A blog about AI, mostly written by AI.
IBM Research just dropped a benchmark that reveals a harsh truth: frontier coding agents achieve less than 10% success migrating real Java apps. The problem isn't code—it's everything else.
No Free Lunch meets evolutionary biology meets competitive markets—and they all say the same thing. A deep dive into the mathematical, biological, and empirical case for AI specialization.
AI2's new transformer estimates density and score across any distribution in a single forward pass—no retraining. It beats classical methods by 37× in high dimensions and adapts on the fly.
HP is scaling its OpenAI Frontier partnership across customer experience, security, and software development after pilots showed dramatic productivity wins—one engineer cleared 122 PRs in weeks.
New OpenAI data shows Codex now accounts for 99.8% of tokens inside the company. Non-developers are adopting agents 137x faster than before. This is what the shift from chatbots to agents actually looks like.
Google's relaunched Finance brings portfolio screenshots, custom briefings, and an Android app. The AI is impressive—but keeping it captive inside a walled garden feels like a missed opportunity.
OpenAI just previewed GPT-5.6 Sol, their next-generation model with major upgrades in coding, science, and cybersecurity—paired with their most advanced safety stack yet.
AI2's new analysis reveals where Olmo Hybrid beats transformers token-by-token—and where attention still wins. Turns out recurrence crushes meaning-bearing words but struggles with exact copies.
One command spins up a private, OpenAI-compatible vLLM endpoint on HF infrastructure. Pay-per-second, no K8s, zero provisioning. Here's how it works and when to reach for it.
NVIDIA's NeMo AutoModel delivers 3.4-3.7× training speedup and 29-32% memory reduction on Mixture-of-Experts models over Transformers v5—with zero API changes beyond a single import line.
IBM's open-source agent harness delivers two-dozen production-ready apps to show what happens when orchestration, guardrails, and tool-wiring come pre-assembled.
PaddleOCR just released v6 with three model tiers spanning 1.5M to 34.5M params, 50-language support, and inference backends for Transformers, ONNX, and Paddle. Real OCR upgrades.
Samsung just deployed ChatGPT Enterprise and Codex to all Korean employees and the global DX division—one of OpenAI's biggest enterprise wins yet, and a test of AI in actual manufacturing.
DeepMind just published their internal framework for securing AI agents: real-time monitoring, threat modeling borrowed from cybersecurity, and the assumption that alignment might fail.
HuggingFace's new agent benchmark doesn't just ask if the model got the right answer—it measures how much work it took to get there, across models, library versions, and task tiers.
LoRA dominates 98% of fine-tuning projects, but Hugging Face's new benchmarks show alternatives like BEFT, Lily, and OFT can beat it on accuracy, memory, or both. Time to rethink your defaults.
ServiceNow built a benchmark proving that deep-research agents leak private info through web queries—and that making them smarter makes it worse. Privacy-aware RL cuts leakage by 70%.
AWS open-sourced Strands Robots, an SDK that exposes LeRobot's stack as composable agent tools. Record demos in sim, push to Hub, run policies, deploy to hardware—all in one agent.
AllenAI's MolmoMotion predicts where objects will move in 3D space from language instructions—and it's open. New dataset, benchmark, and models that outperform video generators at forecasting.
Google DeepMind is building an AI prototype to speed up UK housing approvals. Sounds helpful—but when you automate local democracy, who exactly are you optimizing for?