The latest blog posts to take off on Hacker News. 6,329 posts.
In this post, I will walk through how I learned to implement Flash Attention for 5090 in CUDA C++. The main objective is to learn writing attention in CUDA C++, since many features are not available in Triton, such as MXFP8 / NVFP4 MMA for sm120. I also feel this is a natural next step after learning about matmul…
moved to https://blog.efitz.net/posts/my-experience-creating-software-with-llm-coding-agents-part-2-tips/
I replaced my ChatGPT personalization settings with this prompt a few weeks ago and promptly forgot about it: Be extraordinarily skeptical of your own…
I get it, you’re too good to vibe code. You’re a senior developer who has been doing this for 20 years and knows the system like the back of your hand. Or maybe you’re the star individual contributor who is the only person who can ever figure out how to solve the hard problems. Or maybe you’re the professor who…
The failure mode that stalls “AI for data” efforts or "AI on my APIs" efforts isn’t psychedelic hallucination—it’s confident inaccuracy: plausible answers that are wrong in subtle and costly ways.
The comfort we get when offloading our cognitive load to LLMs is bad for us. Cognitive load should exist, and if we reduce it too much – if we stop thinking – we can actually unlearn how to think. Kids who always choose the easy route and copy their homework from other students eventually find […]
AI turns products from deterministic functions into probabilistic systems. That requires expanding old playbooks (SLOs, funnels, siloed finance), and reasoning in terms of trajectories, Minimum Viable Intelligence thresholds, and data as company operating system.
I’ve been coding entirely with Claude Code for the past two months. At first it was exhilarating. I was speeding through tasks. I was committing like mad.
We reveal AppleStorm, our investigation into how Apple AI’s eco-system quietly transmits messages (WhatsApp, iMessage) sent via Siri to Apple servers, even when it isn’t needed to complete the task.
A methodology for measuring the energy, emissions, and water impact of Gemini prompts shines a light on the environmental impact of AI inference.
In this blog post, we’ll detail how attackers can exploit image scaling on Gemini CLI, Vertex AI Studio, Gemini’s web and API interfaces, Google Assistant, Genspark, and other production AI systems. We’ll also explain how to mitigate and defend against these attacks, and we’ll introduce Anamorpher, our open-source…
"AI first" leaves humans nowhere. Learn about the Bus Factor and how the launch of ChatGPT has changed the landscape on knowledge sharing.
We love TPUs at Google, but GPUs are great too. This chapter takes a deep dive into the world of GPUs – how each chip works, how they're networked together, and what that means for LLMs, especially compared to TPUs. While there are a multitude of GPU architectures from NVIDIA, AMD, Intel, and others, here we will…
The lottery ticket hypothesis explains why massive neural networks succeed despite centuries of theory predicting they should fail
Its modern version (the reverse mode of automatic differentiation) was first published in 1970 by Seppo Linnainmaa
Things are about to get wild
Exploration into providing MCPs with the most powerful of all tools: code.
One in three top 10 Hacker News posts in August 2025 were AI-related. Can we pinpoint when this trend started?
<no value>
ChatGPT and its consequences have been a disaster for the human race.
This article explains how these four things fit together and shows some examples of what they are used for. Derivatives Derivatives are the most fundamental concept in calculus. If you have a funct…
Psychologically Plausible Deep Learning with Differentiable Tversky Similarity
Patterns that keep AI agent systems from falling apart in production: two-tier architecture, stateless design, orchestration, and lessons from UserJot.
Claude Opus 4 and 4.1 can now end a rare subset of conversations. An update on our exploratory research on model welfare.
Discover Imagen 4 Fast, Google's new speed-optimized text-to-image model, now generally available with Imagen 4 and 4 Ultra in the Gemini API.
One of Anthropic's quieter releases recently was their "Security Review," where Claude Code can identify and fix security issues in your code. But how good is it really? In my case, will it find issues with code it helped me write for my newsletter service…
Why automation lags even as models get better
Explore Gemma 3 270M, a compact, energy-efficient AI model for task-specific fine-tuning, offering strong instruction-following and production-ready quantization.
Reading research papers and articles about chain-of-thought reasoning makes me frustrated. There are many interesting questions to ask about chain-of-thought…
From the Zed Blog: Writing code is only one part of effective software engineering.
