Research library
Insights
Analysis of AI, automation, and software: what is changing, what holds up, and what it means for practical decisions.
173 articles · Page 7 of 15
Anthropic Redeploys Claude Fable 5 After 18 Days Under U.S. Export Controls
Anthropic restored Claude Fable 5 on July 1 after Commerce lifted an 18-day export-control hold triggered by an Amazon jailbreak. What the outage reveals.
Claude Science: Anthropic Builds an AI Workbench for the Lab Bench
Claude Science is Anthropic's beta research workbench: 60+ domain skills, HPC orchestration, and BioNeMo integration. A frontier lab goes vertical.
Claude Sonnet 5: Anthropic Cuts the Price and Closes the Gap to Opus
Claude Sonnet 5 ships at $2/$10 launch pricing as the most agentic Sonnet yet. The real story: the gap to Opus 4.8 and a new tokenizer.
GPT-5.6 Sol: OpenAI Gates Its New Flagship at the Government's Request
OpenAI previewed GPT-5.6 Sol, Terra, and Luna—then gated them to ~20 government-approved partners at the White House's request. What it signals.
OpenAI's GPT-5.5-Cyber and Codex Security: Defensive AI Moves Into the Commit Stream
OpenAI's GPT-5.5-Cyber and the Codex Security plugin move defensive AI from the chat window into the commit stream. What the AISI benchmarks really show.
Why the AI Frontier Is Splitting Into Gated Specialists and Open Generalists
In 2026 the AI frontier forked: gated, domain-specific models for high-stakes work versus cheap open-weight generalists. What each means for buyers.
What It Costs to Run AI in Production: A 2026 Pricing Breakdown
What LLM inference actually costs in 2026: per-token rates, reasoning-token overhead, prompt caching, and the self-host crossover point.
GLM 5.2: An Open-Weight Coder That Beats GPT-5.5 on Price, Not the Frontier
Z.ai's MIT-licensed GLM 5.2 edges GPT-5.5 on agentic coding at one-sixth the API cost, but trails Opus 4.8 and loses on terminal coding.
Workflow Automation ROI: A Calculator and a Worked Example
How to compute automation ROI: loaded labor cost, error-reduction value, build cost, and payback period — with a concrete invoice-processing example.
Kimi K2.7 Code: Moonshot Cuts Thinking Tokens 30%, But the Benchmarks Are All Its Own
Moonshot's Kimi K2.7 Code cuts reasoning tokens ~30% and undercuts the frontier on price, but every headline benchmark is first-party and unverified.
AI Governance for Small Businesses: A Practical Checklist
A concrete, non-bureaucratic AI governance checklist for SMBs: data handling, vendor due diligence, shadow AI, human-in-the-loop, and audit logging.
Apple WWDC 2026: Siri Runs on Gemini Now, and Apple Is Fine With That
Apple's rebuilt Siri runs on a custom 1.2T-parameter Google Gemini model at ~$1B/year — in Tim Cook's last keynote, Apple conceded the model race.