Top all-time · #649
Modal
Company blog · modal.com/blog · on HN as thundergolferaiprogramming
Modal is the company blog at modal.com, writing about ai and programming. Since 2022, 96 different Hacker News readers have linked it 236 times, most often “Modal – an end-to-end stack for cloud compute”. It is still writing: the last post was on 2026-10-23, and you can follow it by feed.
Cloud infrastructure for teams that develop, train, and serve AI applications at scale.
Most-linked posts?
- Modal – an end-to-end stack for cloud compute15 people · 16 mentions · first linked 2022-12-23
- Modal – Run code in the cloud without managing your own infrastructure10 people · 11 mentions · first linked 2022-10-20
- GPU Programming Glossary10 people · 10 mentions · first linked 2024-12-12
- Keeping 20k GPUs Healthy5 people · 8 mentions · first linked 2026-01-08
- The Missing Nvidia GPU Glossary5 people · 5 mentions · first linked 2024-12-12
- Linear Programming for Fun and Profit5 people · 5 mentions · first linked 2025-05-07
- How to Achieve Serverless GPUs3 people · 6 mentions · first linked 2026-05-12
- We built a cloud GPU notebook that boots in seconds3 people · 6 mentions · first linked 2025-09-16
- How to serve trillions of tokens for trillion-parameter coding agents3 people · 5 mentions · first linked 2026-09-23
- Speculation Is All You Need3 people · 5 mentions · first linked 2026-06-20
What Hacker News says?
Give https://modal.com a spin -- email me at deven [at] modal.com and happy to help you get set up
Comment linking /
on “We were wrong about GPUs”Submitted DoppelBot: Replace Your CEO with an LLM as a story · 232 points
HN discussion> What about alternatives? Custom models? Apart from the Big 3 (in no particular order): - https://together.ai/ - https://replicate.com/ - https://anyscale.com/ - https://baseten.co/ - https://modal.com/ - https://banana.dev/ - https://runpod.io/ - https://bentoml.com/ - https://brev.dev/ - https://octo.ai/ - https://cerebrium.ai/ ...
Comment linking /
on “Fly.io has GPUs now”Submitted The Missing Nvidia GPU Glossary as a story · 230 points
HN discussionI wonder if new serverless platforms, like Modal [0], are even simpler than monoliths. I've been reading some books on software engineering, and it seems clear that network calls, multiple services, integration points, etc. all cause complexity. However, I do wonder if things like Modal allow you to focus only on writing business logic. No dealing with deploys, network comms, memory, etc. (No, I don't work at Modal,…
Comment linking /
on “In Defense of Simple Architectures (2022)”Submitted 'I paid for the whole GPU, I am going to use the whole GPU' as a story · 154 points
HN discussionIt's not our only focus at Modal but it's a big focus![1] Code agents are the killer use case for LLMs right now, and this complements our GPU inference and training capabilities. I'm quietly betting that agents increase the leverage of deterministic, reproducible devbox tech (eg. Nix, lockfiles, package mirroring), and this will end up being a huge win for us human engineers too. 1. https://modal.com/use-cases/sand…
Comment linking /use-cases/sandboxes
on “A Research Preview of Codex”Submitted We reverse-engineered Flash Attention 4 as a story · 134 points
HN discussion> But I don't think that's what we're talking about here. Actually that is exactly what I'm talking about. I'm not talking about AI beginners putting in some words into a text box, I'm talking about creatives who use workflow managers like ComfyUI to create exactly the output they envision in their minds. In this way, the AI generation is merely a tool to get out whatever is in their head via synthesized means rathe…
Comment linking /blog/comfyui-custom-nodes
on “I Am An AI Hater”Submitted Lambda on hard mode: serverless HTTP in Rust as a story · 131 points
HN discussionWe had fraudsters using modal.com compute to solve this challenge. It's not traditional mining software so it didn't initially get flagged, but we've updated our detection to catch it now[1]. 1. https://modal.com/blog/catching-cryptominers
Comment linking /blog/catching-cryptominers
on “Bitcoin puzzle #66 was solved: 6.6 BTC (~$400k) withdrawn”Submitted Static IPs for Serverless Containers as a story · 125 points
HN discussion- https://modal.com/blog/vprox - https://modal.com/blog/host-overhead-inference-efficiency - https://modal.com/blog/resource-solver
Comment linking /blog/resource-solver
on “Ask HN: What are the best engineering blogs with real-world depth?”Submitted We built a cloud GPU notebook that boots in seconds as a story · 91 points
HN discussionWe use shot-scraper internally to automate keeping screenshots in our documentation up-to-date. Thanks for the tool![1] Agree that Playwright is great. It's super easy to run on Modal.[2] 1. https://modal.com/docs/guide/workspaces#dashboard 2. https://modal.com/docs/examples/web-scraper#a-simple-web-scr...
Comment linking /docs/examples/web-scraper
on “Web Scraping in Python – The Complete Guide”Submitted Linear Programming for Fun and Profit as a story · 62 points
HN discussionFor most people, before it makes sense to just buy all the hardware yourself, you probably should be renting GPUs by the hour from the various providers serving that need. On Modal, I think should cost about $72/hr to serve Kimi K2 https://modal.com/pricing Once that's running it can serve the needs of many users/clients simultaneously. It'd be too expensive and underutilized for almost any individual to use regular…
Comment linking /pricing
on “Kimi K2 is a state-of-the-art mixture-of-experts (MoE) language model”Submitted GPU Glossary as a story · 15 points
HN discussionModal's GPU glossary is a good overview about how GPUs work [0]. Karpathy's LLM overview is a good high level overview on LLMs [1]. 3b1b's video (and subsequent videos) on transformers was excellent at helping me understand the math at a high level [2]. This matrix multiplication optimization worklog helped me understand writing better CUDA (not for beginner intro though) [3]. During this process I also asked ChatGP…
Comment linking /gpu-glossary
on “Show HN: TokenDagger – A tokenizer faster than OpenAI's Tiktoken”Submitted GPU Programming Glossary as a story · 13 points
HN discussionYep, that’s right. Modal ( https://modal.com ) is actually what you’re describing with Python, but Modal is just the compute layer (Unison is going for the whole cloud), and Modal isn’t doing anything to handle version alignment across service boundaries
Comment linking /
on “Unison Cloud”Submitted Building Fast Filesystems for VMs as a story · 11 points
HN discussionThe GPU Glossary mentions that a warp scheduler can context switch https://modal.com/gpu-glossary/device-hardware/warp-schedule... but you said there is no such thing as an SM "context switch between threads". Is there some ambiguity in context switch
Comment linking /gpu-glossary/device-hardware/warp-scheduler
on “The Missing Nvidia GPU Glossary”Submitted Linear Programming for Fun and Profit: Finding Arbitrages in the GPU Market as a story · 11 points
HN discussionNot in love with axolotl but appreciate the advantages. This is an interesting approach, but you can also finetune easily on providers who wrap axolotl like Replicate [1], Modal [2], or if you want to run the infra, LLM Engine [3]. My only gripe with Helix would be that it's smaller than the above and my org would be peeved about data security. The ability to self host is cool, but too much can go wrong too quickly…
Comment linking /
on “How we got fine-tuning Mistral-7B to not suck”Submitted Lambda on hard mode: Inside Modal's web infrastructure as a story · 11 points
HN discussionWe use modal ( https://modal.com/ ). They give us GPUs on-demand, which is critical for us so we are only paying for what we are using. Pricing is about $2/hr per GPU (as a baseline of the costs). Long story short, things get VERY expensive quickly.
Comment linking /
on “Show HN: Lemon Slice Live – Have a video call with a transformer model”Submitted GPU Memory Snapshots: fast container cold boots as a story · 9 points
HN discussionYes. We use Modal ( https://modal.com/ ), and are big fans of them. They are very ergonomic for development, and allow us to request GPU instances on demand. Currently, we are running our real-time model on A100s.
Comment linking /
on “Show HN: Lemon Slice Live – Have a video call with a transformer model”Submitted Checkpoint/restore for sub-second container startup as a story · 9 points
HN discussion
