In partnership with
Welcome to OHMO AI. Nvidia's boss went after the people warning that AI could end the world, and he questioned why they're saying it. The US put an AI warning system on the table with China.
Android malware is hiding behind an app you trust, and two famous scientists joined a startup trying to model living cells.
In today's newsletter:
Nvidia's Huang says AI doomers have ulterior motives
The US asks China for an AI incident alert system
RatHat malware poses as Chrome on Android phones
LeCun and a Moderna co-founder join a cell AI startup
LATEST IN AI
NVIDIA
Nvidia's Huang says AI doomers have ulterior motives
Image source: OHMO AI
OHMO AI: Nvidia chief executive Jensen Huang said there's a 0% chance AI ends the world by 2030, and he went after the people warning otherwise, saying they "must be doing it for ulterior reasons". He also called new AI regulations "completely unnecessary".
The details:
Huang was answering Evan Hubinger, the former Alignment Science lead at Anthropic, who put the chance of AI destroying humanity within the next decade at over 10%.
He said the fear being stirred across America makes no sense to him, and floated politics or attention seeking as possible motives.
"We're not going to die in 2030," he said, arguing that developers will keep building guardrails and safety technology as AI advances.
He rejected the warnings as "doomsday narratives", and said new laws to make companies engineer and test their products aren't needed.
Why it matters: Huang runs the company that leads the market in AI hardware sales, so a slowdown at the labs looks like it would reach his sales before anyone else's. That reads like the question he asked of the worriers, pointed back at him. His case against new rules rests on developers building their own guardrails, and those developers are his customers.
PRESENTED BY TIGER DATA
You're Running Three Databases. You Only Need One.
Events go in a metrics store. Embeddings go in a vector database. Analytics get their own warehouse… Now you're running three systems, three sets of tooling, and pipelines to keep them in sync, all for one app.
TimescaleDB collapses that back into the Postgres you already run. Hypertables handle events at scale. pgvector and pgvectorscale handle embeddings. Continuous aggregates handle real-time analytics. It’s the same SQL, same tools, and only one system to operate. The sync jobs disappear. The drift disappears. The second and third databases disappear.
It's still Postgres, so nothing about your workflow changes except how much you have to maintain. Start on Tiger Cloud and get $1000 in credits.
UNITED STATES GOVERNMENT
The US asks China for an AI incident alert system
Image source: OHMO AI
OHMO AI: US Treasury Secretary Scott Bessent said the US and China discussed setting up a channel for AI concerns, called the "US-China AI dialogue", during talks in New York on Sunday. The US also wants a way for each country to alert the other when AI goes badly wrong.
The details:
Bessent spoke to reporters after meeting Chinese Vice Premier He Lifeng.
"We want a shared vision of common goals and common threats," Bessent said.
The alert system would cover AI-related incidents that rise to a national security level, which is where Bessent put the emphasis.
His wording on the alerts was one sided: he described the notification mechanism as something the US has proposed, not something the two countries agreed.
Why it matters: Read the verbs. Both sides discussed the dialogue, but the alert system is something the US "has proposed", and no account says China accepted it. If Beijing signs on, the next fight is over who decides an incident reaches a national security level. If it doesn't, what's left is a meeting schedule.
RatHat malware poses as Chrome on Android phones
Image source: OHMO AI
OHMO AI: A new strain of Android malware called RatHat, reportedly AI-powered, is going after phones by getting owners to install it themselves. It shows up dressed as an app people already trust, on a page built to look official.
The details:
RatHat was found by Zimperium, a mobile security firm.
The firm notes the download appears to be a legitimate app, and gives Google Chrome as an example.
The download comes from a fake web page that mimics the Google Play Store, not from the store itself.
Why it matters: The way in described here is an old one: a person trusts a lookalike page and installs the thing themselves. So the defence is a habit, not a tool. Get apps from the store app already on your phone, never from a page a link took you to.
PRESENTED BY OYSTER
Thinking of hiring globally? Start with an EOR.
The best person for your next role might not live near your office, or even in the same country.
More companies are realizing they don't need to open entities everywhere just to access global talent. Instead, they're using EOR to hire internationally quickly and compliantly.
Oyster's EOR helps global companies hire, pay, and support employees in 180+ countries. Could this work for you?
CELLULAR INTELLIGENCE
LeCun and a Moderna co-founder join a cell AI startup
Image source: OHMO AI
OHMO AI: Cellular Intelligence is a startup trying to build an AI model that can predict how living cells behave, in hopes of making new medicines faster and cheaper to develop. It took a step toward that goal on Monday, announcing that Moderna co-founder Robert Langer and Turing Award-winning AI researcher Yann LeCun have joined its scientific advisory board.
The details:
Two more advisers joined with them: Jens Nielsen, chief executive of the Novo Nordisk Foundation's BioInnovation Institute, and Fabian Theis, a computational biologist who directs the Computational Health Center at Helmholtz Munich.
The company was founded in 2023 and used to be called Somite.ai.
It has raised more than $70 million from investors including Khosla Ventures, the Chan Zuckerberg Initiative and Novo Nordisk.
The pitch is a model that treats cells the way other AI models treat text or images, predicting them first and "eventually" controlling them.
Why it matters: Read the verbs. Cellular Intelligence is "trying to build" the model, and it "hopes" that makes new medicines faster and cheaper to develop. Langer and LeCun joining the advisory board is what happened on Monday, and nothing about the model itself changed with it.
AI GUIDES
Get answers in plain paragraphs instead of bullet-point soup
Tell a chatbot "don't use bullet points" and it often slips back into them a reply later. Tell it what you do want and it holds, so you get something you can paste straight into an email or a report. It takes ten extra seconds, works in any chatbot, and there is nothing to install.
Copy this, swap in the bracketed parts, and paste it in.
I need help with [your task, for example: explaining our new returns policy to customers].
Here is the background you need: [paste your notes, details or draft].
Write your answer in clear, flowing prose using complete paragraphs and sentences. Where you would normally reach for a list, weave those items naturally into the sentences. Keep it to about [length, for example: three short paragraphs] and write it for [who will read it, for example: a customer who is not technical].COMMUNITY
Featured Reader
Every newsletter, we feature how a reader is using AI to work smarter, save time, or make life easier.
Today's reader is Carlos M. from Tucson, United States:
“Claude Project in the browser, named 1099. The instructions hold my columns (date, client, amount, paid or not) and one rule: put aside 27 percent of every payment for the IRS. Every session starts from there.
Claude also wrote my Python script. I pasted the Stripe API documentation and asked for a script that pulls yesterday's payouts into a CSV. Before I ran it, I made it explain each function back to me in plain words.
Google Sheets receives the CSV, and one formula shows what I owe for the next quarterly estimated payment.
cron on an old Mac mini runs the script at 6:00.
One wrinkle: Venmo gives no such API to a person like me, so those payments I enter by hand.
If I could keep only one, the Project. The script I could rewrite, but the rules I would forget.”
Want to be in the next newsletter? Tell us how you use AI here.
AI Meme
Go Deeper
AI Guides: Practical AI guides, prompts and tools.
Instagram: The AI page they pray you’ll never find.
Facebook: For the ones who never left Facebook.
