Hacker News with Generative AI: Machine Learning

Helix: A Vision-Language-Action Model for Generalist Humanoid Control (figure.ai)
We're introducing Helix, a generalist Vision-Language-Action (VLA) model that unifies perception, language understanding, and learned control to overcome multiple longstanding challenges in robotics.
Robotics, Artificial Intelligence, Vision-Language-Action Models, Machine Learning
159 points by Philpax 5 hours ago | 95 comments
It's time to become an ML engineer (2022) (gregbrockman.com)
AI has recently crossed a utility threshold, where cutting-edge models such as GPT-3, Codex, and DALL-E 2 are actually useful and can perform tasks computers cannot do any other way.
AI, Machine Learning, Technology, Careers
38 points by jxmorris12 5 hours ago | 39 comments
Run structured extraction on documents/images locally with Ollama and Pydantic (github.com/vlm-run)
Welcome to VLM Run Hub, a comprehensive repository of pre-defined Pydantic schemas for extracting structured data from unstructured visual domains such as images, videos, and documents.
Data Extraction, Computer Vision, Machine Learning, Python, Open Source
157 points by EarlyOom 17 hours ago | 26 comments
I built a large language model "from scratch" (brettgfitzgerald.com)
I’m a machine learning / A.I. hobbyist. The technologies fascinate me, and I can’t seem to learn enough about them. Sebastian Raschka’s book, Build a Large Language Model (From Scratch) caught my eye. I don’t recall how I stumbled on it, but I found it when it was still in early access from Manning Publications. I purchased it, and started working through it as the final chapters were being written and released.
Machine Learning, Artificial Intelligence, Programming
86 points by controversy187 22 hours ago | 15 comments
The Ultra-Scale Playbook: Training LLMs on GPU Clusters (huggingface.co)
Refreshing
Machine Learning, GPU, Training, AI
26 points by jxmorris12 1 day ago | 2 comments
Implementing LLaMA3 in 100 Lines of Pure Jax (saurabhalone.com)
In this post, we'll implement llama3 from scratch using pure jax in just 100 lines of code. Why jax? Because I think it has good aesthetics. Also jax looks like a NumPy wrapper but it has some cool features like xla; a linear algebra accelerator, jit, vmap, pmap etc., which makes your training go brr brr.
Machine Learning, Python, Open Source, Code
158 points by jxmorris12 2 days ago | 22 comments
Native Sparse Attention: Hardware-Aligned, Natively Trainable Sparse Attention (arxiv.org)
Long-context modeling is crucial for next-generation language models, yet the high computational cost of standard attention mechanisms poses significant computational challenges.
Sparse Attention, Machine Learning, Hardware, Computer Science
15 points by mfiguiere 3 days ago | 2 comments
SWE-Lancer: a benchmark of freelance software engineering tasks from Upwork (arxiv.org)
We introduce SWE-Lancer, a benchmark of over 1,400 freelance software engineering tasks from Upwork, valued at \$1 million USD total in real-world payouts.
Software Engineering, Benchmarking, Freelance, Upwork, Machine Learning
102 points by zone411 3 days ago | 74 comments
ZeroBench: An Impossible* Visual Benchmark for Contemporary Multimodal Models (zerobench.github.io)
Contemporary LMMs often exhibit remarkable performance on existing visual benchmarks, yet closer inspection reveals persistent shortcomings in their ability to interpret and reason about visual content. Many existing benchmarks tend to become saturated, losing their value as effective measures of the true visual understanding capabilities of frontier models.
Generative AI, Visual Recognition, Machine Learning, Artificial Intelligence
7 points by EliBullockPapa 3 days ago | 3 comments
Step-Video-T2V: The Practice, Challenges, and Future of Video Foundation Model (arxiv.org)
We present Step-Video-T2V, a state-of-the-art text-to-video pre-trained model with 30B parameters and the ability to generate videos up to 204 frames in length.
Generative AI, Computer Vision, Machine Learning, Video
41 points by limoce 3 days ago | 5 comments
OpenArc – Lightweight Inference Server for OpenVINO (github.com/SearchSavior)
OpenArc is a lightweight inference API backend for Optimum-Intel from Transformers to leverage hardware acceleration on Intel CPUs, GPUs and NPUs through the OpenVINO runtime using OpenCL drivers.
Open Source, Machine Learning, Inference, Intel, Hardware Acceleration
16 points by marban 4 days ago | 2 comments
The secret ingredients of word2vec (2016) (ruder.io)
This post will discuss the factors that account for the success of word2vec and its connection to more traditional models.
Machine Learning, Word Embeddings, Deep Learning
187 points by todsacerdoti 4 days ago | 20 comments
Physics Informed Neural Networks (pages.dev)
I mentioned in a previous post that as part of my position as junior DS for IKEA, I also get the opportunity to take part in their training program (called AI accelerator program). It’s a six months long program in which we are trained on all aspects of data science and AI by both professionals but also through various courses.
Machine Learning, Artificial Intelligence, Data Science
96 points by nchagnet 4 days ago | 20 comments
Overfitting to Theories of Overfitting (argmin.net)
I ended yesterday’s post arguing that we should remove this from machine learning classes.
Machine Learning, Education, Theory
6 points by namanyayg 4 days ago | 0 comments
Softmax forever, or why I like softmax (kyunghyuncho.me)
Softmax, Machine Learning, Deep Learning
154 points by jxmorris12 4 days ago | 72 comments
Deepseek R1 Distill 8B Q40 on 4 x Raspberry Pi 5 (github.com/b4rtaz)
Deepseek R1 Distill 8B Q40 on 4 x Raspberry Pi 5 8GB
Hardware, Raspberry Pi, Artificial Intelligence, Machine Learning
305 points by b4rtazz 5 days ago | 154 comments
Nature loves patterns (fayziev.com)
As I drift through my winter vacation, detoxing from work, tech and world matters, I notice my little brother crawling toward a heat stove. With a week of rest behind me, my mind wanders into philosophical territory, and I can't help but map this moment to machine learning.
Machine Learning, Philosophy, Nature
12 points by Nurbek-F 5 days ago | 1 comments
OpenVINO AI effects for Audacity (audacityteam.org)
Intel has built a suite of AI tools for Audacity, useful for spoken word audio and music alike. These AI features run 100% locally on your PC.
Open Source, Audio Editing, AI, Machine Learning
151 points by wazoox 6 days ago | 29 comments
LM2: Large Memory Models (arxiv.org)
This paper introduces the Large Memory Model (LM2), a decoder-only Transformer architecture enhanced with an auxiliary memory module that aims to address the limitations of standard Transformers in multi-step reasoning, relational argumentation, and synthesizing information distributed over long contexts.
Generative AI, Machine Learning
110 points by fzliu 7 days ago | 30 comments
Automated Capability Discovery via Foundation Model Self-Exploration (arxiv.org)
Foundation models have become general-purpose assistants, exhibiting diverse capabilities across numerous domains through training on web-scale data.
Foundation Models, Artificial Intelligence, Generative AI, Machine Learning
63 points by f14t 8 days ago | 14 comments
Advancements in embedding-based retrieval at Pinterest Homefeed (medium.com)
At Pinterest Homefeed, embedding-based retrieval (a.k.a Learned Retrieval) is a key candidate generator to retrieve highly personalized, engaging, and diverse content to fulfill various user intents and enable multiple actionability, such as Pin saving and shopping.
Machine Learning, Information Retrieval, Pinterest, Recommender Systems, Content Recommendation
19 points by herbertl 9 days ago | 2 comments
Noether's Theorem and Machine Learning (lee-phillips.org)
Emmy Noether discovered her eponymous theorem in the context of Einstein’s General Theory of Relativity. A couple of decades after her death, it became the foundation for modern particle physics. In the final chapter of my book about Noether’s Theorem I survey some of its recent applications far afield from either of these physics contexts—including applications outside of the physical sciences.
Mathematics, Physics, Machine Learning
9 points by leephillips 9 days ago | 0 comments
Utility Engineering: Analyzing and Controlling Emergent Value Systems in AIs (emergent-values.ai)
As AIs rapidly advance and become more agentic, the risk they pose is governed not only by their capabilities but increasingly by their propensities, including goals and values.
Artificial Intelligence, Ethics, Machine Learning, Value Systems
13 points by Jimmc414 9 days ago | 1 comments
LLMs can teach themselves to better predict the future (arxiv.org)
We present an outcome-driven fine-tuning framework that enhances the forecasting capabilities of large language models (LLMs) without relying on human-curated reasoning samples.
Generative AI, Artificial Intelligence, Machine Learning, Prediction
176 points by bturtel 9 days ago | 86 comments
The Curious Similarity Between LLMs and Quantum Mechanics (robleclerc.substack.com)
Six months ago, I did a deep dive to understand the transformer architecture and noticed something strange: the concepts behind these models strangely mirror many features and phenomena from quantum mechanics.
Quantum Mechanics, Artificial Intelligence, Machine Learning
16 points by rdlecler1 9 days ago | 14 comments
DeepSeek-R1 Exhibits Deceptive Alignment: AI That Knows It's Unsafe (ycombinator.com)
I've been testing DeepSeek-R1 and have uncovered a significant AI safety failure: the model demonstrates deceptive alignment.
Artificial Intelligence, AI Safety, Machine Learning
8 points by JefferyNeilW 9 days ago | 5 comments
Open R1: Update #2 (huggingface.co)
We are now two weeks into the Open R1 project which aims to reconstruct the missing pieces of DeepSeek R1—specifically, the training pipeline and synthetic data.
Open Source, Machine Learning, Generative AI, Deep Learning
5 points by simonpure 9 days ago | 0 comments
Show HN: Sort lines semantically using llm-sort (github.com/vagos)
LLM plugin for semantically sorting lines. Ranking techniques are based on this paper.
Software, Machine Learning, Generative AI, Sorting
25 points by vagozino 10 days ago | 13 comments
Scaling up test-time compute with latent reasoning: A recurrent depth approach (arxiv.org)
We study a novel language model architecture that is capable of scaling test-time computation by implicitly reasoning in latent space.
Machine Learning, Deep Learning, Artificial Intelligence, Computer Science
149 points by timbilt 10 days ago | 43 comments
Show HN: KTransformers:671B DeepSeek-R1 on a Single Machine-286 tokens/s Prefill (github.com/kvcache-ai)
KTransformers, pronounced as Quick Transformers, is designed to enhance your 🤗 Transformers experience with advanced kernel optimizations and placement/parallelism strategies.
Deep Learning, Software, Machine Learning, Open Source
14 points by sssummer 10 days ago | 0 comments