3
5
Apple Event – September 9 [video] (youtube.com)
1
AI sugar for regular programmers – the silent majority (medium.com/samiezkay)
3
Gemma 3-270M (huggingface.co)
1
Tribe: TRImodal Brain Encoder for whole-brain fMRI response prediction (arxiv.org)
10
GPT-5 System Prompt? (github.com/wyattwalls)
231
Live: GPT-5 (youtube.com)
1
K-State study finds odd twist in galaxy rotations (k-state.edu)
3
ARC-AGI Leaderboard (arcprize.org)
1
A new chapter for June: our founding team is joining Amplitude (june.so)
25
Muvera: Making multi-vector retrieval as fast as single-vector search (research.google)
6
Mistral Small 3.2 (24B-Instruct-2506) (huggingface.co)
2
Meta tried to buy Ilya Sutskever's $32B AI startup, now plans to hire its CEO (cnbc.com)
5
Mistral AI plans to raise $1B (maddyness.com)
1
Geometry Distributions (1zb.github.io)
2
Statement on Anthropic Model Availability (windsurf.com)
25
[flagged] 99% of AI Startups Will Be Dead by 2026 – Here's Why (skooloflife.medium.com)
15
Sam and Jony Introduce Io (openai.com)
11
Lyria 2 (deepmind.google)
1
A deep research agent with 10,000 sources per query (graphthem.com)
2
Perplexity nears second fundraising in six months at $14B valuation (ft.com)
2
Ask HN: Is there a service that offers Common Crawl as an API?
1
DeepSeek R2: 97.3% reduction in unit cost, release imminent + core list (jiuyangongshe.com)
2
Advancing AI systems through progress in perception, localization, and reasoning (meta.com)
1
Mistral: Classifier Factory (mistral.ai)
1
Arc-AGI 2025: A research review (lewish.io)
527
The Llama 4 herd (meta.com)
127
OpenAI releasing new open model in coming months, seeks community feedback (openai.com)
3
Arc Prize 2025 (arcprize.org)
3
SpacetimeDB: Multiplayer at the speed of light (github.com/clockworklabs)
99
ARC-AGI without pretraining (iliao2345.github.io)
1
Redis creator disappointed by Sonnet 3.7 for coding [video] (youtube.com)
2
Arc-AGI-2 Preview (arcprize.org)
3
OpenAI: Deep Research System Card (openai.com)
1
Couldn't create Bing search API resource (microsoft.com)
2
LSTM: The Comeback Story? [video] (youtube.com)
1
ChatGPT vs. DeepSeek: Crazy Chess [video] (youtube.com)
2
Interview with DeepSeek Founder: We're Done Following. It's Time to Lead (thechinaacademy.org)
2
Gemini 2.0 Flash Thinking model (aistudio.google.com)
4
An In-Depth Answer Engine (graphthem.com)
2
PaSa: An LLM Agent for Comprehensive Academic Paper Search (pasa-agent.ai)
3
Meta: Memory Layers at Scale (arxiv.org)
2
Nobel Minds 2024 [video] (youtube.com)
2
AI Startup Perplexity Closes Funding Round at $9B Value (bloomberg.com)
4
GPU Glossary (modal.com)
3
Ilyasut: Pre-training as we know it will end [video] (twitter.com/vincentweisser)
1
Deepmind: Project Mariner (deepmind.google)
2
12 Days of OpenAI: Day 3 [video] (youtube.com)
10
This Video is AI Generated! SORA Review [video] (youtube.com)
3
The EU-Inc Policy Proposal (eu-inc.org)
1
Transit App Launches GPS-Free Underground Subway Navigation System (twitter.com/heshiebee)
17
OpenAI and others seek new path to smarter AI as current methods hit limitations (reuters.com)
1
ProemNews – Add scientific context to any news article (proem.ai)
3
The Super Weight in Large Language Models (arxiv.org)
3
Golden Age of Building (ycombinator.com)
1
AMA with OpenAI's Sam Altman, Kevin Weil, Srinivas Narayanan, and Mark Chen (reddit.com)
1
Scaling Laws for Diffusion Transformers (arxiv.org)
2
Petavoxel-Scale Human Cortex Reconstructed at Nanoscale Resolution [pdf] (cloudfront.net)
1
What Kind of Pencil Do You Use? [pdf] (seths.blog)
1
Actix (Rust) vs. Zap (Zig) vs. Stdlib (Zig): Performance Benchmark in Kubernetes [video] (youtube.com)
1
No Priors Ep. 80 – With Andrej Karpathy from OpenAI and Tesla [video] (youtube.com)
17
Fast Multidimensional Matrix Multiplication on CPU from Scratch (2022) (siboehm.com)
1
Mistral.rs: Faster than llama.cpp on CUDA (github.com/ericlbuehler)
3
Zuck pod interview: Llama 3.1 405B, Meta's open-source AI beats GPT-4 (twitter.com/rowancheung)
2
Projected Language Models: A Large Model Pre-Segmented into Smaller Ones (openreview.net)
2
A Visual Guide to Quantization (maartengrootendorst.com)
2
Azure Llama 3.1 Benchmarks (github.com/azure)
1
Autodiff and Adjoints for Differentiable Physics [video] (youtube.com)
4
llama3.1-405B leaked at 3am the day before on 4chan (twitter.com/kalomaze)
1
Abstraction and Reasoning Corpus via Procedural Example Generation (arxiv.org)
2
Promus Ventures: DeepTech Growth Index vs Nasdaq, S&P (deeptechindex.com)
1
An Adaptive Stochastic Gradient Method with Non-Negative Gauss-Newton Stepsizes (arxiv.org)
2
APIGen: Automated Generation of Verifiable Function-Calling Datasets (apigen-pipeline.github.io)
15
Ladybird Announcement [video] (youtube.com)
2
IISc physicists find a new way to represent 'pi' (iisc.ac.in)
1
Scientists Just Discovered a New Formula for Pi Accidentally [video] (youtube.com)
10
Meta Large Language Model Compiler: Foundation Models of Compiler Optimization (meta.com)
4
OpenAI: Strategic Content Partnership with Time (openai.com)
26
Optimizing AI Inference at Character.ai (character.ai)
3
LiveBench: A Challenging, Contamination-Free LLM Benchmark (livebench.ai)
2
Francois Chollet – LLMs won't lead to AGI $1M Prize to find true solution [video] (youtube.com)
2
General Intelligence (2024) (nonint.com)
1
Build your 8-bit computer from scratch [video] (youtube.com)
322
Microsoft will switch off Recall by default after security backlash (wired.com)
2
A New Chapter for Bartender (surteesstudios.com)
2
Open AI: Scaling and evaluating sparse autoencoders [pdf] (cdn.openai.com)
1
Adaptive Horizon Actor-Critic for Contact-Rich Policy Learning (adaptive-horizon-actor-critic.github.io...
1
Meta FAIR: Contextual Position Encoding, Learning to Count What's Important (arxiv.org)
1
Google responds to leak: Documentation lacks context (searchengineland.com)
2
VALL-E 2: Achieves Human Parity in Zero-Shot TTS (microsoft.com)
1
VALL-E (microsoft.com)
1
Seal Leaderboards–private, expert evaluations of leading frontier models (scale.com)
2
Reproducing GPT-2 (124M) in llm.c in 90 minutes for $20 (github.com/karpathy)
2
14,000 Google Search Ranking Features Leaked (seroundtable.com)
6
Rabbit Gaslit Me, So I Dug Deeper [video] (youtube.com)
9
$30M AI Is Hiding a Scam [video] (youtube.com)
40
26× Faster Inference with Layer-Condensed KV Cache for Large Language Models (arxiv.org)
54
[flagged] Ex-OpenAI employee reported losing 85% of his family's net worth (lesswrong.com)
1
Binary Embeddings: All the AI, 3.125% of the Fat (jina.ai)
4