Hackernews posts about H100
H100 is a high-performance graphics processing unit (GPU) accelerator developed by Nvidia, designed for large-scale artificial intelligence and deep learning workloads.
- Cohere Open-Sources Command A+, a 218B Moe Model That Runs on Two H100s (firethering.com)
- Trump's 25% cut on Nvidia chips to China backfired as Beijing blocks H200 sales (finance.yahoo.com)
- Show HN: Manger – Livestock Management App (www.manger.app)
- The Token Value of $200/mo Plans (link.mail.beehiiv.com)
- Meta confirms 1000s of Instagram accounts were hacked by abusing its AI chatbot (this.weekinsecurity.com)
- MiMo-v2.5-Pro-UltraSpeed: 1T model with 1000 tokens per second (mimo.xiaomi.com)
- RFC 10008: The new HTTP Query Method (www.rfc-editor.org)
- Surprise, pay $1000 (forestwalk.ai)
- Pokemon Emerald Ported to WebAssembly (100k FPS) (pokeemerald.com)
- Snowboard Kids 2 is 100% Decompiled (blog.chrislewis.au)
- Deflock hits 100k ALPRs Mapped in USA (deflock.org)
- An Ohio Valley 100k-watt FM signal is severed in broad daylight (www.radioworld.com)
- Learnings from 100K lines of Rust with AI (2025) (zfhuang99.github.io)
- Federal judge blocks H1B visa $100K fee (www.alaskasnewssource.com)
- Subterranean fungi networks more than 100 quadrillion km in length (www.theguardian.com)
- The oldest surviving animated feature film at 100 (www.bbc.com)
- GPT Guesses Between 1 and 100 (github.com)
- CBS Radio signs off after nearly 100 years of broadcasting (www.cbsnews.com)
- Let's celebrate work that is 100% human-made (www.human-made.work)
- Show HN: ABC Classic 100 Rankings visualised (classic100.gotski.workers.dev)
- DeepSeek v4 Pro 1.6T model post-trained by Huawei on 1000 Ascend 910C chips (www.tomshardware.com)
- Why Julia's GPU Accelerated Ode Solvers Are 20x-100x Faster Than Jax and PyTorch (www.stochasticlifestyle.com)
- ShinyHunters hacked 100 orgs by exploiting an Oracle PeopleSoft 0-day (www.theregister.com)
- Trump's $100k H-1B visa fee is unlawful, US judge rules (www.reuters.com)