721
Trinity large: An open 400B sparse MoE model
A deep dive into the open weight Trinity Large model, covering architecture, sparsity, training at scale, and why we shipped Preview, Base, and TrueBase checkpoints.
New
The latest blog posts to take off on Hacker News. 3,237 posts.
722
Prism
Prism is a free LaTeX-native workspace with GPT-5.2 built in, helping researchers write, collaborate, and reason in one place.
723
AI2: Open Coding Agents
SERA is the first in our family of Open Coding Agents, achieving state-of-the-art performance at low cost.
724
Kimi Released Kimi K2.5, Open-Source Visual SOTA-Agentic Model
Kimi K2.5 defines Visual Agentic Intelligence. Trained on 15T tokens, it introduces SOTA visual coding and autonomous agent swarm. Read the full tech blog.
725
There is an AI code review bubble
Today everybody's doing AI code review. Here's how Greptile's viewpoint is differentiated - independence, autonomy, and feedback loops.
726
Qwen3-Max-Thinking
Qwen offers comprehensive functionality spanning chatbot, image and video understanding, image generation, document processing, web search integration, tool utilization, and artifacts.
727
Unrolling the Codex agent loop
A technical deep dive into the Codex agent loop, explaining how Codex CLI orchestrates models, tools, prompts, and performance using the Responses API.
728
GPTZero finds 100 new hallucinations in NeurIPS 2025 accepted papers
GPTZero's analysis 4841 papers accepted by NeurIPS 2025 show there are at least 100 with confirmed hallucinations
729
Qwen3-TTS family is now open sourced: Voice design, clone, and generation
Qwen offers comprehensive functionality spanning chatbot, image and video understanding, image generation, document processing, web search integration, tool utilization, and artifacts.
730
Your brain on ChatGPT: Accumulation of cognitive debt when using an AI assistant
This study explores the neural and behavioral consequences of LLM-assisted essay writing. Participants were divided into three groups: LLM, Search Engine, and …
731
ANN v3: 200ms p99 query latency over 100B vectors
Our latest ANN release supports scales of 100+ billion vectors in a single search index, with 200ms p99 query latency at 1k QPS and 92% recall.
732
Claude's new constitution
A new approach to a foundational document that expresses and shapes who Claude is
733
How AI destroys institutions
Civic institutions—the rule of law, universities, and a free press—are the backbone of democratic life. They are the mechanisms through which complex societies encourage cooperation and stability, while also adapting to changing circumstances. The real superpower of institutions is their ability to evolve and adapt…
734
Our approach to age prediction
ChatGPT is rolling out age prediction to estimate if accounts are under or over 18, applying safeguards for teens and refining accuracy over time.
735
The assistant axis: situating and stabilizing the character of LLMs
Who is the Assistant? We investigate the character that most modern language models inhabit when interacting with users.
736
Keeping 20k GPUs healthy
How we do active and passive monitoring on hyperscalers and neoclouds.
737
How scientists are using Claude to accelerate research and discovery
Labs at Stanford and MIT built Claude-powered systems that run genome-wide studies in minutes and automate gene cluster interpretation.
738
FLUX.2 [Klein]: Towards Interactive Visual Intelligence
Introducing FLUX.2 [klein], our fastest image models. Unified generation and editing in under a second, on hardware you already own. Apache 2.0 on 4B.
739
Our approach to advertising
OpenAI plans to test advertising in the U.S. for ChatGPT’s free and Go tiers to expand affordable access to AI worldwide, while protecting privacy, trust, and answer quality.
740
LLM Structured Outputs Handbook
A practical handbook for developers using LLMs to get structured outputs.
741
Pocket TTS: A high quality TTS that gives your CPU a voice
Our mission is to build and democratize artificial general intelligence through open science.
742
Furiosa: 3.5x efficiency over H100s
Meet the RNGD Server, delivering scalable, energy-efficient AI inference at data center scale with FuriosaAI’s Renegade accelerator platform.
743
Scaling long-running autonomous coding
We've been experimenting with running coding agents autonomously for weeks at a time.
744
Claude Cowork exfiltrates files
Claude Cowork is vulnerable to file exfiltration attacks via indirect prompt injection as a result of known-but-unresolved isolation flaws in Claude's code execution environment.
745
Why we don’t use AI
We get asked about AI a lot. Whether we’re going to add it to Yarn Spinner, whether we use it ourselves, what we think about it. Fair questions. Time to write it all down. Yarn Spinner doesn’t use the technology that’s currently being called AI. We don’t have generative AI features in the product, and we don’t use…
746
We can't have nice things because of AI scrapers
In the past few months the MetaBrainz team has been fighting a battle against unscrupulous AI companies ignoring common courtesies (such as robots.txt) and scraping the Internet in order to build up their AI models . Rather than downloading our dataset in one complete download, they insist on loading all of…
747
Using proxies to hide secrets from Claude Code
How to use HTTP proxies to prevent credential exposure in Claude Code sandboxes.
748
vLLM large scale serving: DeepSeek 2.2k tok/s/h200 with wide-ep
How vLLM reaches 2.2k tokens per second per H200 for DeepSeek-style MoE serving with Wide-EP, async scheduling, dual-batch overlap, disaggregated serving, CUDA
749
Mozilla's open source AI strategy
The future of intelligence is being set right now, and the path we’re on leads somewhere I don’t want to go. We’re drifting toward a worl
750
