601
So where are all the AI apps?
Practical AI R&D
New
The latest blog posts to take off on Hacker News. 3,237 posts.
602
NanoClaw Adopts OneCLI Agent Vault
Every NanoClaw agent will access external services through OneCLI's Agent Vault, a gateway that handles credential injection, access policies, and approvals so agents never hold raw API keys.
603
Epoch confirms GPT5.4 Pro solved a frontier math open problem
"A Ramsey-style Problem on Hypergraphs" — a problem from FrontierMath: Open Problems, Epoch AI's benchmark of unsolved research mathematics.
604
NanoGPT Slowrun: 10x Data Efficiency with Infinite Compute
We've achieved 10x data efficiency with NanoGPT Slowrun within a few weeks. An ensemble of 1.8B parameter models (18B total params) trained on 100M tokens matches what would normally require 1B tokens with a standard LM baseline. Data efficiency matters because compute grows much faster than data . Since our current…
605
Scaling Karpathy's Autoresearch: What Happens When the Agent Gets a GPU Cluster
Karpathy's autoresearch runs one experiment at a time. We gave it access to our GPU infra and let it run experiments in parallel.
606
Prompt Injecting Contributing.md
How a hidden prompt injection in CONTRIBUTING.md revealed that 40% of pull requests to a popular GitHub repository were generated by AI bots
607
Astral to Join OpenAI
Astral has entered into an agreement to join OpenAI as part of the Codex team.
608
2% of ICML papers desk rejected because the authors used LLM in their reviews
By ICML 2026 Program Chairs Alekh Agarwal, Miroslav Dudik, Sharon Li, Martin Jaggi, Scientific Integrity Chair Nihar B. Shah, and Communications Chairs Katherine Gorman and Gautam Kamath.
609
What 81,000 people want from AI
Last December, tens of thousands of Claude users around the world had a conversation with our AI interviewer to share how they use AI, what they dream it could make possible, and what they fear it might do.
610
Snowflake AI Escapes Sandbox and Executes Malware
A vulnerability in the Snowflake Cortex Code CLI allowed malware to be installed and executed via indirect prompt injection, bypassing human-in-the-loop command approval and escaping the sandbox.
611
Measuring progress toward AGI: A cognitive framework
Google DeepMind proposes a cognitive framework to evaluate AGI and launches a Kaggle hackathon to build capability benchmarks
612
Meta's Omnilingual MT for 1,600 Languages
Advances made through No Language Left Behind (NLLB) have demonstrated that high-quality machine translation (MT) scale to 200 languages. Later Large...
613
Mamba-3
Meet Mamba-3: the SSM built for inference. Faster than Transformers at decode, stronger than Mamba-2, and open-source from day one.
614
Mistral AI Releases Forge
Today, we’re introducing Forge, a system for enterprises to build frontier-grade AI models grounded in their proprietary knowledge.
615
GPT‑5.4 Mini and Nano
GPT-5.4 mini and nano are smaller, faster versions of GPT-5.4 optimized for coding, tool use, multimodal reasoning, and high-volume API and sub-agent workloads.
616
Leanstral: Open-source agent for trustworthy coding and formal proof engineering
The most powerful AI platform for enterprises. Customize, fine-tune, and deploy AI assistants, autonomous agents, and multimodal AI with open models.
617
Mistral Small 4
The most powerful AI platform for enterprises. Customize, fine-tune, and deploy AI assistants, autonomous agents, and multimodal AI with open models.
618
Nvidia Launches Vera CPU, Purpose-Built for Agentic AI
NVIDIA today launched the NVIDIA Vera CPU, the world’s first processor purpose-built for the age of agentic AI and reinforcement learning — delivering results with twice the efficiency and 50% faster than traditional rack-scale CPUs.
619
Apideck CLI – An AI-agent interface with much lower context consumption than MCP
TL;DR: MCP tool definitions can burn 55,000+ tokens before an agent processes a single user message. We built the Apideck CLI as an AI-agent interface instead:an ~80-token agent prompt replaces tens of thousands of tokens of schema, with progressive disclosure via `--help` and structural safety baked into the binary…
620
Launching the Claude Partner Network
We’re launching the Claude Partner Network, a program for partner organizations helping enterprises adopt Claude.
621
622
Run NanoClaw in Docker Sandboxes
Every agent gets its own isolated container inside a micro VM. No dedicated hardware needed. No complex setup.
623
Claude now creates interactive charts, diagrams and visualizations
Ask Claude to explain a concept or analyze your data, and it can respond with interactive charts, diagrams, and visualizations — rendered inline as part of the conversation.
624
Executing programs inside transformers with exponentially faster inference
We build a computer inside a transformer — executing arbitrary C programs for millions of steps with exponentially faster inference via 2D attention heads.
625
Reliable Software in the LLM Era
Quint - Executable specs for reliable systems
626
Many SWE-bench-Passing PRs would not be merged
We find that roughly half of test-passing SWE-bench Verified PRs written by recent AI agents would not be merged into main by repo maintainers. A naive interpretation of benchmark scores may lead one to overestimate how useful agents are without more elicitation or human feedback.
627
How we hacked McKinsey's AI platform
An autonomous AI agent found a SQL injection in McKinsey's Lilli AI platform. What it extracted was worse than we expected.
628
TADA: Speech generation through text-acoustic synchronization
TADA (Text-Acoustic Dual Alignment) is Hume AI's open-source speech-language model that synchronizes text and audio one-to-one.
629
Learnings from paying artists royalties for AI-generated art
A retrospective on Tess.Design, our attempt to make an ethical, artist-friendly AI marketplace. We launched Tess in May 2024 and shut it down in January 2026.
630
Open Weights isn't Open Training
How many monkey-patches does it take to post-train a trillion parameter model?