Hacker Papers

1. 1026
   

   CvT: Introducing Convolutions to Vision Transformers

   Haiping Wu et al. · 2021

   Hacker News threads

   • Mar 2022Transformers, originally designed to handle language, are taking on vision
   • Jul 2022
   • Dec 2025The Illustrated Transformer

   arXiv ↗
2. 1027
   

   Escaping the Big Data Paradigm with Compact Transformers

   Ali Hassani et al. · 2021

   Hacker News threads

   • Mar 2022Transformers, originally designed to handle language, are taking on vision
   • Dec 2023Deep Dive into the Vision Transformers Paper (ViT)
   • Jul 2025FastVLM: Efficient Vision Encoding for Vision Language Models

   arXiv ↗
3. 1028
   

   Addressing "Documentation Debt" in Machine Learning Research: A Retrospective Datasheet for BookCorpus

   Jack Bandy, Nicholas Vincent · 2021

   Hacker News threads

   • Feb 2023Ask HN: Does Google use the text inside gdocs and Gmail for training AI models?
   • Jul 2023The Pile: a dataset for language modeling [pdf]
   • Sep 2023

   arXiv ↗
4. 1029
   

   Pattern-defeating Quicksort

   Orson R. L. Peters · 2021

   Hacker News threads

   • Jul 2021Do Low-Level Optimizations Matter? (2020)
   • May 2023Ask HN: What is new in Algorithms / Data Structures these days?
   • Jun 202310~17x faster than what? A performance analysis of Intel x86-SIMD-sort (AVX-512)

   arXiv ↗
5. 1030
   

   TruthfulQA: Measuring How Models Mimic Human Falsehoods

   Stephanie Lin et al. · 2021

   Hacker News threads

   • Oct 2021Dublin Core's dirty little secret (2010)
   • Feb 2023ArxivGPT: Chrome extension that summarizes arxived research papers using ChatGPT
   • Aug 2026If this is true, the hyperscalers are toast

   arXiv ↗
6. 1031
   

   Efficiently Modeling Long Sequences with Structured State Spaces

   Albert Gu et al. · 2021

   Hacker News threads

   • Sep 2023Ask HN: Transformer alternatives that could have emergent properties when scaled
   • Feb 2024A decoder-only foundation model for time-series forecasting
   • Feb 2025The FFT Strikes Back: An Efficient Alternative to Self-Attention

   arXiv ↗
7. 1032
   

   Hyper-Fast Positive Energy Warp Drives

   Erik W. Lentz · 2021

   Hacker News threads

   • Nov 2022Physical Warp Drives (2021)
   • Oct 2023Supernova in 1006
   • Jun 2024A Study Says Warp Drives Might Be Real–and We'll Find Them with Lasers

   arXiv ↗
8. 1033
   

   Deep Symbolic Regression for Recurrent Sequences

   Stéphane d'Ascoli et al. · 2022

   Hacker News threads

   • Jan 2022Deep symbolic regression for recurrent sequences
   • Feb 2022I asked GPT-NeoX-20B a hundred arithmetic questions
   • Mar 2022GPT-3 can run code

   arXiv ↗
9. 1034
   

   Memory-assisted prompt editing to improve GPT-3 after deployment

   Aman Madaan et al. · 2022

   Hacker News threads

   • Mar 2022GPT-3 can run code
   • Dec 2022Building a Virtual Machine Inside ChatGPT
   • Jan 2023What Is AGI-Hard

   arXiv ↗
10. 1035
    

    Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time

    Mitchell Wortsman et al. · 2022

    Hacker News threads

    • Dec 2024Beyond Gradient Averaging in Parallel Optimization
    • Mar 2026NanoGPT Slowrun: Language Modeling with Limited Data, Infinite Compute
    • Jun 2026Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model
    • Jun 2026Rio de Janeiro's "homegrown" LLM appears to be a merge of an existing model

    arXiv ↗
11. 1036
    

    Self-Consistency Improves Chain of Thought Reasoning in Language Models

    Xuezhi Wang et al. · 2022

    Hacker News threads

    • Mar 2022GPT-3 can run code
    • Sep 2023A GPT-4 capability forecasting challenge
    • Feb 2026The path to ubiquitous AI (17k tokens/sec)

    arXiv ↗
12. 1037
    

    Flamingo: a Visual Language Model for Few-Shot Learning

    Jean-Baptiste Alayrac et al. · 2022

    Hacker News threads

    • May 2022DeepMind: A Generalist Agent
    • Feb 2024The killer app of Gemini Pro 1.5 is using video as an input
    • Oct 2025OpenTSLM: Language models that understand time series

    arXiv ↗
13. 1038
    

    Suppressing quantum errors by scaling a surface code logical qubit

    Rajeev Acharya et al. · 2022

    Hacker News threads

    • Jul 2022A bird's-eye view of quantum error correction and fault tolerance
    • Dec 2023IBM releases first-ever 1k-qubit quantum chip
    • Apr 2026Securing Elliptic Curve Cryptocurrencies Against Quantum Vulnerabilities [pdf]

    arXiv ↗
14. 1039
    

    Game-theoretic statistics and safe anytime-valid inference

    Aaditya Ramdas et al. · 2022

    Hacker News threads

    • Jun 2023Annoying A/B testing mistakes every engineer should know
    • Aug 2024Ask HN: Resources about math behind A/B testing
    • Jan 2025Lies. Damned Lies. P-value thresholds

    arXiv ↗
15. 1040
    

    Flow Matching for Generative Modeling

    Yaron Lipman et al. · 2022

    Hacker News threads

    • May 2025Veo 3 and Imagen 4, and a new tool for filmmaking called Flow
    • Jun 2025Normalizing Flows Are Capable Generative Models
    • Jul 2026Show HN: Flow Matching model inference in C

    arXiv ↗
16. 1041
    

    Efficiently Scaling Transformer Inference

    Reiner Pope et al. · 2022

    Hacker News threads

    • Jun 2023AudioPaLM: A Large Language Model That Can Speak and Listen
    • Jul 2023Chiplet ASIC supercomputers for LLMs like GPT-4
    • Feb 2024Groq runs Mixtral 8x7B-32k with 500 T/s

    arXiv ↗
17. 1042
    

    Scaling Laws for Generative Mixed-Modal Language Models

    Armen Aghajanyan et al. · 2023

    Hacker News threads

    • Feb 2023Planning for AGI and Beyond
    • Mar 2023Google Releases Paper on Embodied LLM (562b parameters)
    • Apr 2025Liquid: Language models are scalable and unified multi-modal generators

    arXiv ↗
18. 1043
    

    Tracr: Compiled Transformers as a Laboratory for Interpretability

    David Lindner et al. · 2023

    Hacker News threads

    • Feb 2023Why Can't Little Chatty Do Math?
    • Sep 2023I made a transformer by hand (no training)
    • Dec 2023Phi-2: The surprising power of small language models

    arXiv ↗
19. 1044
    

    G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment

    Yang Liu et al. · 2023

    Hacker News threads

    • Jun 2023The Darwinian argument for worrying about AI
    • Dec 2024Garvata: Observability and Debugging for AI Agent Stack
    • May 2025Positional preferences, order effects, prompt sensitivity undermine AI judgments

    arXiv ↗
20. 1045
    

    Visual Instruction Tuning

    Haotian Liu et al. · 2023

    Hacker News threads

    • Dec 2023Vowels and Diphthongs in Sperm Whales
    • Apr 2024Suggest HN: Rabbit r1 AI device for blind or visually impaired people?
    • May 2024PaliGemma: Open-Source Multimodal Model by Google

    arXiv ↗
21. 1046
    

    Simple and Controllable Music Generation

    Jade Copet et al. · 2023

    Hacker News threads

    • Jun 2023MusicGen-Looper: Generate fixed-bpm loops from text prompts
    • Mar 2025Show HN: Sonauto API – Generative music for developers
    • Oct 2025Should LLMs just treat text content as an image?

    arXiv ↗
22. 1047
    

    Efficient Guided Generation for Large Language Models

    Brandon T. Willard, Rémi Louf · 2023

    Hacker News threads

    • Aug 2023Show HN: LLMs can generate valid JSON 100% of the time
    • May 2024Mistral Fine-Tune
    • Jan 2026LLM Structured Outputs Handbook

    arXiv ↗
23. 1048
    

    Emergent Linear Representations in World Models of Self-Supervised Sequence Models

    Neel Nanda et al. · 2023

    Hacker News threads

    • Jul 2024Large models of what? Mistaking engineering achievements for linguistic agency
    • Nov 2024Evaluating the world model implicit in a generative model
    • Dec 2024Alignment faking in large language models

    arXiv ↗
24. 1049
    

    The Benefits of a Concise Chain of Thought on Problem-Solving in Large Language Models

    Matthew Renze, Erhan Guven · 2024

    Hacker News threads

    • Mar 2025Sketch-of-Thought: Efficient LLM Reasoning
    • Apr 2026Talk like caveman
    • Jul 2026Autistic Caveman: Why use many token when correct token do trick

    arXiv ↗
25. 1050
    

    LLMs Can't Plan, But Can Help Planning in LLM-Modulo Frameworks

    Subbarao Kambhampati et al. · 2024

    Hacker News threads

    • Oct 2024Use Prolog to improve LLM's reasoning
    • Apr 2025Chain of Recursive Thoughts: Make AI think harder by making it argue with itself
    • Dec 2025If you're going to vibe code, why not do it in C?

    arXiv ↗

Data from the Hacker News archive (ODC-By), arXiv and OpenAlex. Not affiliated with Hacker News or Y Combinator.