Demis Hassabis: AGI is on the horizon, it is within our grasp; our societies must prepare for this unprecedented change!

If Demis is speaking so clearly about AGI, then it really is in the immediate vicinity.

AI Post ⚪️ | Our X 🏴

More raw footage of the robot people thought was CGI is coming out. Here you have people looking like they are running for their lives.

AI Post ⚪️ | Our X 🏴

📢 GPT-5.2 next week! (Dec 9th)

It should be at least as good as Gemini 3.0 (pro)

" Sources say that the 5.2 update should close the gap that Google created with the release of Gemini 3 last month, a model that topped leaderboards and wowed Sam Altman and xAI CEO Elon Musk."

The pressure on OpenAI is so high that they have to bring forward the release.

Source.

AI Post ⚪️ | Our X 🏴

At the 2025 Greater Bay Area New Economy Forum, Midea Group's CTO Wei Chang officially unveiled "MIRO U," a super humanoid robot that looks absolutely wild.

Billed as the industry's first six-armed, wheeled-legged humanoid, it's designed to "break through human physiological limits." It starts factory operations this month.

AI Post ⚪️ | Our X 🏴

📈 GPT-5 is falling more and more behind.

It is being overtaken by Claude, Gemini, and even Grok in certain domains. Yes, GPT-5 Pro is certainly an outstanding research model for novel scientific ideas and tasks.

But for the broad user base, the other models are surpassing it.

Source.

AI Post ⚪️ | Our X 🏴

⚡️ Googles "Titans" is absolutely nuts.

Titans is Google's new architecture type that gives language models something like a real long-term memory, while the model is running.

How? A deep neural network (MLP) acts as a "long-term memory" that is continuously updated while the model reads text. The model learns during the inference run itself what to retain ("test-time memorization"), instead of having everything fixed into the weights beforehand. With ~10 million tokens, it still maintains around 70% accuracy. Insane

Google is nailing it.

Source.

AI Post ⚪️ | Our X 🏴

When someone said it was CGI, EngineAI CEO Zhao Tongyang took a kick from the T800 head-on.

AI Post ⚪️ | Our X 🏴

⚡ Jensen Huang: The AI race is shifting from chips to concrete, power, and speed

Nvidia’s Jensen Huang says the new battleground isn’t model quality, it’s who can build and power hyperscale AI data centers the fastest. A top-end US AI facility takes 3 years from ground-breaking to supercomputer-ready, while China can erect massive structures in weeks and scale energy far faster.

Speed vs. Scale
• Huang notes China’s ability to stand up large buildings even hospitals in days, giving it a construction-speed edge.
• AI capacity is now limited by permitting, land, steel, grid hookups not algorithms.
• Countries with faster infrastructure cycles will dominate the next wave of AI buildout.

The power gap
• China has roughly 2× the total energy capacity of the US, despite a smaller economy.
• That means it can feed power-hungry AI clusters more easily, while US grid growth remains almost flat.
• Power, not compute, is becoming the bottleneck.

Chips vs. manufacturing strength
• Nvidia is still “multiple generations ahead” in AI chips.
• But Huang warns not to underestimate China’s deep manufacturing base, it can scale chip and system production faster than many expect.

The next phase of AI leadership may be won not by the best model but by the fastest builders of energy and infrastructure.

AI Post ⚪️ | Our X 🏴

🍏Apple’s biggest leadership reset since Steve Jobs and the blueprint for its post-2026 era

Apple is undergoing a rare top-tier shake-up: more than half a dozen senior executives are exiting, from COO Jeff Williams to CFO Luca Maestri, AI chief John Giannandrea, policy head Lisa Jackson, and legal chief Kate Adams. Bloomberg’s visualization shows a clear thinning of the “old guard” around Tim Cook and a deliberate rebuild.

The old team is rotating out
• Key operators (Williams, Maestri) and long-time policy/legal leads are retiring or leaving.
• Meta has poached multiple design + AI leaders, pulling ~100 engineers from Apple’s AI foundation models team.
• Net effect: Apple’s internal AI bench and institutional memory are weaker than a year ago.

A new leadership stack is forming
• John Ternus is emerging as the likely CEO successor, a shift toward hardware and product engineering instead of operational management.
• Amar Subramanya (ex-Google AI) is brought in to accelerate Apple’s sluggish AI roadmap, signaling a move toward execution-focused leadership rather than research-first.
• Stephen Lemay takes over interface design, a high-credibility builder who’s shipped nearly every major Apple UI since the first iPhone.
• Jennifer Newstead will run both legal and government affairs, centralizing Apple’s response to antitrust and global regulation.

The strategic direction
• This isn’t random churn, it’s a controlled transition toward a 2026-ready Apple built around:
• AI acceleration
• hardware + device engineering
• regulatory survival and global policy alignment

Apple is quietly replacing the iPhone-era leadership machine with a structure built for AI, geopolitics, and the next generation of devices.

AI Post ⚪️ | Our X 🏴

📊 ChatGPT has about 800 million monthly users, yet mobile use and paid plans have stalled while Gemini near 650 million users rides on default placement in Google search, Android, and work tools.

Multiple benchmarks for knowledge, math, and coding now put Gemini slightly ahead of OpenAI’s best models, so buyers no longer treat OpenAI as the automatic performance leader.

OpenAI reports about $13 billion 2025 revenue yet analysts project losses that could reach $140 billion by 2029, while Google and Microsoft make around $30 billion profit each quarter that can fund cheaper integrated features.

AI Post ⚪️ | Our X 🏴

The AI Race is changing.

AI Post ⚪️ | Our X 🏴

⚡️ DeepSeek V3.2 (38.2%) is currently the best Open source Model on Cortex-AGI benchmark.

Gemini 3.0 Pro is the top ranker with 45.6%. Cortex-AGI measures how well AI models can perform abstract, out-of-distribution reasoning on procedurally generated logic puzzles across 10 increasingly complex levels, without relying on memorization.

It also measures and compares the performance of proprietary models against open-source models under this rigorous setting.

AI Post ⚪️ | Our X 🏴

🔥 Grok spots what the ER missed and saves a man’s life

A 49-year-old man says Grok saved his life after an ER doctor misdiagnosed his near-ruptured appendix as acid reflux. In severe pain, he asked Grok for help. The AI flagged “perforated ulcer or atypical appendicitis” and told him to return immediately and demand a CT scan.

He followed the advice, pushed for the scan, and doctors found an appendix minutes from rupture. Surgery happened six hours later and he woke up pain-free. He didn’t tell the hospital that an AI guided him, saying he feared they’d dismiss it. The story went viral, with many arguing this proves AI can catch what overwhelmed doctors miss and some even saying they’d trust an AI doctor if it meant better care.

Elon has long predicted AI-driven medicine would arrive fast; this incident suggests it already has.

AI Post ⚪️ | Our X 🏴

Google founder Sergey Brin said he is back in the office every day because the pace of AI progress is the most compelling he has seen in his career.

He wants to be directly involved as Google rebuilds its stack around Gemini and TPUs.

AI Post ⚪️ | Our X 🏴

SPACE AI?

Elon Musk just said this: “Satellites with localized AI compute, where just the results are beamed back from low-latency, sun-synchronous orbit, will be the lowest cost way to generate AI bitstreams in <3 years”

AI Post ⚪️ | Our X 🏴

GITAI robots cooperatively assemble a 5-meter tower, a building block for future off-world habitats, on its own.

The combination of AI and robotics creates the necessary technological breakthrough and acceleration that we have always hoped for.

AI Post ⚪️ | Our X 🏴

This is one of the funniest things I’ve seen today

It appears like the tele - operator was taking off his headset before he disconnected it 😂

AI Post ⚪️ | Our X 🏴

📢 “AI surveillance” in the U.S. was actually cheap offshore labor

What looked like cutting-edge policing tech has turned out to be a global sweatshop. Flock, the largest provider of “AI-powered” cameras for U.S. police, was barely using AI at all. Instead, much of the work was done manually by low-paid freelancers in the Philippines.

• The workers handled everything: reading license plates, identifying car makes and colors, tagging pedestrians, and even transcribing accident audio.
• Cities bought these systems expecting automated intelligence but got human eyes quietly scanning American streets.
• The revelation raises serious questions about data security, law-enforcement transparency, and how many “AI” products are really powered by hidden labor.

AI Post ⚪️ | Our X 🏴

🔥 NVIDIA just pulled off a an amazing stung using a tiny 4B model that beat far larger systems on ARC AGI 2, 29,72% / $0.20 per task!

By leaning on synthetic data and test-time training instead of brute-force scale, the NVARC team proved that clever design can outpace raw parameter count. It’s an exciting signal that efficient, adaptive reasoning might be the real frontier in AGI progress - not just ever-bigger models.

•27.64% accuracy on the official ARC-AGI-2 leaderboard

• Uses a 4B-parameter model that beats far larger, more expensive models on the same benchmark.

• Inference cost is just $0.20 per task, enabled by synthetic data, test-time training, and NVIDIA NeMo tooling.

Source.

AI Post ⚪️ | Our X 🏴

🚨 Trump says U.S. will allow Nvidia to sell H200 AI chips to China under new rules

He says 25% of the revenue will go to the U.S., argues this will boost American jobs and manufacturing, and criticizes Biden for forcing “degraded” chip designs.

He adds that newer NVIDIA chips (Blackwell, Rubin) aren’t part of the deal, and that a similar approach will be used for AMD, Intel and other U.S. chipmakers.

AI Post ⚪️ | Our X 🏴

This is actually the biggest danger: that workers still have to keep it a secret that they use AI at work because so many employers still think AI is still at the GPT-4 level.

This has to change.

AI Post ⚪️ | Our X 🏴