8 min read
AA-Robotics v0.1: GPT-6 Astra 69, Humanoids Under 2
Artificial Analysis AA-Robotics v0.1: GPT-6 Astra 69, Sol 65, Gemini 4 Argon 55 on a robot arm; no model reaches 2/100 on humanoids. What is verified.
Blog · updated daily
The AI blog.
New AI, explained the day it ships. Frontier model releases, agent guides, Claude Code deep dives, MCP servers, AI careers, and global policy — 3,100+ articles from the instructors who've taught 350,000+ students.
Browse by topic →
8 min read
Accenture Claude Cap: $29 a Month, Power Users Hit
Accenture now funds $29 per person monthly for Claude Enterprise; users spending $500 to $1,000 need project budgets. Memo details and what is unconfirmed.
8 min read
AI Interviewer vs AI Interviewee: Two Bots, One Loop
A comedian sent his AI agent to face an AI recruiter. The bots haggled over salary, booked a follow-up alone and got stuck saying goodbye. What it shows.
9 min read
AI21 Nemotron 3.5 Beats Opus 5.5 at 33x Less: What Is Real
AI21 says post-trained Nemotron 3.5 Super VL models match Claude Opus 5.5 for Gong, Deel and Alice at lower cost. Aggregators quote 88.5%; AI21 says 72%.
8 min read
Amazon Bee AI Wearables: $75 Frontier, $100 Eclipse
Amazon Bee Frontier ($75) and Eclipse ($100) record your day and summarize it; Bee Pro costs $12/month. What is confirmed, what the privacy claims mean.
8 min read
Anthropic + METR: Employee-Like Access, 3 Evaluator Tiers
Anthropic onboarded six METR staff with employee-like access and defined three evaluator tiers. What can be published, redacted and what is still open.
11 min read
Anthropic Q3 Revenue $16.8B: Claimed vs Verified
Anthropic reportedly told IPO investors Q3 revenue was $16.8B, up 45%. Here is what is sourced, what is unconfirmed, and what it means for builders.
8 min read
Scion vs Mononoke: Attribution Dispute, Facts Checked
Scion vs Mononoke: a Meta alum alleges uncredited copying; an Anthropic author says Sapling was credited and a footnote is coming. What is verified.
8 min read
claude-plugins-official: Install, Submit, Trust Guide
anthropics/claude-plugins-official guide: /plugins vs /external_plugins, install command, submission form, immutable names, security warning.
9 min read
Apertus 1.5 Report: Open 8B/70B With 262K Context
Apertus 1.5 adds image, audio, 262K context and thinking mode to the fully open Swiss 8B and 70B models. Report summary.
10 min read
Apple + Meta Muse Spark Open: Mac Studio Deal, Cost, RAM
Apple and Meta partner so Muse Spark Open (816B) runs on Mac Studio M5 Ultra. Costs from $10,799, RAM claims compared, and what is still unconfirmed.
9 min read
MLX RDMA on Mac Studio: 36.7 tok/s on 4 M5 Ultras
Apple research: MLX plus RDMA over Thunderbolt 5 runs an 816B model on four Mac Studio M5 Ultras at 36.7 tokens/s. Benchmarks, setup and limits.
9 min read
Plan or Retrieve First? Apple Tests Tool Discovery
Apple's tool discovery study on ToolBench: retrieve first wins for single-tool queries, plan first wins for multi-tool ones, at double the inference cost.
8 min read
Apple SSR: 90% Faster Reasoning for Small Search Agents
Apple's selection-based structured reasoning cuts per-turn latency over 90% in small multimodal search agents. How SSR works, and what the page leaves out.
8 min read
Ternus: Apple Memory "Crisis," Chinese Chips on the Table
Apple CEO John Ternus told TIME the memory shortage is a crisis and all supply options should be on the table, plus what he said about Siri, Gemini and AI.
8 min read
AWS Trust in Depth: 4 Layers for Safer AI Agents (2026)
AWS Trust in Depth gives AI agents four layers: scope, know, bound, observe, built on AgentCore Policy and automated reasoning. What it does, and what to check.
9 min read
AWS AI-DLC Workflows: 7 Coding Agents, 33 Stages (2026)
awslabs/aidlc-workflows adds approval gates and audit trails to Claude Code, Codex, Cursor, Kiro and more. Install steps, limits and verdict.
8 min read
Bluesky Attie Social Vibe Coding: Sites, Games, Quizzes
Bluesky's Attie adds social vibe coding for sites, games and polls on AT Protocol. What was announced at Disrupt, the models allowed, and open questions.
…
Page 1 of 174 · 3131 articles