It's Wednesday, October 7th. OpenAI's Decisions API gives apps a faster way to choose among answers you define. Claude now works inside the Google Doc, Sheet, or Slide you have open.
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 Let An App Make The Small Calls Faster
Image from OpenAI Developers
OpenAI's new Decisions API turns text or images into a fixed answer your app can act on, such as a category, a yes/no probability, or a score. It is in public beta for developers.
An app often needs to choose a path before it needs to write a reply. OpenAI's Decisions API gives developers a dedicated endpoint for that first step. The developer supplies the question and, for categories or scores, the allowed answers. The API returns a typed result and probability information. OpenAI says this is about 10 times faster than the Responses API. AIC has not benchmarked that claim.
Try a yes/no check, such as whether a support request describes a damaged item, and use its estimated probability to decide when a person should review it.
Route an inquiry among departments or task queues using fixed choices, with an “other” option for cases that do not fit.
Score urgency against levels you define, then set thresholds using labeled examples from your own workflow.
This is an API for a builder to wire into an application. It currently supports gpt-6-luna only.
Most automation fails in the handoff between noticing something and choosing what happens next. Decisions API puts that choice in a small, explicit step. You define possible outcomes before the model sees the input, then keep uncertain cases available for human review. The test is whether its probabilities lead to better decisions on your own examples.
TOOL SPOTLIGHT
2️⃣ Claude Can Work Inside The File You Have Open
Image from Claude
Anthropic has put Claude in a sidebar for Google Docs, Sheets, and Slides. It can answer questions about the open file and make edits there, with changes recorded in Google version history.
The new Claude for Google Workspace puts the assistant beside the document, sheet, or deck a person is already editing. After a Marketplace install, open Extensions, Claude, then Open Claude. The beta requires a Claude Pro, Max, Team, or Enterprise plan, plus permission from the Google account's Workspace admin to install the app.
In Docs, ask Claude to tighten a long brief, reorganize sections, or turn notes into a first draft.
In Sheets, ask it to explain or repair a formula, clean inconsistent data, or build a model in the open workbook.
In Slides, ask it to split a crowded slide, add a slide, or restyle an existing deck.
By default Claude reads only the file that is open. It can use other connected sources only if the reader enables those connectors.
This is most useful when the source material and the finished work belong in the same file. A team could shorten a brief, inspect a spreadsheet formula, or clean up a deck without rebuilding it in another chat. Keep the default “Ask before edits” mode for consequential documents and check version history after a larger change.
🔥 What 120 Chapters Are Building Through October 31
You just read about two tools. Here is somewhere to point them. Hack for Humanity is four 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
🧰 Two Skills I Keep Reaching For
A lot of times, I want an agent to ask better questions or do better research before it makes anything. These are two skills I keep coming back to.
1. Grill Me, by Matt Pocock
Grill Me interviews you until a fuzzy idea becomes a clearer decision.
Use it when:
You have a product, project, or piece of writing in mind, but have not settled what it should do.
Your agent is ready to build and you still have questions to answer.
2. Last 30 Days by Matt Van Horn
Last30Days searches recent discussion across Reddit, X, YouTube, Hacker News, Polymarket, and the web, depending on which sources you have enabled.
Use it when:
You want to hear what people are actually saying about a topic.
You need fresh examples or angles before you write or build.
After installing a skill, ask your agent, “What phrase or slash command turns this on?” In agents that support slash commands, try /grill-me or /last30days [your topic]. Then you can call the right skill as soon as you need it.
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.
🧪 LAB
Researching what frontier AI can do and how to keep it safe.
RESEARCH ENGINEER, CLIMATE MODELING - Allen Institute for AI | Seattle, WA | $146,880-$220,320
ALIGNMENT RED TEAM - RESEARCH ENGINEER/RESEARCH SCIENTIST - AI Security Institute | London, UK | £65,000-£145,000, level-dependent
RESEARCH SCIENTIST, APPLIED WHITE-BOX METHODS - FAR.AI | Berkeley, CA; Hybrid (US) | $150,000-$250,000
🏢 INDUSTRY
Putting AI into products and the work around them.
APPLIED AI ENGINEER - Laurel | Remote, US; Hybrid in SF, LA, or NYC | $254,000-$328,000 + Equity
STAFF SOFTWARE ENGINEER (AI GOVERNANCE) - Coder | United States, Remote | $193,000-$294,000 by location tier + Equity + Bonus
SENIOR STAFF SOFTWARE ENGINEER - AI GOVERNANCE - OneTrust | San Francisco, CA | $193,050-$257,400
⚖️ POLICY
Working on the rules and institutions around AI.
MODEL POLICY MANAGER, AGENTIC SAFETY - OpenAI | San Francisco, CA | $207,000-$335,000 + Equity
POLICY ADVOCACY - POLITICAL RESEARCH - Anthropic | San Francisco, CA | $230,000-$265,000
SENIOR SUBJECT MATTER EXPERT (SME) - AI GOVERNANCE & SECURITY POLICY - RIVA Solutions | Alexandria, VA | $175,000-$200,000
✨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.