🤯 AI built cancer vaccine saves a dog

One guy, ChatGPT + protein-folding AI, and a dog with months to live. The story sounds fake, but it actually happened. An Australian tech worker adopted a rescue dog that was diagnosed with aggressive cancer.
Veterinarians said the dog likely had only a few months left.

Instead of giving up, he tried something extreme. He essentially attempted to build a custom cancer vaccine himself.

What he did:

• Paid about $3,000 to sequence the tumor’s DNA
• Used ChatGPT to analyze the mutations
• Used AlphaFold to model the cancer proteins
• Identified potential drug targets
• Designed a personalized mRNA cancer vaccine

Important detail: He had zero formal biology background.

After months of regulatory paperwork and veterinary approvals, the experimental vaccine was finally produced and injected. What happened next surprised everyone.

Within weeks:

• The tumor began shrinking dramatically
• The dog’s health started improving

Meanwhile, major pharmaceutical companies are spending billions running clinical trials trying to achieve the exact same thing, personalized cancer vaccines.

AI is starting to give individuals access to tools that used to exist only inside research labs. And this might be a preview of where medicine is heading: AI-assisted, fully personalized treatments designed for a single patient.

Source.

@aipost 🏴

⚠️Elon Musk just delivered the clearest death sentence for every hybrid company on the planet.

Musk: “One laptop with a spreadsheet can outperform a skyscraper of several hundred human computers. Now, if even a few cells in that spreadsheet were done manually, you would not be able to compete with a spreadsheet that was entirely a computer.”

One biological operator in a digital workflow throttles a supercomputer down to the speed of human typing. A hybrid company is a digital spreadsheet waiting on a human to do the math. The fully algorithmic entity demolishes the hybrid model because it operates at total computational velocity with zero biological friction.

Musk: “What this means is that companies that are entirely AI will demolish companies that are not.”

The greatest delusion of the current business cycle is the belief that traditional companies will slowly and safely transition into the AI era. There is no transition. There is replacement. Your competitor is a fully autonomous network executing decisions in milliseconds. Your company still requires a human to approve an email.

Your survival rate is exactly zero. Today’s enterprise is proud of its massive headcount. Tomorrow’s winner is horrified by it. The future Fortune 500 won’t be companies with a hundred thousand employees. It’ll be trillion-dollar entities run by a handful of operators and an army of autonomous AI agents.

The laptop already won. The skyscraper just doesn’t know it’s empty yet.

@aipost 🏴

Australian tech entrepreneur Paul Conyngham explains how he used ChatGPT/AlphaFold (spent $3,000 with no biology background) to create a custom MRNA vaccine to treat his dog’s cancer tumors.

@aipost 🏴

🧠 GPT-5.4 just “solved” a 20-year-old math problem… by finding a forgotten paper

A new story circulating in AI circles claims GPT-5.4 cracked a decades-old math problem from the FrontierMath benchmark, a set of research-level problems designed to challenge top mathematicians. But the twist makes the story even more interesting.

Instead of inventing a completely new proof, the model found an obscure 2011 preprint paper that already contained the key idea. The benchmark author didn’t know the paper existed, so the problem had been considered unsolved.

What happened:

• Researchers tested GPT-5.4 Pro on FrontierMath, one of the hardest math benchmarks
• The model solved a Tier-4 research problem, something previous AI systems couldn’t do
• While reasoning through the problem, it surfaced a little-known academic paper from 2011
• That paper contained the method needed to solve it

In other words: the AI didn’t invent new math, it rediscovered forgotten math.

Why this matters:

• FrontierMath problems can take weeks or months for mathematicians
• AI models solved 2% of them in 2024
• GPT-5.4-level systems now solve 38%

That’s a massive jump in research-level reasoning.

This might be the real superpower of AI: A system that can scan decades of academic literature, find forgotten ideas, and apply them instantly to unsolved problems.

There may be thousands of “unsolved” discoveries already hiding in old papers.

Source.

@aipost 🏴

🇨🇳 A robot dancing ballet.

@aipost 🏴

The robot thought his white shoes was the soccer ball 🤣

@aipost 🏴

Optimus serving water in Austin Downtown.

@aipost 🏴

🗣 Yann LeCun says today's humanoid robot demos look impressive, but the robots are very stupid in the real world.

"The missing piece is not hardware; it is AI that can't reason, plan, or adapt like humans"

Robot companies are betting AI will make them smart enough to sell at scale within 3-5 years.

It's a big bet.

@aipost 🏴

🗣 Marc Andreessen on future of jobs.

“Everybody wants to talk about job loss, but really what you want to look at is task loss.”

In 10 years, the key job will be instructing AI how to build products, not humans building them directly.

@aipost 🏴

🔔The alliance for secure AI launched jobloss.ai

A real-time tracker of AI-driven layoffs across the U.S. These jobs are disappearing. The numbers are growing and every single day.

@aipost 🏴

ChatGPT released skills for white collar work.

@aipost 🏴

❗️Ex-NFL player allegedly asked ChatGPT what to do after killing his girlfriend

Former Darron Lee is facing a first-degree murder charge and prosecutors say ChatGPT chats are part of the evidence. According to court filings, Lee’s girlfriend Gabriella Perpetuo was found dead in their Tennessee home in February.

But before calling for help, investigators say Lee allegedly consulted ChatGPT.

Prosecutors claim he asked questions like:

• what injuries from a fall look like
• what to do if someone is unresponsive
• how to explain the situation to police

Authorities argue the searches show he was trying to stage the death as an accident, possibly a fall in the bathroom. Medical reports cited in court say the victim had severe injuries inconsistent with an accident, and investigators say the scene appeared to have been cleaned up.

Lee has denied the allegations, and his defense argues the case is circumstantial. But prosecutors are now using the AI chat history as part of the timeline and intent evidence in the case.

Source.

@aipost 🏴

🔔 Anthropic announces The Anthropic Institute to address powerful AI's societal challenges as progress accelerates.

- Led by co-founder Jack Clark as Head of Public Benefit, it combines Frontier Red Team, Societal Impacts, and Economic Research.
- Focuses on AI's effects on jobs, economies, resilience, threats, values, and governance amid expected dramatic advances by 2028.
- Founding hires: Matt Botvinick for AI-law work, Anton Korinek for economic transformations, Zoë Hitzig linking economics to models.
- Unique access to frontier AI data for candid public reporting and external partnerships.

The Institute aims to inform responses to transformative AI's upsides and risks through interdisciplinary efforts.

https://www.anthropic.com/news/the-anthropic-institute

@aipost 🏴

Trump: The hot thing right now is AI.

@aipost 🏴

Something big is coming:

“A massive AI breakthrough is coming in the first half of 2026 and Morgan Stanley says most of the world isn’t ready for it.“

Morgan Stanley warns that a massive AI capability jump driven by unprecedented compute scaling at U.S. labs could arrive in early 2026, triggering rapid productivity gains, job disruption, and severe power shortages as intelligence becomes the key economic resource.

Source.

@aipost 🏴

In a NYT test, 54% of users preferred AI-generated text to human-generated text.

Source.

@aipost 🏴