?: COMPL-AI study measures LLMs’ compliance with EU’s AI act
A new study suggests that leading AI models may meet the requirements of the European Union’s AI Act in some areas, but probably not in others.
Machine Learning Research
Claude Controls Computers: Anthropic empowers Claude Sonnet 3.5 to operate desktop apps, but cautions remain
API commands for Claude Sonnet 3.5 enable Anthropic’s large language model to operate desktop apps much like humans do. Be cautious, though: It’s a work in progress.
Data Points
Python overtakes JavaScript as top programming language on GitHub: Robots finally have enough training data to fold your laundry
ChatGPT now includes an AI search engine. Upgrading data centers leaves behind too much trash. OmniParser works with vision models to read computer screens. Gallup poll shows most big companies’ workforces haven’t embraced AI.
Data Points
What counts as an open source AI model?: Claude, Gemini, and o1 models added to GitHub Copilot
Researchers find hallucinations in Whisper transcriptions. White House directs how security agencies should use AI. Jailbreaking LLM-enabled robots could have devastating effects. Google’s synthetic podcast tool Illuminate specializes in research papers.
Machine Learning Research
Synthetic Data Distorts Models: Could training on generated output doom AI’s future?
Training successive neural networks on the outputs of previous networks gradually degrades performance. Will future models succumb to the curse of recursive training?
Machine Learning Research
Benchmark Tests Are Meaningless: The problem with training data contamination in machine learning
The universe of web pages includes correct answers to common questions that are used to test large language models. How can we evaluate new models if they’ve studied the answers before we give them the test?
Tech & Society
No Work for Coders: Could coding assistants take over software development?
AI coding assistants are brewing codebases that once were the sole province of human programmers. Will AI systems take over software development?
Tech & Society
Innovation Can’t Win: Bureaucracy chokes AI growth as lawmakers tighten grip
Politicians and pundits have conjured visions of doom to convince lawmakers to clamp down on AI. What if terrified legislators choke off innovation in AI?
Tech & Society
AI Burns All the Energy: Will AI’s growing power demands drain the grid?
The globe’s growing AI infrastructure requires huge amounts of electricity, possibly more than power providers can generate responsibly. Could AI models suck energy resources dry?
Letters
The Psychology of AI Doom: Why do people who are well informed about AI worry about unrealistic dangers? There are incentives for doomsaying.
Welcome to our special Halloween issue of The Batch, in which we probe fears, anomalies, and shadows of AI.
The Batch Newsletter
Trick or treat! AI Devours Energy, Innovation Can’t Win, Models Collapse, Benchmark Tests Are Meaningless, No Work for Coders
The Batch AI News and Insights: Welcome to our special Halloween issue of The Batch, in which we probe fears, anomalies, and shadows of AI.