2911
Paving the way to efficient architectures: StripedHyena-7B
StripedHyena-7B is an open-source model built beyond the Transformer architecture, matching Llama-2 and Mistral 7B while running faster on long sequences.
Topics /
AI & Machine Learning
The most popular blog posts about ai & machine learning on Hacker News. 6,325 posts.
2912
SnapFusion: Text-to-Image Diffusion Model on Mobile Devices Within Two Seconds
Text-to-image diffusion models can create stunning images from natural language descriptions that rival the work of professional artists and photographers. However, these models are large, with complex network architectures and tens of denoising iterations, making them computationally expensive and slow to run. As a…
2913
Hand Tracking for Mouse Input (2023)
The other day I saw the launch of Apple Vision Pro, the whole thing was very interesting, but one thing that caught my attention was the finger input. It seems very intuitive, by using the finger pinching as sort of like a cursor or mouse input. I figured I want to try it out, so I took it upon myself to create it.
2914
2915
How to build a brain – An introduction to neurophysiology for engineers
How does the brain work? While neuroscience provides insights on small-scale processes and overall behavior, we don’t know much for sure about the learning and processing algorithms in between. In this blog post, I will give an introduction to what we know. Let’s take an engineering perspective and play evolution: How…
2916
Problems the AI industry is not addressing adequately
As a supporter, I would love not to feel this way
2917
2918
2919
LLMs pose an interesting problem for DSL designers
The most exciting part of Programming Languages (PL) research for me has always been in Programming Language Design.
2920
Running Kimi K3 on MI355X at Better Performance per Dollar Than B300
Running Kimi K3 at ~952 tok/s/node, AMD continues to prove its case as the winner in performance per dollar.
2921
Choosing an AI model: one prompt, 11 models, different results
Netlify now runs any OpenRouter model, including Kimi K3, GLM 5.2 and DeepSeek V4. We tested 11 of them on the same build prompt to see how they differ.
2922
Managing context on the Claude Developer Platform
New context editing and memory tools enable Claude agents to handle long-running tasks without hitting context limits.
2923
Kevin Kelly's 2016 AI Predictions, 10 Years Later
Kelly predicted AI heavyweights while advocating for a future of open weights
2924
2925
The New AI Superpowers: Focus and Followthrough
Burnout is on the rise again, with an ironic twist.
2926
2927
OpenAI technical goals
OpenAI’s mission is to build safe AI, and ensure AI’s benefits are as widely and evenly distributed as possible.
2928
Replit's new AI Model now available on Hugging Face
At Replit, our mission is to empower the next billion software creators. Yesterday, we strengthened our commitment by announcing that Replit AI is now free for all users. Over the past year, we’ve witnessed the transformative power of building software collaboratively with the power of AI. We believe AI will be part…
2929
Building a deeper understanding of images
Posted by Christian Szegedy, Software EngineerThe ImageNet large-scale visual recognition challenge (ILSVRC) is the largest academic challenge in c...
2930
2931
Q-learning is not yet scalable
Over the past few years, we've seen that next-token prediction scales, denoising diffusion scales, contrastive learning scales, and so on, all the way to the point where we can train models with billions of parameters with a scalable objective that can eat up as much data as we can throw at it. Then, what about…
2932
How I Used Deep Learning to Train a Chatbot to Talk Like Me
Can you see the resemblance?
2933
Bringing K/V context quantisation to Ollama
K/V context cache quantisation has been added to Ollama. This enables significant reductions in VRAM usage, allowing users to realise the potential of expanded context sizes and run larger models at their existing context sizes.
2934
2935
Introducing Superalignment
We need scientific and technical breakthroughs to steer and control AI systems much smarter than us. To solve this problem within four years, we’re starting a new team, co-led by Ilya Sutskever and Jan Leike, and dedicating 20% of the compute we’ve secured to date to this effort. We’re looking for excellent ML…
2936
There's No Fire Alarm for Artificial General Intelligence
What is the function of a fire alarm?
2937
The Hater's Guide to the AI Bubble
Hey! Before we go any further — if you want to support my work, please sign up for the premium version of Where’s Your Ed At, it’s a $7-a-month (or $70-a-year) paid product where every week you get a premium newsletter, all while supporting my
2938
Slightly reducing the sloppiness of AI generated front end
I've been trying to make progress on a herculean task; I want decent looking programs for personal use which I can generate quickly using AI agents yet I'm a person without taste controlling an AI without taste. I found "one weird trick" that does enough for me, the results are not extra nice but it doesn't make me…
2939
Voicebox: Generative AI model for speech that generalizes across tasks
Voicebox is a state-of-the-art speech generative model based on a new method proposed by Meta AI called Flow Matching. By learning to solve a text-guided speech infilling task with a large scale of data, Voicebox outperforms single-purpose AI models across speech tasks through in-context learning.
2940
Real-time LLM Inference on Standard GPUs: 3k tokens/s per request
Today, Kog AI launches a tech preview of the Kog Inference Engine (KIE): 3,000 output tokens/s per request on 8× AMD MI300X GPUs and 2,100 on 8× NVIDIA H200 (FP16, no speculative decoding). This preview runs a 2B model, with support for large third-party MoE models coming next at similar speeds.