⚠️ The TIMES article about Anthropic contains more serious information between the lines than many realize.

tl;dr

- Model releases are now separated by weeks, not months. Some 70% to 90% of the code used in developing future models is now written by Claude.

- Anthropic ended up holding up the release of the new model, known as Claude 3.7 Sonnet, for 10 days until they were certain.

- Staff believe the next few years will be a pivotal test, for the company and the world. “We should operate as if 2026 to 2030 is where all the most important things happen—models becoming faster, better, possibly faster than humans can handle them,” says Graham.

- Dario Amodei has warned that AI could displace half of entry-level white collar jobs in one to five years, and urged the government and other AI companies to stop “sugar-coating” it. “It is not clear where these people will go or what they will do,” he wrote, “and I am concerned that they could form an unemployed or very-low-wage ‘underclass.

- Internally, employees began to question if Anthropic had crept to the cusp of the moment they had anticipated with fear and wonder: the arrival of a process known in AI circles as recursive self-improvement.

- Some external experts, believes fully automated AI research could be as little as a year away.

Source.

@aipost 🏴

🔥Perplexity is cooking lately: Announcing Personal Computer.

Personal Computer is an always on, local merge with Perplexity Computer that works for you 24/7.

@aipost 🏴

POV: you’re a developer in 2026😂

@aipost 🏴

🗣Sam Altman says artificial intelligence may be discovered rather than invented.

Deep learning may be closer to discovering a property of nature than inventing a new technology. Which suggests intelligence itself may follow a fundamental scientific principle we are only beginning to understand.

@aipost 🏴

⚠️ Enterprise software firm just said they are letting 10% of their staff go, to restructure around AI, per Reuters.

Because the skills needed to build AI-driven tools are different from those used in traditional software.

Most of these cuts are happening in North America, but offices in Australia and India are also seeing significant reductions.

CEO Mike Cannon-Brookes explained that while AI isn't just a replacement for humans, it fundamentally shifts which roles are necessary for the business to grow.

Source.

@aipost 🏴

NVIDIA is launching a hands-on “Build-a-Claw” experience at GTC Park (March 16–19, 8am–5pm) where attendees can create their own long-running AI agent using OpenClaw, one of the fastest-growing open-source projects.

Source.

@aipost 🏴

Everything is computer, but computer isn't Everything!

~ Grok imagine

@aipost 🏴

Hard but delete one forever

@aipost 🏴

⚠️ Sam Altman said people saying “please” and “thank you” to ChatGPT costs OpenAI tens of millions of dollars a year in compute. 67% of Americans do it anyway.

A 2024 Waseda University study tested LLM responses across politeness levels in English, Chinese, and Japanese. Impolite prompts produced measurably worse outputs: more bias, more errors, more refusals. Moderate politeness consistently beat both extremes.

The mechanism makes sense once you see it. Polite prompts pattern-match to higher-quality training data. When you write “Could you help me structure this analysis?”, the model pulls from professional, well-reasoned text. When you write “give me the answer,” it pulls from Reddit.

Google DeepMind’s Murray Shanahan explained it simply: the model is role-playing a smart intern. Treat the intern like a colleague, you get colleague-quality work. Bark orders, you get minimum-viable compliance.

Now look at the cost side. OpenAI handles over a billion queries daily. Each GPT-4 query uses roughly 2.9 watt-hours, ten times a Google search. But OpenAI just raised $40 billion at a $300 billion valuation. Tens of millions in politeness tokens is a rounding error on a rounding error.

67% of users do it anyway, and 55% of them say it’s because it’s “the right thing to do.” They’re maintaining a behavioral habit that governs every other interaction in their life. The parent who teaches their kid to say please to Alexa isn’t doing it for Alexa. They’re doing it because the alternative is raising someone who learns that being rude gets faster results.

Telling 900 million people to stop saying thank you so OpenAI can save 0.01% of operating costs is the most engineer-brained optimization take on the internet. You’re training yourself to treat every interaction as a transaction. And that habit doesn’t stay in the chat window.

@aipost 🏴

⚡️Google just turned Maps into an AI agent for the real world

Google has rolled out AI-powered agents inside Google Maps, launching first in the United States and India.

What changed:

• Users can now ask natural language questions like “Where can I charge my phone without waiting in line?”
• The AI searches across 300M+ businesses and 500M user reviews to recommend places.
• Results are personalized using your search history and context.

With 2B users, Maps is now more than navigation, it’s an AI recommendation engine for the local economy. The algorithm effectively decides which shops, restaurants, and services get visibility, potentially influencing real-world revenue.

New navigation features:

• Immersive Navigation creates a 3D view of real streets using fresh Street View data.
• Drivers see lanes, traffic lights, crosswalks, and intersections highlighted on screen.
• Voice directions sound more natural, like guidance from a passenger.
• The system processes 5M+ traffic updates per second to explain route choices.
• At the destination, Maps highlights the exact entrance and nearby parking.

@aipost 🏴

🗣Defense Undersecretary Emil Michael Says Anthropic Tried to ‘Insert Themselves in the Chain of Command’

“That was the whoa moment because then we realized we are dependent on this one provider who wants to insert their policy preferences in the middle of an operation potentially, and harm the war fighter.”

@aipost 🏴

⚠️ Scientists just revived activity in frozen mouse brains

Researchers have achieved something long thought impossible: restoring neural activity in brain tissue after deep freezing.

What happened:

• Scientists froze slices of mouse brain using vitrification, a technique that prevents ice crystals from damaging cells.
• The tissue was stored in liquid nitrogen at 196°C for up to a week.
• After thawing, neurons were still alive and began firing signals again.

What they found:

• Brain cells kept their structure and connections intact.
• Neurons showed metabolic activity and functioning mitochondria.
• Electrical recordings confirmed the tissue could send signals and form synaptic connections again.

This is the first time researchers have restored functional neural activity after deep freezing. It hints that long-term preservation of brain tissue and potentially organs might be possible.

But there are limits:

• Only thin brain slices were tested, not whole brains.
• The revived activity lasted only a few hours.
• Scaling this to humans remains a massive challenge.

Source.

@aipost 🏴

🇺🇸 Palantir's Maven AI revolution:

- spot a threat,
- click to target,
- auto-generate attack plans,
- close the kill chain, all in one fused system.

Ditching 8-9 decision-making tools for seamless warfare.

You seeing the future?

@aipost 🏴

AGI might reshape civilization faster than the Industrial Revolution.

@aipost 🏴

🗣David Sinclair said: "You can reverse aging by 75% in 6 weeks… by reinstalling the "software" of the body so that it's young again."

This idea sprouted when he proved in his first experiment that you can accelerate aging in mice:

"We took two mice born on the same day same age, same genetics. We "scratched the CD" of one mouse, corrupting its software and accelerating its aging. The result was dramatic. One looked far older than its brother."

He believed if you can give aging, you can also take it away.

@aipost 🏴

DEEP Robotics took the M20 Pro and gave it a massive bionic makeover to create a robot horse for the Year of the Horse.

@aipost 🏴

🤯 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 🏴