571
Hacking Diffusion into Qwen3 for the Arc Challenge
A deep dive into ongoing efforts adapting language models for diffusion-based ARC solving.
New
The latest blog posts to take off on Hacker News. 1,227 posts.
572
Things that helped me get out of the AI 10x engineer imposter syndrome
Curing Your AI 10x Engineer Imposter Syndrome
573
Genie 3: A new frontier for world models
Today we are announcing Genie 3, a general purpose world model that can generate an unprecedented diversity of interactive environments. Given a text prompt, Genie 3 can generate dynamic worlds that you can navigate in real time at 24 frames per second, retaining consistency for a few minutes at a resolution of 720p.
574
Lack of intent is what makes reading LLM-generated text exhausting
There’s something icky about LLM-generated text when you think it’s written by a human. I think I finally put my finger on one reason why I feel this way.
575
Qwen-Image: Crafting with native text rendering
GITHUB HUGGING FACE MODELSCOPE DEMO DISCORD We are thrilled to release Qwen-Image, a 20B MMDiT image foundation model that achieves significant advances in complex text rendering and precise image editing. To try the latest model, feel free to visit Qwen Chat and choose “Image Generation”. The key features include…
576
AI promised efficiency. Instead, it's making us work harder
AI tools were supposed to free up our time—but they’re increasing our cognitive load and making us less productive. Here’s what’s really happening (and how to use them without burning out).
577
Perplexity is using stealth, undeclared crawlers to evade no-crawl directives
Perplexity is repeatedly modifying their user agent and changing IPs and ASNs to hide their crawling activity, in direct conflict with explicit no-crawl preferences expressed by websites.
578
Typed languages are better suited for vibecoding
My >10 year old programming habits have changed since Claude Code launched. Python is less likely to be my go-to language for new projects anymore. I am managing projects in languages I am not fluent in—TypeScript, Rust…
579
One Dataset. No Warning. Google Took Everything. You're Not Safe Either
I’m an independent developer who built Punge, an on-device NSFW image detection app. The idea was simple: give people tools to manage…
580
Persona vectors: Monitoring and controlling character traits in language models
A paper from Anthropic describing persona vectors and their applications to monitoring and controlling model behavior.
581
Tokens are getting more expensive
"language models will get cheaper by 10x" will not save ai subscriptions from the short squeeze
582
Cerebras Code
We are launching two new plans designed to make AI coding faster and more accessible: Cerebras Code Pro ($50/month) and Code Max ($200/month). Both plans give you access to Qwen3-Coder , the world’s leading open-weight coding model—running at speeds of up to 2,000 tokens per second , with a 131k-token context window …
583
Does the Bitter Lesson Have Limits?
The bitter lesson says general methods and compute win. But compute is a constraint, and human-crafted logic often ships cheaper and sooner.
584
Deep Agents
Build deep agents that plan and execute complex tasks over longer time horizons. Learn the architecture behind Deep Research and Claude Code.
585
Facts will not save you – AI, history and Soviet sci-fi
AI, history, and Soviet sci-fi
586
Gemini 2.5 Deep Think
Deep Think utilizes extended, parallel thinking and novel reinforcement learning techniques for significantly improved problem-solving.
587
AI is a floor raiser, not a ceiling raiser
A reshaped learning curve
588
Gemini Embedding: Powering RAG and context engineering
Explore how organizations leverage Gemini Embedding for context-aware AI systems improving performance, accuracy, and efficiency across diverse industries.
589
Releasing weights for FLUX.1 Krea
Krea announces the open release of FLUX.1 Krea, including model weights, training details, and resources to run and fine-tune the model.
590
OpenAI's "Study Mode" and the risks of flattery
Serious learning requires friction, frustration... and other humans
591
Stargate Norway
We’re launching Stargate Norway—OpenAI’s first AI data center initiative in Europe under our OpenAI for Countries program. Stargate is OpenAI’s overarching infrastructure platform and is a critical part of our long-term vision to deliver the benefits of AI to everyone.
592
I know when you're vibe coding
I shouldn't have to care about this. I don't want to care about how someone's code gets into the IDE. Whether you wrote it by hand, copied it from a forum, prom
593
Ollama's new app
Ollama's new app is now available for macOS and Windows.
594
Vibe code is legacy code
Code that nobody understands is tech debt
595
Critical vulnerability in AI coding platform Base44 allowing unauthorized access
New discovery underscores security implications of AI-powered development and the rise of Vibe Coding Platforms.
596
Study mode
Introducing study mode in ChatGPT, a new learning experience that helps you work through problems step by step, guiding students with questions, scaffolding, and feedback for deeper learning.
597
My 2.5 year old laptop can write Space Invaders in JavaScript now (GLM-4.5 Air)
I wrote about the new GLM-4.5 model family yesterday—new open weight (MIT licensed) models from Z.ai in China which their benchmarks claim score highly in coding even against models such …
598
Stop selling “unlimited”, when you mean “until we change our minds”
Claude Code joins the AI pricing bait-and-switch
599
Robot hand could harvest blackberries better than humans
The soft robotic gripper developed at the U of A received a U.S. patent.
600