1021
Typesafe AI raises $870M at $7.5B
TypeSafe AI is an AI lab building machine-native intelligence infrastructure for automation, designed to make decisions within software. Try our first System One Model, Jev, in early access.
Topics /
AI & Machine Learning
The most popular blog posts about ai & machine learning on Hacker News. 6,306 posts.
1022
The slow collapse of critical thinking in OSINT due to AI
This isn’t a rant against AI. I use it daily, ChatGPT, Copilot, Claude, Gemini. They’re in my workflow like everyone else’s. But the tradecraft is slipping. Analysts are skipping the hard parts. They’re trusting GenAI to do the heavy cognitive lifting, and it’s changing how we operate at a foundational level.
1023
Exploring 12M of the 2.3B images used to train Stable Diffusion
What images were used to train Stable Diffusion's text-to-image AI model? I worked with Simon Willison to make a data browser to explore and search a sample of 12 million images.
1024
Memory has grown to nearly two-thirds of AI chip component costs
High-bandwidth memory (HBM) accounts for 63% of AI chip component costs, up from 52% in Q1 2024. Epoch AI's breakdown of component cost shifts across major chip designers.
1025
The Weird Nerd comes with trade-offs
A metascience post of sorts that argues we should take human capital more seriously
1026
No elephants: Breakthroughs in image generation
When Language Models Learn to See and Create
1027
"A computer can never be held accountable"
This legendary page from an internal IBM training in 1979 could not be more appropriate for our new age of AI. A computer can never be held accountable Therefore a …
1028
Understanding is the new bottleneck
Agents can write code faster than we can absorb it. Here's why it still matters for humans to understand what they build — and some techniques for doing that efficiently: explainer docs, quizzes, micro-worlds, and shared spaces.
1029
Introducing Keras 2
Keras was released two years ago, in March 2015. It then proceeded to grow from one user to one hundred thousand.
1030
Decomposing language models into understandable components
Interpretability research decomposing a transformer layer into thousands of interpretable features that can steer model behavior.
1031
Go is an ideal language for AI-assisted software engineering
As AI shifts software engineering from writing to reviewing, discover how Go's strict compiler and unified toolchain ensure reliable AI-generated code.
1032
Mozilla.ai: Investing in Trustworthy AI
We’re committing $30M to build Mozilla.ai: A startup — and a community — building a trustworthy, independent, and open-source AI ecosystem. We’re o
1033
I Am An AI Hater
I am an AI hater. This is considered rude, but I do not care, because I am a hater.
1034
OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?
What is “the true cost of a token”?
1035
Controlling the nuclear fusion plasma in a tokamak with reinforcement learning
Successfully controlling the nuclear fusion plasma in a tokamak with deep reinforcement learning
1036
Accelerated PyTorch Training on M1 Mac
In collaboration with the Metal engineering team at Apple, we are excited to announce support for GPU-accelerated PyTorch training on Mac. Until now, PyTorch training on Mac only leveraged the...
1037
Classical music generation with recurrent neural networks
Writeup for my first major machine learning project.
1038
Street Fighter II's AI Engine
By today’s standards, the Artificial Intelligence in Street Fighter 2 World Warrior isn’t very sophisticated. These days, when most people talk about AI they’re talking about mach…
1039
PyTorch Internals: Ezyang's Blog
This post is a long form essay version of a talk about PyTorch internals, that I gave at the PyTorch NYC meetup on May 14, 2019.
1040
Born Against, or why hobby programming communities are against LLM usage
I came across a GH thread related to chess engine development that made me think of why hobby programming communities are increasingly hostile toward LLM development. While the thread doesn’t give a lot of insight into answering the question, it prompted me to think about it a bit. I’ve seen similar sentiments…
1041
1042
Making
TLDR: I gain a lot of fulfillment by making things. I don't consider things built by others at my request to be made by me, and are therefore much less fulfilling. And then I feel sad. This article starts strong and then heads off into the weeds.
1043
Gemma 3n preview: Mobile-first AI
Introducing Gemma 3n – the latest Google open model for accessible AI, featuring unique flexibility, privacy, and expanded multimodal capabilities on mobile devices.
1044
What are transformer models and how do they work?
Explore the fundamentals of transformer models, which have revolutionized natural language processing.
1045
Can LLMs learn from a single example?
We’ve noticed an unusual training pattern in fine-tuning LLMs. At first we thought it’s a bug, but now we think it shows LLMs can learn effectively from a single example.
1046
Phi-4: Microsoft's Newest Small Language Model Specializing in Complex Reasoning
Today we are introducing Phi-4, our 14B parameter state-of-the-art small language model (SLM) that excels at complex reasoning in areas such as math, in...
1047
Don't use cosine similarity carelessly
Cosine similarity - the duct tape of AI. Convenient but often misused. Let's find out how to use it better.
1048
Jeff Dean’s Lecture for YC AI [video]
Jeff Dean is a Google Senior Fellow in the Research Group, where he leads the Google Brain project. He spoke to the YC AI [https://ycombinator.wpengine.com/yc-ai/] group this summer. Here are his slides:
1049
EmbeddingGemma 2: An open, lightweight multimodal embedding model
Introducing EmbeddingGemma 2, an open multimodal embedding model optimized for privacy-first use cases
1050
ChatGPT Health
ChatGPT Health is a dedicated experience that securely connects your health data and apps, with privacy protections and a physician-informed design.