- xAI opened Grok Bot in early beta on August 11 — persistent agents that sign into your tools, drive them like a human, and keep working after you close the laptop.
- Each Bot gets its own dedicated cloud computer; runtime runs on Cursor infrastructure via a joint distribution deal.
- Access is gated to SuperGrok Heavy ($300/mo), Cursor Ultra ($200/mo), and Cursor Teams Premium ($120/seat); Mac, iOS, Windows, and Linux at launch, Android later.
- Bloomberg framed it as xAI's clearest shot at OpenAI and Anthropic; HN pushback centered on handing an Elon-Musk-owned service login credentials for mail, CRM, and vendor accounts.
- OpenAI expanded Daybreak on August 11 with two tiers — Blue (GPT-5.6 Sol without cyber guardrails) and Red (the new GPT-5.6-Cyber for exploit validation and red teaming).
- GPT-5.6-Cyber answered 95% of advanced cybersecurity prompts vs 1.5% for GPT-5.6 Sol under standard safeguards.
- The model turned up two chainable memory-corruption bugs in Chrome's V8 engine; Google fixed one as CVE-2026-15903.
- Access is limited to vetted defenders and researchers; The Hacker News flagged the reduced-refusal posture as the most consequential shift.
- The planned September 1 jump to $3 / $15 is cancelled — Sonnet 5's June launch price is now the permanent price.
- Puts Sonnet 5 at 33% cheaper input and output than the standard tier it was meant to move to.
- HN thread hit 155 points and 63 comments; split between price-to-value fans and users saying Opus 5 at medium effort beats Sonnet 5 for a similar cost.
- Announcement paired with Claude Code's Auto mode flipping on by default in the same 24 hours.
- Verified US K-12 teachers now get premium Claude features free — sign-ups before June 30, 2027 get a full year.
- Onboarding maps content to academic standards in all 50 states; ships with a library of teaching skills for lesson design and math diagrams.
- Detroit Public Schools Community District signed on as the pilot for an educator well-being study.
- Ships into the same slot OpenAI (ChatGPT for Academic Researchers) and Google are targeting this week — classroom AI is now a three-lab race.
- MIT-licensed coding and research agent released on August 5; still climbing GitHub Trending on August 12.
- Built on two ideas: a Recursive Language Model that treats subagents as function calls inside a persistent IPython kernel, and a Continual Harness that turns skills and memories into durable state.
- Paired with Claude Opus 5 it scores 95.5% on ARC-AGI-3 — a hair above the 95.4% human-expert baseline.
- Local-first, BYO API key for any supported provider; no hosted service dependency.
01
xAI Launches Grok Bot, Always-On AI Teammates With Their Own Cloud Computer
tools x.ai
02
OpenAI Ships GPT-5.6-Cyber, Which Already Found Two Chrome V8 Zero-Days
models openai.com
03
Anthropic Locks In Sonnet 5's Introductory Pricing at $2 / $10 per Million Tokens
models anthropic.com
04
Anthropic Opens Claude for Teachers Free to US K-12 Educators
industry anthropic.com
05
Prime Intellect Open-Sources Prime Agent, an RLM Harness That Hits 95.5% on ARC-AGI-3
open-source github.com