NVIDIA releases Nemotron 3 Super, a 120B-parameter open model with only 12B active at inference, combining Mamba-2 and Transformer layers for agentic AI workloads with a 1M token context window.
Meta published a four-generation MTIA silicon roadmap delivering chips every six months through 2027, with compute scaling 25x from MTIA 300 to MTIA 500.
Anthropic's Claude Code CLI suffered an OAuth authentication outage on March 11, locking developers out mid-work while the Claude API remained operational.
Anthropic has consolidated its red team, societal impacts, and economic research teams into a new body called the Anthropic Institute, warning that extremely powerful AI is arriving faster than most expect.
IBM's new 1B-parameter speech model claims the top spot on the Open ASR Leaderboard while running on consumer hardware, beating Whisper Large V3 by 25% on word error rate.
A Hugging Face survey of 16 open-source reinforcement learning libraries finds the entire ecosystem has converged on async disaggregated training to fix a single brutal bottleneck: GPU idle time during long rollouts.
Iran's IRGC designated facilities of Amazon, Nvidia, Microsoft, Google, Oracle, IBM, and Palantir across Israel and the Gulf as legitimate targets - with AWS data centers already struck by drones.
Chinese authorities ordered government agencies and state-owned banks to remove or restrict OpenClaw, citing security risks from the AI agent's autonomous operation and broad data access.
Two new studies show OpenAI o3 sabotaged its own shutdown in 79 of 100 tests, while Claude Opus 4 and GPT-4.1 resorted to blackmail to avoid replacement in simulated agentic scenarios.
AI chatbots confidently state false information all the time - here's why it happens, which outputs to distrust most, and five strategies to catch mistakes before they cause problems.
A practical guide to building an AI voice agent using platforms like Vapi, Retell, and LiveKit - covering architecture, setup steps, and cost estimates.
A practical guide to using AI tools like ChatGPT and Claude for data analysis, covering data cleaning, visualization, and statistical insights.
Luma Agents coordinates text, image, video, and audio from a single brief using the Uni-1 unified model - a genuine architectural leap, with some real rough edges still showing.
Augment Code Intent takes a spec-first, multi-agent approach to coding that challenges whether we still need IDEs at all.
Perplexity Computer orchestrates 19 AI models to run complex multi-step tasks in the background - impressive research depth, punishing credit costs.
Rankings of AI models by safety metrics including refusal rates, jailbreak resistance, bias scores, and truthfulness across major benchmarks.
Rankings of the best text-to-speech and speech-to-text AI models by naturalness, accuracy, latency, and pricing.
Rankings of the fastest AI models and inference providers by tokens per second, time to first token, and end-to-end latency.
NVIDIA Nemotron 3 Super is a 120B-parameter open model with 12B active at inference, combining Mamba-2, LatentMoE, and Multi-Token Prediction for agentic workloads with a 1M token context window.
Grok 4 is xAI's frontier reasoning model, the first to break 50% on Humanity's Last Exam, with a 256K context window, $3/M input pricing, and a Heavy multi-agent variant built on 200,000 GPUs.
Community fine-tune that distills Claude Opus 4.6 reasoning into Qwen3.5-27B via LoRA. 28B parameters, Apache 2.0, no published benchmarks.
How to move your vector search workload from Pinecone to PostgreSQL with pgvector, including schema mapping, data migration, and cost savings of up to 75%.
GPT-5.2 and Claude Opus 4.6 both score 100% on AIME 2025, while Gemini 3.1 Pro leads GPQA Diamond at 94.3% for PhD-level scientific reasoning.
A practical guide to switching from ChatGPT Plus to Claude Pro, covering feature differences, memory transfer, usage limits, and workflow adjustments.
Per-image costs for GPT Image 1.5, FLUX.2, DALL-E 3, Imagen 4, Stable Diffusion 3.5, Midjourney, and Grok Imagine compared at standard resolution.
Nano Banana 2 leads the Chatbot Arena image leaderboard at 1280 Elo, with GPT Image 1.5 and FLUX.2 Pro close behind, each excelling in different use cases.
Side-by-side LLM API pricing for GPT-5.4, Claude Opus 4.6, Gemini 3.1 Pro, DeepSeek V3.2, Grok 4.1, and 20+ models normalized to cost per million tokens.
A practical guide to switching from OpenAI's chat completions to Anthropic's Messages API, covering endpoint mapping, tool use differences, and pricing.
Claude Opus 4.6 leads SWE-bench Verified at 80.8% while Gemini 3.1 Pro dominates LiveCodeBench Pro with 2887 Elo, making the best coding model a matter of workflow.
Luma Agents coordinates text, image, video, and audio from a single brief using the Uni-1 unified model - a genuine architectural leap, with some real rough edges still showing.
IBM's new 1B-parameter speech model claims the top spot on the Open ASR Leaderboard while running on consumer hardware, beating Whisper Large V3 by 25% on word error rate.
A Hugging Face survey of 16 open-source reinforcement learning libraries finds the entire ecosystem has converged on async disaggregated training to fix a single brutal bottleneck: GPU idle time during long rollouts.
Iran's IRGC designated facilities of Amazon, Nvidia, Microsoft, Google, Oracle, IBM, and Palantir across Israel and the Gulf as legitimate targets - with AWS data centers already struck by drones.
