It's Wednesday, September 23rd. Three AI labs shipped four models in two days. Grok 4.7 landed Monday. GPT-6 Sol and Luna followed Tuesday morning, with Claude Opus 5.5 arriving that same afternoon. Below, we walk through how to actually test all four against your own work instead of trusting a benchmark chart. Then: PayPal is routing checkout through Meta's shopping agent Muse, and one AIC builder explains how she keeps her AI agents synced through a shared "brain" in Notion.
In Today’s Top Tools, we spotlight some of the most innovative, creative AI apps that we recommend adding to your stack.
TOOL SPOTLIGHT
1️⃣ Four Models Shipped in Two Days. Test Them Before You Switch.
Image from OpenAI
Grok 4.7, GPT-6 Sol, GPT-6 Luna, and Claude Opus 5.5 all landed inside 48 hours. Run one real task through each before you change anything.
OpenAI released GPT-6 Sol and Luna on Tuesday, with API prices 50% lower than GPT-5.6 promotional pricing. Anthropic released Claude Opus 5.5 on Tuesday and says it matches Fable 5.1 on most tasks at 40% lower cost than Opus 5. Grok 4.7 arrived Monday, aimed at coding and knowledge work. Run one task in each, then compare quality, editing time, speed, and cost.
Try one realistic task in each model. Check each provider’s release page for current access.
Compare answer quality, revision effort, and response speed.
If you use a paid plan or API, compare the final cost too.
Run one task across all four, then check each provider’s release page for current access and model details.
Three launches in two days make this a good moment to retest your defaults. Run the same recurring task in GPT-6 Sol and Luna, Claude Opus 5.5, and Grok 4.7. Compare the output, editing time, response speed, and API cost if that applies to you. Benchmarks can help you shortlist, but your own work is the better test. Keep your current setup until a new model earns the switch.
TOOL SPOTLIGHT
2️⃣ PayPal and Shopify Both Wired Muse Into Checkout on the Same Day
Image from Meta
PayPal is bringing checkout to Meta Muse across its global merchant network. Muse launched with Stripe Link, and Shop Pay is expected to follow.
Yahoo Finance reports PayPal will let people shop and check out with Muse across PayPal’s global merchant network. Meta describes Muse as a personal AI agent for finding products and completing purchases. Before you approve a purchase, check the merchant, total price, delivery, and payment details.
Ask Muse to find a specific product and compare options against a budget or list of requirements.
Review the merchant, product, price, and delivery details before approving a purchase.
Try a low-stakes shopping research task first, then decide whether you want the agent to reach checkout.
Muse Agent can search, compare, and help complete a purchase. You stay in control of the final approval.
PayPal gives Muse another route to checkout, but reach is only part of the story. People need to know which merchant the agent is using, what the full price includes, and when it is about to place an order. For merchants, this partnership could add a route to buyers who shop through AI. For shoppers, the test is whether Muse saves time and keeps product, delivery, and payment choices clear before they approve.
🔥 What 120 Chapters Are Building Through October 31
You just read about two tools. Here is somewhere to point them. Hack for Humanity is two weeks in, with 120 AIC chapters in 50 countries building through October 31. Each chapter picked its own problems, drawn from what 250,000 people named during Humans in AI Week in June: job displacement, technical literacy, accessibility. No sponsor brief, no product to validate, no predetermined answer. MLH, Google, Lovable, Snowflake, and Databricks are backing the builds without steering them. Find one happening near you or start one in your city.
🎓AIC AUTOMATIONS
Keep Your Agent Brain in the Cloud
A shared notebook helps connected agents pick up where the last one stopped.
I first kept my Agent Brain in a folder of notes on my computer. That worked when I was using local tools, but cloud agents couldn’t reach those notes while my computer was closed. So I moved the shared notes to Notion, where connected agents can still use them.
Think of an Agent Brain as a shared notebook. Agents can look up useful context and leave updates for one another.
What goes in the Agent Brain?
Skills: instructions for tasks I do often.
Agents: the agents I use and where they run.
Memory: useful context, decisions, and lessons.
Daily Activity: what the agents have done.
Browsing: research they have collected.
Vault Files: links to notes and documents.
Step 1: Set up the shared notebook
Start with one Notion page. Ask your AI assistant to create a simple structure and explain what belongs in each section:
Help me set up a simple Agent Brain in Notion for the AI agents I use. Create sections for Skills, Agents, Memory, Daily Activity, Browsing, and Vault Files. Explain what belongs in each one. Use my actual notes, and mark anything you can’t confirm as “not set up yet.” If you can’t create a Notion page from this chat, give me an outline I can paste in.
Step 2: Connect an agent
Once the page exists, tell each agent how to use it:
Before starting work, check the relevant Agent Brain sections. After each action, add a short Daily Activity note with what you did and the result. Save lasting decisions and lessons in Memory. Link to source files when possible. Don’t make up details or save passwords and API keys. Tell me if you can’t access the page.
What changes?
The setup is agent-agnostic. If you use multiple AI models or subscriptions, connected agents can read the same context and pick up where another agent left off.
Pro tip: keep it updated automatically
Instinct writes its logs every 30 minutes. For local tools like Claude Code and Codex, sync their notes to Notion so cloud agents can see them too.
One way to put AI to work, every week. From Joy Dong
In this section, we feature a few standout opportunities at leading AI companies, non-profits, policy groups, and other organizations.
🔬 Work at a Lab/AI Safety
Roles at frontier labs, safety organizations, and research institutions
RESEARCH ENGINEER / SCIENTIST, ALIGNMENT - Anthropic | San Francisco, CA | $350,000-$500,000
RESEARCHER, ALIGNMENT INTERPRETABILITY - OpenAI | San Francisco, CA | $295,000-$500,000 + equity
RESEARCH SCIENTIST, CONTROL - Apollo Research | San Francisco, CA or London, UK | $204,000-$385,000 SF, £136,000-£258,000 London, + equity
TASK DEVELOPMENT ENGINEER - METR | Berkeley, CA | $260,937-$385,490
MEMBER OF TECHNICAL STAFF, NORTH MODELLING (EVALS) - Cohere | Remote, London, Europe or Toronto | No UK range published. Canada CA$250,000-CA$535,000, US CA, NY and WA $205,000-$380,000, + equity
🚀 Work in Industry
Roles at funded AI startups and private companies
SOFTWARE ENGINEER, AI PRODUCT - Figma | San Francisco, CA or New York, NY | $153,000-$376,000
AI ENGINEER, PRODUCT - Brex | San Francisco, CA | $171,000-$240,000
STAFF PRODUCT MANAGER, AGENTIC AI APPLICATIONS - Databricks | Mountain View or San Francisco, CA | $172,200-$236,850 zone 1, $172,200-$258,400 local
MACHINE LEARNING ENGINEER - Stripe | South San Francisco, CA | $212,000-$318,000
SOFTWARE ENGINEER, AI SDK - Vercel | San Francisco, CA or New York, NY, hybrid | $196,000-$294,000 San Francisco base range
🏛️ Work in AI Policy / Governance
Roles in government, think tanks, NGOs, and policy organizations
AI POLICY RESEARCH RESIDENT - RAND Corporation | Washington, DC | $110,900-$183,300
RESEARCH MANAGER, UK FELLOWSHIPS - GovAI | London, UK | £83,000-£105,000
POLICY ADVISOR / SENIOR POLICY ADVISOR - Encode | Washington, DC | $110,000-$180,000
POLICY ADVISOR - MIRI | Washington, DC | $120,000-$300,000
✨AIC OPPORTUNITIES
Current ways to build with The AI Collective.
DIRECTOR OF SF BAY AREA PROGRAMS - The AI Collective | SF Bay Area, in-person for events | Full-time, paid | $60,000-$80,000
EDUCATION OPERATIONS LEAD - The AI Collective | Remote | Volunteer | 5-10 hours a week
EXECUTIVE ASSISTANT - The AI Collective | Remote, LatAm preferred, PST hours | $1,500 a month
PARTNERSHIPS LEAD, GRANTS - The AI Collective | Remote | Volunteer, part-time | ~5-7 hours a week
🫵 Want your message in front of 200,000 AI builders?
Our partners and sponsors get exclusive placements across the newsletter and access to AIC's in-person network — demo nights, dinners, hackathons, and forums across 180+ chapters.
For all inquiries, send us a note at [email protected].
The AI Collective is built by volunteers across 180+ chapters in 40 countries.
Thank you to the thousands of volunteers around the world who make this work possible. We truly could not do this without you.
🧑💻 About the Editors
About Noah Frank
Noah is a researcher, innovation strategist, and ex-founder thinking and writing about the future of AI and the workforce. His work and body of research explores the economics of emerging technology and organizational strategy. Outside of AIC, Noah heads research for Centaurian AI.
About Lindsay Gross
Lindsay is an AI engineer, researcher, and writer focused on how AI systems behave in practice and what it takes to make them safe. Her work sits at the intersection of AI safety, governance, and product design, and at AIC she writes about the questions that matter most as these systems scale.