Welcome back, Superhuman. Anthropic’s stock starts trading today in one of the biggest IPOs ever, but a new coalition of critics isn’t exactly popping champagne. Plus, a new startup claims it can test your health with just a few drops of blood. Does that sound familiar to anyone else?
Today: Gemini drops Neon, how to animate yourself into a reaction GIF with ChatGPT Images, and get the latest prompts and trending social posts.
TODAY IN AI
1. A startup claims to measure your health using a few drops of blood: Helios analyzes roughly 1,000 proteins in your blood to learn where your health is headed, claiming it can even spot issues before symptoms appear. It works like this: you collect a blood sample using their proprietary device, then receive insights from the Helios AI coach. The startup is founded by PhDs Omar Abudayyeh and Jonathan Gootenberg, along with Yousef Badran. Currently waitlist-only. See how it works.
2. Google reveals another frontier model, its third in about a week: Google just released Gemini 4 Neon, a model that sits between Carbon, Google’s most powerful model, and Hydrogen, its faster, cheaper option. Neon ranks #3 on the Artificial Analysis Intelligence Index, falling below Claude Fable 5.2 and GPT-6.2 Astra. It’s live now in AI Studio, the Gemini API, Antigravity, and the Gemini app for paid Pro and Ultra subscribers.
3. OpenAI rolls out Live Dictation for ChatGPT, plus an accounting plugin: Dictation in ChatGPT now turns your speech into text in real time as you talk, instead of waiting until you’re done to start transcribing. It’s live on all plans across iOS, Android, web, and desktop. Separately, the lab rolled out ChatGPT for Accounting, a plugin that connects to popular accounting tools to complete invoicing, month-end reporting, and other workflows.
PRESENTED BY DEASY LABS
In our benchmark, that came from one change: removing duplicate files. Deasy scans and enriches your files before they go to AI, catching old versions, contradictions, and sensitive details before they become answers. All in minutes.
Publish to Databricks, Snowflake, or Collibra
Push metadata back into SharePoint so Copilot finds the right file
Build a queryable vector database
TUTORIAL
How to animate yourself into a reaction GIF with ChatGPT Images
Open ChatGPT > Explore > Images, then upload a clear photo of yourself.
Ask ChatGPT to animate the uploaded photo into a looping GIF. ChatGPT Images new update allows you to turn a prompt or image into an animated GIF.
Sample Prompt: “Turn me into a seamless reaction GIF showing [reaction]. Keep my face, hairstyle, outfit, and background consistent. Use 6 clear animation beats: neutral > notice > react > exaggerated peak > recover > return smoothly to the first pose. Keep the camera fixed and make the loop seamless.”
Specify the reaction, motion intensity, camera framing, and background behavior so only the intended elements move.
Open the generated GIF and adjust the playback speed until the reaction feels natural and readable.
If the face, pose, or loop drifts, describe the exact change in the same chat and regenerate the GIF with tighter consistency instructions.
FROM THE FRONTIER
A coalition of researchers, founders, and scholars rally against Anthropic right as it goes public
Today is Anthropic’s big day. In just five years, the lab has gone from unknown start-up to one of the largest IPOs ever — but not everyone is celebrating. In a new open letter, a coalition of tech and academic leaders argue that Anthropic is not a safe company.
The MisAnthropic Statement claims the lab’s actions don’t match its branding. They argue that Anthropic positions itself as a safe AI company while also chasing recursive self-improvement, which, if achieved, could cause the tech to spiral out of human control. Nearly 90 leaders have signed the statement, including:
Thomas Wolf: Co-founder, Hugging Face
Igor Babuschkin: Co-founder, xAI
Sheldon Lee Glashow: Nobel Laureate in Physics, Harvard
Oliver Hart: Nobel Laureate in Economics, Harvard
Dan Hendrycks: Executive Director, Center for AI Safety
Some facts support the argument. Anthropic CEO Dario Amodei penned a personal essay in mid-September, urging frontier labs to pace AI’s progress. Since then, his lab has shipped five new models, including Opus 5.5 and Fable 5.2, along with about a dozen new products.
Anthropic’s never shied away from its risks, though. The lab dedicated more than 90 pages of its IPO filing to potential risk factors, roughly double the norm. It also consistently updates a frontier safety roadmap. Now that Anthropic is public, expect its actions to come under even more scrutiny.
PRESENTED BY INCOGNI
Scammers don’t guess your number — they buy it from data brokers selling your phone, home address, and relatives’ info. That’s how a random call turns into phishing or impersonation.
Incogni removes your data from these databases and keeps removing it automatically, so your number and address stay off these lists.
IN THE KNOW
What’s trending on socials & headlines today

Meme of the day
🚚 Poppin’ Wheelies: Have you ever seen a 5-ton construction vehicle pop a wheelie? Now you have. The craziest part is that a human isn’t driving — AI is (1.6M views).
🍿 Artificial’s Trailer: Andrew Garfield is back. The actor who portrayed Facebook co-founder Eduardo Saverin is starring as Sam Altman in an upcoming movie that already looks Oscar-worthy. Watch the trailer.
📝 103 Use Cases: This post can reinvent how you use AI. Morning Brew co-founder Alex Lieberman wrote a monster list of 103 ways to use AI, most of which were impossible a year ago (3.6K bookmarks).
🖌 Fable Paints: Microsoft Paint was notoriously impossible to draw with. But not for Fable 5.2. The LLM crafted this masterpiece in just 2.5 hours after someone handed it the virtual paintbrush (6K upvotes).
🧮 Accounting Bots: SpaceXAI claims to have five Grok Bots working on its accounting team, and shared the secrets behind how they trained and onboarded the agents before letting them work independently.
PRODUCTIVITY

🌟 Jason Harness*: Open-source tool that turns your AI agent into a 9-role outbound sales team. It does the work, and you approve every message.
🔵 BlueDot Phone Calls: An AI-powered notetaker for your phone calls
👀 Squint: Explain anything on your screen by simply dragging a box beside it.
🪼 Smart Jelly: An AI agent that works through your tasks overnight.
🍋 Lemon: A voice-powered AI agent that turns voice into done tasks.
PROMPT STATION
Turn every recurring task into a system
Use Case: Find the repeatable tasks in your work day and create templates, triggers, SOPs, or automations for each.
Prompt: Act as a productivity strategist and operations designer. Turn my recurring work into simple, repeatable systems that reduce decision fatigue, save time, and ................IMAGE
Turn a casual office selfie into an editorial team photo

Use Case: Take a casual group picture and transform it into something worthy of a company website without having to hire a professional photographer.
Prompt: [upload reference photo] Use the uploaded casual group photo as the identity and composition reference and transform it into a polished editorial team portrait suitable ..............EXTRAS

Want more?
Learn how to use AI at work with our AI Academy
Check out our Top Prompts and Top 125 AI Tools
Level up with our newsletters for marketers, programmers, cybersecurity and IT
Grow customers & revenue: Join companies like Amazon, HubSpot, and Salesforce. Showcase your product to our 2 million+ readers and followers on socials. Get in touch.
What did you think of today's email?
Until next time — Zain, Theodore, & the Superhuman AI team.






