Latest Artificial Intelligence News & Technical Dispatches

Fable 5.1 (158 HN pts, PhiloLabs), General Intuition ($6B valuation), and NVIDIA NOOA — three radically different approaches to world modeling for autonomous agents. Head-to-head on simulation fidelity, compute cost, causal reasoning, and enterprise production readiness.

Google's Gemini 3.8 Flash launched to 863 HN points on September 2 — its highest-rated model launch of 2026. The Flash Cyber variant embeds 27M security advisories into a dedicated Mixture of Security Experts layer, achieving 89.4% zero-day detection recall.

Meta's Muse Spark 1.3 launched to 429 HN points — the company's highest-rated AI launch of 2026. Generates 512x512 images in 0.8 seconds on consumer GPUs with 4.2 FID, released under CC BY-NC 4.0.

WebLLM runs 27B models in-browser at 45 tok/s — 86% of Ollama's native speed. Head-to-head comparison on latency, VRAM, privacy, deployment complexity, and which wins for your production agent pipeline.

Gemini 3.8 Flash scored 863 HN points — Google's highest-rated AI launch of 2026. Deep dive into the Flash Cyber security-first variant, its 89.4% zero-day recall on SECURE-bench, and what the 3.8 architecture means for enterprise AI.

OpenAI launches GPT-5.6 Sol API with sub-100ms time-to-first-token latency and 180 tok/s throughput at $2.50 per million input tokens. The fastest inference launch in OpenAI's history positions Sol as the premium choice for real-time agent applications requiring instant responses.

Evaluating AI agents is fundamentally different from evaluating LLMs. Agents make tool calls, follow multi-step plans, use external data, and produce outputs that are hard to score with static benchmarks. This guide covers production-grade eval harnesses for task completion, tool accuracy, latency, cost, and regression detection.

The Model Context Protocol registry surpassed 10,000 registered MCP servers in September 2026, marking a pivotal milestone for the AI agent ecosystem. From 2,000 servers in January to 10,000 in September, MCP has become the standard protocol for AI agent tool integration.

RAG, fine-tuning, and agentic retrieval are the three main approaches to injecting knowledge into LLMs. This comprehensive comparison covers accuracy, latency, cost, and maintenance trade-offs with a decision framework for enterprise use cases in 2026.

Meta releases Llama 4.5 with 405 billion parameters, 410 tok/s throughput, 128K context window, and Apache 2.0 license. At $0.15 per million tokens on API providers, Llama 4.5 reshapes the economics of open-weight AI for enterprise deployments.

Speculative decoding has matured from research paper to production standard in 2026. Medusa heads, Eagle draft models, and self-speculation techniques cut inference latency by 2.5x without quality loss. Deep dive into architecture, benchmarks, and deployment trade-offs.

OpenCode's launch hit 1,274 HN points — the highest for any AI coding tool in 2026. 14,200 GitHub stars in 48 hours, 7K-token system prompt (79% less than Claude Code), and 47 community MCP integrations.

LiteLLM (1.8ms overhead, 200+ providers) vs Unify (3.2ms, dynamic latency benchmarking). Which proxy should power your production AI stack? Complete benchmark data, config files, and decision matrix.

Docker Sandboxes GA'd September 1 with 180ms cold starts, gRPC connection pool API, and native OpenCode integration. $0.002/min with volume discounts at 10K+ minutes/month.

Enterprise LLM costs dropped from $200 to $30 per million tokens through 5 optimization layers. This production playbook covers prompt compression, semantic caching, speculative decoding, multi-model routing, and batch processing with runnable TypeScript code.