Posts that commenters keep linking to, ranked by how many different people have cited them in Hacker News comments. 593 posts.
We conduct a randomized controlled trial to understand how early-2025 AI tools affect the productivity of experienced open-source developers working on their own repositories. Surprisingly, we find that when developers use AI tools, they take 19% longer than without—AI makes them slower.
Secure and private AI processing in the cloud poses a formidable new challenge. To support advanced features of Apple Intelligence with larger foundation models, we created Private Cloud Compute (PCC), a groundbreaking cloud intelligence system designed specifically for private AI processing. Built with custom Apple…
Facebook owner Meta’s dangerous algorithms and reckless pursuit of profit substantially contributed to the atrocities perpetrated by the Myanmar military against the Rohingya people in 2017, Amnesty International said in a new report published today.
Anthropic's latest interpretability research: a new microscope to understand Claude's internal mechanisms
Reproducibility is a bedrock of scientific progress. However, it’s remarkably difficult to get reproducible results out of large language models. For example, you might observe that asking ChatGPT the same question multiple times provides different results. This by itself is not surprising, since getting a result from…
We’ve trained a model called ChatGPT which interacts in a conversational way. The dialogue format makes it possible for ChatGPT to answer followup questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.
When we amplify the "Golden Gate Bridge" feature, Claude's responses focus on the bridge. For a short time, anyone can interact with this model.
The Waymo Driver autonomously navigates tens of thousands of rider-only miles across San Francisco, Phoenix, Los Angeles, and Austin daily. It can navigate common scenarios, like adhering to a crossing guard directing traffic, as well as more unique interactions like avoiding a swerving vehicle. As the Waymo Driver…
We identified coordinated distillation attacks using ~24,000 fraudulent accounts to extract Claude's capabilities.
This post presents WaveNet, a deep generative model of raw audio waveforms. We show that WaveNets are able to generate speech which mimics any human voice and which sounds more natural than the best existing Text-to-Speech systems, reducing the gap with human performance by over 50%.
We investigate the internal mechanisms used by Claude 3.5 Haiku — Anthropic's lightweight production model — in a variety of contexts, using our circuit tracing methodology.
OpenAI is a non-profit artificial intelligence research company. Our goal is to advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return. Since our research is free from financial obligations, we can better focus on a positive…
Track Claude Code's daily performance on SWE-Bench-Pro as we collect an Opus 5.5/high baseline.
We’ve created OpenAI LP, a new “capped-profit” company that allows us to rapidly increase our investments in compute and talent while including checks and balances to actualize our mission.
We’re introducing a neural network called CLIP which efficiently learns visual concepts from natural language supervision. CLIP can be applied to any visual classification benchmark by simply providing the names of the visual categories to be recognized, similar to the “zero-shot” capabilities of GPT-2 and GPT-3.
We’ve trained a neural network called DALL·E that creates images from text captions for a wide range of concepts expressible in natural language.
A technical report on three bugs that intermittently degraded Claude's responses: what happened, why fixes took time, and what we're changing.
We propose measuring AI performance in terms of the length of tasks AI agents can complete. We show that this metric has been consistently exponentially increasing over the past 6 years, with a doubling time of around 7 months. Extrapolating this trend predicts that, in under a decade, we will see AI agents that can…
We built software to mask gender in technical engineering interviews. See what we learned and why it matters.
Technical details on how we tested Claude Mythos Preview, a model strikingly capable at computer security tasks, and what we found.
An increasing number of the machine learning (ML) models we build at Apple each year are either partly or fully adopting the Transformer…
Get more out of Claude Code: tell Claude what it can't see, give it a check to run, and enforce the rules that can't have exceptions.
We’re introducing Jukebox, a neural net that generates music, including rudimentary singing, as raw audio in a variety of genres and artist styles. We’re releasing the model weights and code, along with a tool to explore the generated samples.
Compare AI model performance on AA-Omniscience v1.1: Knowledge and Hallucination Benchmark. A benchmark measuring factual recall and hallucination across various economically relevant domains.
We call on all AI labs to immediately pause for at least 6 months the training of AI systems more powerful than GPT-4.
We’ve trained a large-scale unsupervised language model which generates coherent paragraphs of text, achieves state-of-the-art performance on many language modeling benchmarks, and performs rudimentary reading comprehension, machine translation, question answering, and summarization—all without task-specific training.
Anthropic research on data poisoning attacks shows a few hundred malicious documents can "backdoor" large language models of any size.
Algorithm recovers speech from the vibrations of a potato-chip bag filmed through soundproof glass.
It turns out we've all been using MCP wrong. Most agents today use MCP by exposing the "tools" directly to the LLM. We tried something different: Convert the MCP tools into a TypeScript API, and then ask an LLM to write code that calls that API. The results are striking.
