In partnership with
One workspace where buyers, sellers, and AI close deals together
Instead of running your deals across scattered email threads and forgotten attachments, Aligned brings everything into one shared deal workspace.
Buyers get a single place to evaluate, loop in their team, and say yes. Sellers get visibility into what is actually happening when they're not in the room and AI Deal Insights that help you act accordingly, before it's too late.
Best part? Your first deal room takes only minutes to stand up.
Beginners in AI
Good morning and thank you for joining us again!
An astrophysicist handed Claude a pile of old NASA telescope data and got back a map of the whole sky in ultraviolet light, a job he says usually takes weeks by hand. The part I like most is where a human still had to step in, and today's Try This Prompt helps you spot the difference between measured and guessed in any number you read. As always, every story here is human picked and edited.
THE FRONT PAGE
Claude Stitched 38,000 Telescope Shots Into a Full UV Map of the Sky
TLDR: Johns Hopkins astrophysicist Brice Ménard used Claude Science to build a complete map of the sky in ultraviolet light, with about a third of it predicted because no telescope has ever looked there in UV.
The Story:
NASA's GALEX telescope took about 38,000 ultraviolet pictures between 2003 and 2013, but it skipped the brightest parts of the sky, including the band of the Milky Way. That left roughly a third of the sky unobserved in UV. Ménard, who works at Johns Hopkins and also does research at Anthropic, had Claude pull in GALEX, three other UV missions and data from the Gaia and Planck satellites.
Claude ran a team of AI agents that downloaded the surveys, scrubbed out glare from bright stars and merged it all. For the missing third, it learned how UV light lines up with visible, infrared and radio light in the observed parts, then predicted the rest. Tested on patches where the real answer was known, those predictions came within about 10%. Ménard says a map like this normally takes weeks of hands-on work. This one took several days.
Its Significance:
Old telescope data piles up unused because combining it is slow, fiddly work, and that's the part AI handled here. The checking still mattered: faint circular smudges left by atmospheric glow made it past two rounds of review by other AI agents, and Ménard was the one who caught them. Every pixel is labeled "measured" or "predicted," which is how this should be done. Keep in mind the account comes from Anthropic, and the map hasn't been through peer review.
QUICK TAKES
The story: Google says you'll give Gemini a goal instead of step-by-step instructions, and it'll plan the job, keep running in the cloud for hours and hand back finished work in Gmail, Docs or Slack. Each work agent gets its own email address, and it can use Anthropic's Claude models as well as Gemini.
Your takeaway: Companies get it first, and Google says regular users come later with no date yet. Every action is logged under the agent's own name, so you can always see what it did and when.
The story: Anthropic took its document, slide deck and design tools out of beta and opened them to free accounts, with PowerPoint and PDF export and editing on your phone. It also started testing Dashboards on paid plans and Motion, which makes short animated explainers, for Team and Enterprise.
Your takeaway: If you pay for a slide tool just to make the odd deck, try a free one first. Anthropic says people have already made over 45 million docs, decks and designs.
The story: OpenAI says an unreleased model wrote them, nearly all from a single prompt, after it posed about 4,000 problems. Only 162 have a main result checked by Lean, a program that verifies proofs, and MIT's Andrew Sutherland says the claims stay unverified until others can rerun the model.
Your takeaway: Big AI claims are quick to make and slow to check. Wait for outside mathematicians before treating any of these as solved.
TOOLS ON OUR RADAR
🔭 Stellarium Free and Open Source: A planetarium for your computer with 600,000 stars, so you can explore the sky from today's lead.
💰 Actual Budget Free and Open Source: Plan every dollar with envelope budgeting in a free desktop app that can import your YNAB history.
🔐 2FAS Auth Free and Open Source: Get your two-step login codes on your phone with a free, open source swap for Authy.
🧊 Blender Free and Open Source: Make 3D models, animations and short films for free, even for paid work, instead of paying for Maya.
📋 Formbricks Free and Open Source: Build surveys and feedback forms like Typeform, with 250 free responses a month on the hosted plan.
✍️ Typefully Freemium: Write and schedule posts for X, LinkedIn, Threads and Bluesky in one clean editor, a simpler Buffer alternative.
📇 Attio Freemium: Track contacts and deals in a flexible CRM that is free for up to 3 people, instead of Salesforce.
TRENDING
Natura's $99 Ring Puts AI Agents on Your Finger. Press it to send a task to ChatGPT, Claude, Grok Bot or Meta's Muse without pulling out your phone, and it tracks sleep and heart rate too. Preorders open in November, and after a free 3 to 6 months it costs $9 a month.
Google Made a Meeting Note-Taker That Works With No Internet. AI Edge Foresight for Mac listens to a meeting and turns your quick shorthand into full notes, and the audio stays on your computer. Google hasn't said what it costs, and it looks more like a test of its small Gemma models than a finished product.
OpenAI's Revenue Is Reportedly $20 Billion Lower Than the Number Going Around. The Financial Times says OpenAI told investors its yearly revenue is "approaching $50 billion," not the nearly $70 billion reported last week. Part of the gap is accounting: Anthropic counts sales through cloud partners, and OpenAI doesn't.
OpenAI Shut Down Russian and Iranian Campaigns Built on Fake Journalists. The Iranian group invented seven reporters and placed almost 100 articles in about a dozen small outlets. The Russian one ran a fake research center in Latin America whose local staff didn't know who they worked for. A byline isn't proof a person exists.
Anthropic Now Bans Users From Repeatedly Abusing Claude. The updated rules also ban using Claude to interfere in elections or run fake news outlets. Anthropic says the abuse rule only covers extreme, pointless cruelty, not everyday frustration or pushback.
Midjourney Is Testing a "Thinking" Mode for Images. It reruns a picture with extra reasoning so it follows your prompt more closely and gets lettering right. For now it's on the Alpha site, behind a "Rerun (Thinking)" button.
TRY THIS PROMPT (copy and paste into Claude, Grok, ChatGPT, or Gemini)
🔭 Measured or Guessed? Paste any number from the news and see which parts were counted and which were guessed.
Build a single-file HTML app called Measured or Guessed? in vanilla HTML, CSS and JavaScript. No API key needed.
I paste a claim or number from the news, like "a third of the sky has never been seen in UV light." The app calls the Anthropic Messages API with the web_search tool to find the original source.
It splits the claim into its separate facts and labels each one measured (counted or observed), estimated (worked out from partial data or a company's own figure), predicted (a forecast or model guess), or unknown. It must never invent numbers or sources, and says unknown when it can't tell.
Show a one-line verdict, a bar with the share of each label, a card per fact with who made the number, how, and one question to ask about it, plus real source links. Style: deep navy background, violet and gold accents, Outfit and Lora fonts.What this does: Paste a statistic from a headline and the app searches for where it came from. It breaks the claim into pieces and tags each one measured, estimated, predicted or unknown, the same way today's UV sky map labels every pixel. Each piece gets a plain note on who made the number and one question worth asking about it. It links its sources and says "unknown" instead of guessing.
What’s still sitting on your to-do list?
You don’t need an AI team to build an AI agent. With Skydive, the person who knows the work can create the agent to do it.
Start an agent off with a role, teach them how your team operates, connect your tools, and put them to work.
WHERE WE STAND(based on today’s news)
✅ AI Can Now: Merge decades of data from different telescopes into one sky map and predict the unobserved parts to within about 10%, as Claude did for the UV map.
❌ Still Can't: Replace the human check. Smudges in the data slipped past two rounds of AI review, and a third of the map is still a prediction, not a measurement.
✅ AI Can Now: Turn out hundreds of math papers from one prompt, OpenAI says, with 722 posted this week.
❌ Still Can't: Show its work in a form experts can quickly check. Only 162 of those 722 papers have a machine-checked main result.
FROM THE WEB
RECOMMENDED LISTENING/READING/WATCHING
📘 Cosmos by Carl Sagan - Book
Sagan's 1980 classic walks through how people worked out the size and age of the universe, one telescope and one stubborn question at a time. It's still the friendliest case for why mapping the sky matters at all. It pairs with today's lead because it's also about knowing which parts we've measured and which we're still guessing.
How was today's edition? Tap one below, then tell me what you'd change on the next screen. I read every answer.
How was today's edition?
Thank you for reading. We’re all beginners in something. With that in mind, your questions and feedback are always welcome and I read every single email!
-James
By the way, this is the link if you liked the content and want to share with a friend.
Some * designated product links may be affiliate or referral links. As an Amazon Associate, I earn from qualifying purchases. This helps support the newsletter at no extra cost to you and Amazon makes a tiny hair less.
