This website uses cookies

Read our Privacy policy and Terms of use for more information.

In partnership with

God of Prompt: The Move List

Find what’s slowing your GTM.

Most GTM problems aren’t one big failure. They’re one or two weak spots quietly dragging down the rest.

The free GTM Audit scores your go-to-market across four dimensions, then sequences a 90-day roadmap for your stage.

Drop your URL. Get your score in about 10 minutes.

Welcome back. Wednesday was wild. Nvidia lost the founder who built the AI chip empire, Google made a near-top model free for everyone, and Anthropic's IPO is reportedly drowning in orders. Hot take: the tools you use every day just got cheaper and shakier at the same time. Have a backup model ready.

Today: Jensen Huang's exit, free Gemini 4 Hydrogen, Anthropic's ~$1T IPO, the researcher who says he can tell which chip trained a model, and a 20-minute test to find out if a free model can do your paid one's job.

TODAY IN AI

1. 💥 Jensen Huang quits as Nvidia CEO: Huang resigned on Nov 11 as president, CEO and director, effective immediately. CFO Colette Kress is interim CEO. The stock fell 13% at the open and closed down 4.3%. Reports tie the exit to an alleged affair with a woman US agencies suspect was a Chinese intelligence recruit; none of it is proven, and Nvidia says no secrets were compromised. Congress wants testimony, and over 1,600 employees reportedly signed a petition to bring him back.

Why it matters: the company whose chips run nearly every AI tool you use lost its founder overnight. Watch the Nov 17 earnings call: that's where price and supply noise will show up first.

2. ⚡ Gemini 4 Hydrogen is now free in the Gemini app and Search: Released Nov 12, it's the new default in Search, AI Studio and the API, and free users get it through auto model selection. API price is $1.50 in and $7.50 out per million tokens, with a 1M-token context. Google says it's 2.5x faster than Gemini 3.8 Flash; Artificial Analysis ranks it just below Sonnet 5.5 and GPT-6.1 Sol at about a third of Sonnet's cost. It's Google's third new model in six weeks.

Why it matters: near-top smarts now cost you nothing in the app. Run your hardest weekly prompt through it next to your paid model today (the tutorial below shows how).

3. 📈 Anthropic's IPO is reportedly ~6x oversubscribed at about $1T: Bloomberg reported on Nov 12 roughly six times more orders than shares on offer, days before listing. Price and date aren't confirmed. On the roadshow Dario Amodei said AI could cure cancer and Alzheimer's within 10 years; doctors pushed back, and the claim isn't in the prospectus. Same day, nonprofits sued OpenAI, Anthropic, Google, Meta and SpaceX to halt frontier AI.

Why it matters: the company behind the tool many of you use daily is about to answer to public shareholders. Plans and prices tend to follow. Not investment advice, just a reason to know your Plan B.

What this means for you: the AI stack is moving fast at the top. Don't bet your workflow on one model. Test a free backup this week.

You Build the Future. We Protect It.

WithCoverage replaces traditional brokers with AI-native insurance experts and attorneys. Get 40+ competing carriers, flat fees instead of commissions, and one platform as you scale.

Join 1,000+ businesses like Gopuff, Opendoor, and Polymarket.

Get a free, confidential policy review to uncover savings and better protect your business.

FROM THE FRONTIER

An OpenAI researcher says he can tell which GPU trained a model

The claim. Will Depue posted on Nov 12 that he found GPU-specific fingerprints in how model weights start out, by searching about six trillion chip-and-seed combinations against open-weights models. He claims DeepSeek, Moonshot, MiniMax and Ant Group trained on export-banned Nvidia Blackwell chips. The post hit 5.1M views.

The catch. It's one researcher, and nobody has independently reproduced it yet. It also lands as OpenAI reportedly joins Anthropic in backing tighter chip export controls on China, which is unconfirmed.

My call: open models are about to lose their anonymity

If the method holds, where a model was trained becomes checkable. That matters if your business runs a Chinese open model: expect compliance questions from clients and partners. I'd change my mind if independent labs can't reproduce it.

IN THE KNOW

𝕏 Watcher.Guru @WatcherGuru

Over 1,600 Nvidia employees sign a petition asking the board to bring Jensen Huang back.

♥ 31K · ~790K views · View on X

𝕏 Will Depue @willdepue

Claims forensic proof that DeepSeek, Moonshot, MiniMax and Ant trained on banned Blackwell chips, by fingerprinting the GPU in the weights.

♥ 35K · ↻ 7.3K · 5.1M views · View on X

𝕏 Kalshi @Kalshi

BREAKING: Dario Amodei says AI could cure cancer and Alzheimer's within 10 years.

♥ 16K · 1M views · View on X

TOOLBOX

🔗 n8n: build AI workflows you own. Send every meeting summary to your CRM and a follow-up email without touching it.

🎥 Arcads: write a script, pick an AI actor, get a UGC-style ad back. Test new angles without booking creators.

🕷️ Firecrawl: turns any website into clean text your AI can read. Pull a competitor's pricing page into a sheet in minutes.

TUTORIAL

Build a free backup model in 20 minutes: test Gemini 4 Hydrogen against your paid AI

1. Pick your 3 most-used prompts, the ones you run every week.

2. Open Hydrogen for free in the Gemini app (auto mode) or AI Studio, and keep your current tool open in the next tab.

3. Run each prompt on the same real input in both. Save both answers.

4. Paste both answers into the judge prompt below and let it score them blind.

5. Mark each prompt as "switch", "backup" or "keep paid". Now if your main tool changes price or goes down, you already know what works elsewhere.

You are a strict judge. Here is a task and two answers, A and B. You don't know which model wrote which.
Task: [paste the prompt]
Answer A: [paste]
Answer B: [paste]

1. Score each answer 1-10 on accuracy, following instructions, and how usable it is without edits.
2. List any facts, numbers or steps that are wrong or missing in either answer.
3. Pick a winner, or say TIE if the gap is under 1 point.
Be specific. Do not reward length.

PROMPT STATION

AI Vendor Risk Check

When news like this week's makes you wonder what breaks if one AI tool changes price, plans or goes down.

## Role
You are an operations lead who audits how a small business depends on AI tools.
## Task
Find where my business is exposed if one AI provider raises prices, changes plans or has an outage.
## Context
My business: [what you sell, team size]
AI tools I pay for and what each does: [tool, plan, monthly cost, tasks it handles]
Workflows that would stop without AI: [list]
## Output
1. A table: tool, tasks it covers, monthly cost, impact if gone (high / medium / low), free or cheaper backup.
2. My top 3 single points of failure, each with a 1-hour fix.
3. A one-week plan to test the backups on real work.
Be concrete. Name specific alternatives. Flag anything you are unsure about.

👉 Browse more prompts like this. Built for ChatGPT, Claude, Gemini, Grok.

Want one written for your business instead? Describe the task in a sentence and the prompt generator writes it, now UNLIMITED with your context profiles inside NjectAI.

IMAGE PROMPT

Bold Quote Card for Social

When you want a customer review or a strong line from your post to stop the scroll.

A square social media quote card.
Exact text: "[your quote, max 12 words]" and below it, smaller: "— [name or brand]". No other words.
Layout: the quote centred in large heavy sans-serif type, left-aligned lines, generous margins, a thin accent line above the name.
Colours: [background colour] background, [text colour] text, one [accent colour] detail.
Style: clean modern graphic design, flat, high contrast, readable on a phone.
No photos, no logos, no extra icons.

👉 More image prompts in the library. Works in ChatGPT, Grok Imagine and similar image models.

Step-by-step guides for Claude, ChatGPT and Gemini, written for people who run a business. Browse the guides

Hit reply and tell me which job you handed to AI this week. I read every reply.

🔑 God of Prompt

P.S. If you do one thing today, run your most-used prompt through free Gemini 4 Hydrogen and the judge prompt above. Whatever happens at Nvidia or Anthropic next week, you'll have a backup that already works.

Reply

Avatar

or to participate