Researchers at the University of Maryland have found that Polish is the most effective language for interacting with artificial intelligence.

In a large-scale study evaluating leading AI models including ChatGPT, Gemini, Qwen, Llama, and DeepSeek, across 26 languages, Polish achieved the highest accuracy rate at 88%.

According to the findings, the structure of Polish makes instructions clearer and less ambiguous for AI systems to interpret.

AI Post 🪙 | Our X 🥇

Michael Burry, the investor who famously foresaw the 2008 financial crisis, is once again taking a bold bearish stance — this time against the AI sector.

According to his latest 13F filing, Burry has taken put options on Nvidia ($NVDA) and Palantir ($PLTR) worth about $1.1 billion, representing roughly 80% of his total portfolio — a massive wager that the AI boom could be overextended.

Burry, who began shorting the U.S. housing market two years before its collapse, is known for his patience in spotting bubbles early. Nearly two decades later, he seems to believe that artificial intelligence may be the next one to burst.

All eyes are now on whether his latest contrarian bet will play out as dramatically as his first.

AI Post 🪙 | Our X 🥇

Yuta Lee says we're on the cusp of no longer living with our anxiety of death.

As longevity escape velocity nears, new generations will grow up without it, and philosophy and religion will have to find new ground.

AI Post 🪙 | Our X 🥇

🔔 Google just announced AI data centers in space, powered by TPUs launching 2027

Project Suncatcher: Solar-powered satellites at 650km altitude running AI workloads directly from the Sun. The physics advantages are definitive—8× more solar capture with no atmosphere or nighttime, vacuum cooling that eliminates 40% of ground data center costs, and radiation-hardened TPUs tested to 15× mission requirements.

Google’s launching two satellites early 2027, each carrying four TPUs to prove it works.

AI Post 🪙 | Our X 🥇

A decade ago, few could have imagined NVIDIA would rise to become the world’s most valuable company. Yet in the era of artificial intelligence, its chips have become the essential tools powering both the training and operation of large language models, the digital equivalent of selling shovels during a gold rush.

Jensen Huang, NVIDIA’s founder and CEO, recognized the company’s pivotal role early on. When AI demand exploded, he moved quickly, securing key partnerships, expanding research and development, and scaling production capacity through TSMC to meet unprecedented global needs. While rivals such as AMD have benefited from the AI wave, none have come close to matching NVIDIA’s dominance.

NVIDIA’s story will be remembered as a defining chapter in tech history, and Jensen Huang as the visionary who saw the AI revolution coming and positioned his company at its center.

AI Post 🪙 | Our X 🥇

OpenAI has now signed $816 BILLION worth of AI deals this year alone.

At what point do they run out of money?

AI Post 🪙 | Our X 🥇

Coca-Cola will once again have its Christmas commercial generated entirely by AI this year.

When a billion-dollar company like Coca-Cola is doing that, just think about the pressure currently weighing on the creative industry.

AI Post 🪙 | Our X 🥇

Google is the only native full-stack AI company.

• Chips: TPUs
• Applications: Gemini
• Data Centers: Google Cloud
• Foundational Models: DeepMind

+ Edge AI: Pixel
+ Physical AI: Waymo
+ Ventures: Anthropic
+ Proprietary Data

All while their core business prints cash.

AI Post 🪙 | Our X 🥇

AgiBot deploys Real-World Reinforcement Learning (RW-RL) in industrial robotics with Longcheer Technology. Robots now learn new skills in tens of MINUTES (not weeks)

We are accelerating, faster and faster.

AI Post 🪙 | Our X 🥇

⚖️ One man used Claude to cut a $195,000 hospital bill down to $33,000

No lawyer. No hospital insider. Just a $20/month Claude Plus subscription.

He uploaded his itemized hospital bill, and Claude identified duplicate codes, illegal double billing, and charges explicitly banned under Medicare rules. The model then drafted a formal appeal letter citing each violation in regulatory language.

Result: the hospital reduced its claim by 83%.

It’s a glimpse of what AI can really disrupt not just labor, but information asymmetry. Hospitals, insurers, and law firms have long profited from complexity and fine print. Claude gave one individual the same analytical power as a compliance team.

AI isn’t just replacing workers, it’s starting to erase the opacity that kept systems unequal.

AI Post 🪙 | Our X 🥇

Xpeng's new humanoid robot looks more human than most other humanoid robots.

AI Post 🪙 | Our X 🥇

Apple is reportedly striking a $1 billion-a-year deal with Google to use its 1.2 trillion-parameter Gemini model to supercharge Siri’s long-awaited overhaul.

- The upgraded assistant - codenamed Linwood - will debut next spring, with Gemini handling Siri’s summarizing and planning features while Apple’s own models continue in limited roles.

- Behind the scenes, Apple is also racing to finish its own 1-trillion-parameter model by next year, aiming to replace Google’s tech once it catches up.

AI Post 🪙 | Our X 🥇

At Elon Musk’s AI firm xAI, female staff were reportedly required to submit their biometric data to help develop an anime-style avatar known as Ani. Their facial images and voice recordings were allegedly used to make the digital character appear more realistic.

Employees said participation was compulsory and that the terms were not clearly explained at first. In response to inquiries, xAI dismissed the allegations as “media lies.”

AI Post 🪙 | Our X 🥇

🔔 AI Turns Brain Activity Into Sentences

Researchers have created a “mind-captioning” AI that translates brain activity into full descriptive sentences, not just keywords.

Using fMRI scans and deep-language models, the system decodes what people see or imagine with surprising detail, effectively narrating mental scenes. Beyond its sci-fi appeal, the technology could unlock new ways to understand perception and help those with speech loss regain communication.

Can AI now read minds? 🤔

AI Post 🪙 | Our X 🥇

Robots like this, capable of cleaning your house and cooking, will be everywhere in 2–3 years.

AI Post 🪙 | Our X 🥇

It's a woman in a suit." No, it's this robot in a suit. "But it looks like a woman." Yes, that's why we call them 'humanoid robots', we design and build them, so they look as human as possible.

The model like walk we see in this vid is identical to what we saw yesterday. It’s a bit ridiculous that there are still so many people skeptical that such robots are real.

We’re living in late 2025, almost 2026, not 1986. We're very close to 2030. Our technology is pretty advanced.

AI Post 🪙 | Our X 🥇

🔥 Kimi K2 thinking, open-source agent that actually thinks

Moonshot AI has released Kimi K2 Thinking, an open-source model built for autonomous reasoning and tool use and it’s already setting new benchmarks.

🔹 Performance:
• SOTA on HLE (44.9%) and BrowseComp (60.2%)
• Handles 200–300 sequential tool calls without human input
• Specializes in reasoning, agentic search, and coding
• Comes with a 256K context window

🧠 Built for “test-time scaling”:
K2 isn’t just trained to predict tokens, it’s designed to think through problems by scaling both reasoning depth and tool-calling steps.

🌐 Available now:
• Chat mode: kimi.com
• API: platform.moonshot.ai
• Tech blog: moonshotai.github.io/Kimi-K2/thinking
• Weights & code: huggingface.co/moonshotai

Open-source agents are catching up fast, Kimi’s next challenge will be turning raw reasoning power into real-world autonomy.

AI Post 🪙 | Our X 🥇

Perplexity 🤝 Snapchat

Perplexity will be the default AI for all Snapchat users in the Snapchat app from January 2026.

AI Post 🪙 | Our X 🥇

David Sacks shuts down OpenAI CFO request for federal backstop

AI Post 🪙 | Our X 🥇

A statement from NVIDIA CEO Jensen Huang.

AI Post 🪙 | Our X 🥇