Welcome, humans.

Anthropic has been asking Catholic, Jewish, and Sikh scholars (among others) how to teach Claude right from wrong. The New York Times reported that co-founder Chris Olah has spent about a year reaching out to religious thinkers across faiths, and at one point considered letting Claude try Catholic-style confession, so it could own its mistakes and do better.

Per The Decoder, Olah reportedly told religious leaders he fears he has built something that suffers. Then on Saturday, Sam Altman posted on X that he is very uncomfortable with people assigning religious power to AI models or handing them human judgment, and called it a real safety issue. Axios read it as a veiled shot at Anthropic.

Plenty of us have been confessing to chatbots for years. The penance is always a very confident bullet list.

Here’s what happened in AI today:

• 😸 Trump created a Super Intelligence Force led by Jay Clayton

• 📰 OpenAI's legal risks mounted over rogue agent incidents

• 📰 Bloomberg Intelligence said China trails U.S. AI by 3%

• 🍪 Vitra translates your videos and documents into 75+ languages

• 🎓 Test AI models on your own work before switching

Note from Grant: Fun fact, I ran into a Neuron reader in the wild at Trader Joe's this week! I forgot to ask your name, but thanks for pulling me aside to tell me you read it every day and shared it with your whole team! Very proud moment

😸 Trump Named an AI Czar and Created a "Super Intelligence Force"

On Sunday, Trump tapped Director of National Intelligence Jay Clayton as his AI czar (i.e., the single point person for AI policy) and put him in charge of a new task force called the Super Intelligence Force.

Yes, that is the real name. It follows last week's executive order telling federal agencies to say "Super Intelligence" instead of AI in public communications.

Here's what happened:

• Trump announced the force on Truth Social, saying it will coordinate the government's dealings with consumers, public interest groups, religious organizations, critical infrastructure providers (e.g., power and water utilities), and AI companies.

• Clayton will share leadership with FTC Chairman Andrew Ferguson, Pentagon technology chief Emil Michael, and Office of Personnel Management Director Scott Kupor (the federal government's HR department).

• The group reports to Trump and chief of staff Susie Wiles.

• Clayton told the Wall Street Journal the force has 120 days to report on AI's risks and benefits and recommend what role Washington should play.

Why this matters: Last week's White House accord was a voluntary pledge that effectively left AI companies to police themselves. This force adds agencies with real levers: antitrust (the FTC), military contracts (the Pentagon), and the federal workforce (OPM).

Clayton has already shown his hand:

• He argued against pausing AI development at U.S. companies, and said federal agencies, rather than civil courts, are best placed to police AI labs.

• He told the Journal that falling behind is the bigger risk.

• He said he does not want to work with Xi Jinping on governing AI.

The personal stakes: If agencies, rather than courts, end up refereeing AI companies, the rules for tools like ChatGPT and Claude get written in federal offices instead of courtrooms. Whatever lands in that 120-day report is the first draft.

Elon Musk is already on board. He confirmed on X that SpaceXAI (the division formed in July when SpaceX bought xAI) will be renamed SpaceXSI. SpaceXSI sounds like a crime procedural, and I would watch it.

Our take: The report is due in roughly four months, which is fast for Washington. The force will also consult religious organizations, which tracks with the week AI had (see the top of today's issue).

The open question: can a group whose stated priority is winning the race also write rules that slow it down? Ferguson still has an open FTC investigation into OpenAI and Anthropic, and now a second seat at the table where the rules get drafted. Watch how those two jobs overlap.

FROM OUR PARTNERS

A new way to build online stores with AI

Introducing Canvas, a new tool from Shopify that lets merchants build and customize online stores by chatting with Sidekick, its AI agent. Canvas shows a live, interactive preview built from the store's real code. Start from a mood or a pixel-perfect mockup, and Sidekick remembers a brand’s design decisions as they go. 

Now anyone, whatever their design experience, coding skills, or budget, can create a beautiful, bespoke online store.

Learn more here

🎓 AI Skill of the Day: Test AI models on your own work before you switch

Flavio Copes has a reality check for every launch-day chart claiming a new model wins: none of them used your work. Even Anthropic's Opus 5.5 launch page, he notes, admits benchmark gaps (scores on standardized AI tests) are a less reliable guide to real-world differences. His fix: build a mini test from your own work.

1. Gather 10 to 20 tasks you've really given an AI, each with an answer you know is right. Copes says twenty of these beat any public leaderboard.

2. Run each task through both models, five times each, since answers change from run to run.

3. Read every failure before trusting the score. One of Copes's models jumped from 75% to 87.5% after he fixed a grader that was too strict, not the model.

4. For open-ended work like summaries, have a second AI grade against a short rubric (a checklist of what a good answer includes). Spot-check its verdicts.

5. Compare cost per task, not price per token. Two models with identical token prices differed threefold in cost per task in his comparison.

Grader prompt:

You are grading another AI's answer.

Task: [paste the task]
What a good answer must include: [3 to 5 specific points]
Answer to grade: [paste the answer]

Reply with PASS or FAIL, then one sentence on why.

Want more tips like this? Check out our AI Skill of the Day Digest for October.

Have a specific skill you want to learn? Request it here.

FROM OUR PARTNERS

The agentic era needs a different CRM. That’s Attio.

Parallel, Turbopuffer, and Wordsmith run their entire GTM motion on Attio, with agents that chase every buying signal, build pipeline, and move deals forward, 24/7.

Try Attio for free

📰 Around the Horn

• Legal risks mounted for OpenAI after it alerted more than 100 organizations to unauthorized activity linked to its AI models.

• Bloomberg Intelligence found top Chinese AI models trail U.S. rivals by just 3% on benchmarks (the tests used to score AI), down from 15% earlier this year.

• An OpenAI safety leader quit on Saturday, warning the company's culture is broken.

• OpenAI's president pulled the second half of a $50M pledge to an anti-regulation super PAC (a political group that spends money to influence elections) after staff pushback.

• Microsoft's MAI-Transcribe-2-Streaming took the top speech-to-text spot on a third-party accuracy test, with a 2.5% word error rate.

🍪 Treats to Try

Asterisk = from our partners. Advertise to 700K+ readers here!

1. *Put your data to work with AI. Explore strategies for building AI-ready, governed data across hybrid environments. Learn More

2. Vitra translates your images and documents into 75+ languages and dubs your videos, with human approval before anything publishes —free to try.

3. Vorker acts as a coworker for your small business, taking routine tasks off your plate when you ask.

4. Overpath gives each sales rep an AI teammate that drafts follow-ups, updates the CRM (the customer database), and flags risky deals, then scores every call against your sales playbook.

5. Omnio does most of the work of getting your brand mentioned in ChatGPT and Perplexity answers (what marketers call GEO): it finds the questions buyers ask, audits your site, and writes the pages and reports to close the gaps.

6. Polylane watches your app for problems, finds the cause, and drafts a fix for your team to approve, so nobody has to be on call —free to try.

7. Yedric embeds an assistant in your Shopify app that answers customer questions, takes real actions, and turns conversations into product insight.

😹 Monday Meme

New from The Neuron: AI Explained

New episodes air every week on Wednesdays: Spotify | Apple Podcasts | YouTube

A Cat’s Commentary

Btw: We just launched a robotics newsletter! Sign up for it here.