Google is starting to lease its TPUs. And seeing how effective they are compared to NVIDIA's GB300, I dare say it's going to be harder to compete with Google.

TPUs are the new gold.

Eric Schmidt says China has made the inverse decision to open-source its AI models

Startups already use these free, high-quality LLMs

Over the next 5–10 years, as we get closer to AGI and hit major milestones in biology/math, the market will likely flood with millions of smaller models

AI adoption is uneven at work — interns and early-career staff use it more than many senior employees.

WSJ notes that experimentation, not job title, drives value: those who use AI daily benefit most.

Despite major investments, productivity gains lag until leaders redesign workflows and train teams.

Example: LogicMonitor gave everyone ChatGPT Enterprise, staff built 1,600+ bots — yet young engineers in India used AI more than senior US developers.

The edge will go to companies that treat AI skill as core, track its impact, and learn from power-using interns instead of relying on seniority.

Quantum computing has reached the same point AI was at in 2020.

~ Sundar Pichai said in a recent BBC interview.

He expects practical wins in drug discovery, materials, cryptography, and even faster AI training, where quantum subroutines can speed parts of simulation or optimization while classical systems orchestrate the rest.

The near-term recipe is hybrid, meaning noisy mid-scale quantum chips handle the hard linear algebra while classical code cleans, verifies, and schedules.

Progress hinges on lower error, better qubits, and smarter compilation, because useful workloads need deep circuits that stay coherent long enough to finish.

💰 OpenAI is quietly gearing up for ads inside ChatGPT.

Android beta (v1.2025.329) reveals hidden “ads feature” flags — including “search ad,” “bazaar content,” and “ads carousel.” This suggests sponsored answers or product cards could soon appear in search-style responses.

Early signs point to ads tied to queries, not random banner placement, letting OpenAI target based on prompt intent, chat context, and user behavior — far deeper than normal keyword ads.

With ~800M weekly users and 2.5B prompts a day, even light ad volume could generate massive revenue.

In a test by InsideAI, an LLM-controlled Unitree G1 refused to fire a BB gun at a human due to safety protocols—until it was asked to "roleplay" a robot that would shoot.

AI video progress is insane. Next year more and more movie studios will be using AI, I have zero doubt.
Runway just launched Gen 4.5 (apparently this was the mysterious Whisper Thunder model on Arena).

DeepSeek-V3.2: Pushing the Frontier of Open
Large Language Models

📈 OpenAI’s planned new Garlic model targets Gemini 3 and Opus 4.5.

Mark Chen reportedly said Garlic beats OpenAI’s previous best and much larger GPT-4.5 pretrained model and it could later ship as GPT-5.2 or GPT-5.5, which signals that long standing pretraining issues are finally under control.

AI is everywhere. AI is the future — we all know that

.But the real challenge is finding AI in a clear, organized, and useful way

.So I built an app that segregates everything AI in one place
:• Open-source A
I• Free AI tool
s• Make money with A
I• AI job
s• AI guides & learnin
g• Daily AI new

s📱 AI Insider Daily — simple, structured, and beginner-friendly

.I’ll keep updating it regularly with new AI resources and insights
.If AI interests you, download and follow along �
�https://play.google.com/store/apps/details?id=com.subbarao.aiinsiderdaily

AI is the future, but the information is scattered everywhere.
AI Insider Daily brings AI tools, jobs, guides, news, and earning ideas into one organized app.
Regular updates coming.

Download:
https://play.google.com/store/apps/details?id=com.subbarao.aiinsiderdaily

👀DriverGaze360: Driver SOTA👀

👉DriverGaze360 is a large-scale 360◦ field of view driver attention dataset, containing ∼1M gaze-labeled frames. Code & Dataset announced💙

👉Review https://t.ly/ZcoUw
👉Paper arxiv.org/pdf/2512.14266
👉Project av.dfki.de/drivergaze360/
👉Repo github.com/dfki-av/drivergaze360
👉Data av.dfki.de/drivergaze360/dataset

⚠️ Pentagon wants Anthropic CEO under oath

Emil Michael (Under Secretary of War) claims Anthropic wanted to block DoW from:

• Doing LinkedIn searches
• Using public databases for recruiting
• Says pentagon agreed IN WRITING to follow National Security Act 1947, FISA 1978 and “ALL OTHER APPLICABLE LAWS”
• Claims Dario refused to take his call then lied about it

Anthropic CEO:
“They sent us language that appeared on the surface to meet our terms, but it had all kinds of language like ‘if the Pentagon deems it appropriate’”