Quiet AI Alpha

AI Signal Today 1.Businesses are no longer optimizing only for Google. They are starting to optimize for AI answers. 2.Anthropic just deepened its compute dependence with Google and Broadcom, signing for multiple gigawatts of next-gen TPU capacity expected to come online from 2027. 3.Gemma 4 just crossed roughly 2 million downloads in its first week, and Google’s AI Edge Gallery shows local, on-device AI is moving out of toy territory. 4.OpenAI is broadening from model company to institution builder, releasing a people-first industrial policy paper and launching a new Safety Fellowship. The pattern across all four is clear: distribution is shifting compute is concentrating local AI is accelerating governance is becoming part of the product story

@sharbel 🙃

> you bought a TV > you aimed it at your couch > it aimed back > every second of every show > your TV has been taking thousands of screenshots of your screen per hour > matching them. logging them. selling them. > 78 million households > here's the full story:

@StartupArchive_ Really interesting perspective from Bezos. Choosing to focus on growing the pie for everyone instead of taking more for himself shows a long-term mindset that not many leaders have. Makes you think about what truly drives real impact.

Jeff Bezos explains why he didn’t take additional equity building Amazon Jeff is asked why he only paid himself $80,000 per year and never took additional equity during his tenure as CEO of Amazon. He responds: “I asked the comp committee of the board not to give me any comp. My view was I was a founder. I already owned a significant amount of the company, and I just didn’t feel good about taking more. I felt I had plenty of incentive. I owned more than 10% of the company, and earlier — before it was diluted by various things — more than 20% of the company. I just felt how could I possibly need more incentive?” Jeff continues: “Most founders own big chunks of the company. They’re more like owner-operators. The way they increase their wealth is not by getting more equity. They just want to make the equity they have more valuable. And so I just would have felt icky about it. And I’m actually very proud of that decision.” Jeff is especially proud of how much wealth he’s created for other people: “Somebody needs to make a list where they rank people by how much wealth they’ve created for other people — instead of the Forbes list where it ranks you by your own wealth. Amazon’s market cap is $2.3 trillion today. I own about $200 billion-ish of it. So if you take $2.3 trillion and subtract out the piece I kept for myself, then I’ve created something like $2.1 trillion of wealth for other people. That should put me pretty high on some kind of list. And that’s a better list — how much wealth have you created for other people?” Video source: @nytimesevents (2024)

This is exciting. Newton's GPU-accelerated physics engine could really speed up robot training. Open-source, differentiable, and already used by Disney. Looking forward to seeing how it advances real-world robotics. Great collaboration between NVIDIA, DeepMind and Disney Research.

🚨 NVIDIA, Google DeepMind, and Disney open sourced a physics engine that trains robots 70x faster than anything that existed before. It's called Newton. A GPU-accelerated physics simulator that teaches robots how to walk, grab, balance, and move like humans. Before they ever touch the real world. No proprietary licenses. No closed-door research. Three of the most powerful companies on Earth just gave this away. Here's what this thing actually does: → Simulates thousands of robots in parallel on a single GPU → 70x faster than existing robotics simulators → Robots learn to walk on snow, gravel, sand. Handle cups, fruit, fabric. → Differentiable physics. The robot improves itself through the simulation. → Works with MuJoCo, NVIDIA Isaac Lab, and Isaac Sim out of the box → Write simulation code in Python. Get CUDA-level GPU speed automatically. Here's the wildest part: Disney is already using this to build real theme park robots. The Star Wars BDX droids that walked on stage at NVIDIA GTC were trained with Newton. They move, gesture, and express emotion. Not CGI. Real physical robots. Disney's director of research said Newton "unlocks the potential of reinforcement learning for robotic systems of unmatched complexity." Robotics simulation software costs companies hundreds of thousands per year. Training a single humanoid takes weeks on traditional engines. Newton does it in hours. pip install newton 2.6K GitHub stars. 1,487 commits. Linux Foundation project. Apache 2.0 License. 100% Open Source.

@UnslothAI Crazy future ahead👀 Check these also : x.com/i/status/20414…

You can now fine-tune Gemma 4 with our free notebooks! 🔥 You just need 8GB VRAM to train Gemma 4 locally! Unsloth trains Gemma4 1.5x faster with 50% less VRAM. GitHub: github.com/unslothai/unsl… Guide: unsloth.ai/docs/models/ge… Gemma-4-E4B Colab: colab.research.google.com/github/unsloth…

@HowToAI_ OpenDataLoader PDF processes complex documents at high speed on CPU, supporting Markdown, JSON with bounding boxes, and HTML. It handles tables, formulas, and scanned PDFs via OCR, with strong benchmark results. Useful for local RAG setups. Any experiences with it?

🚨 Someone just open-sourced a tool that converts pdfs to markdown at 100 pages per second. It's called OpenDataLoader. It runs entirely on CPU and handles complex layouts, tables, and nested structures like a senior dev 100% Free.

@googlegemma Check these Guys also: x.com/quietaialpha/s…

@tibo_maker Insane growth to $274k MRR with pure organic + product-led magic The human-sounding articles update is exactly what everyone was begging for. Listening to feedback and shipping fast = why you're winning. Keep crushing it Tibo!

Outrank now at $274k mrr with 2,500+ websites using it 🤩 and we're not slowing down - rolling out a major update TODAY: 👉 articles that sound like humans wrote them + updated image models (not the average human btw, more like SEO boss) we read every piece of feedback from the last 6 months most of it said the same thing: make the content sound more human - so we fixed it 🫡 if you want SEO on autopilot, there's nothing else like it

@rubenhassid 🙂

How to set up Claude Cowork completely in 1 hour: (duplicate my exact folder, files, and prompts) 1. Download the Claude desktop app. 2. Click the Cowork tab at the top. 3. You need a Pro plan ($20/mo). Worth it. ----- → 0-10 min: Build your 'Claude Cowork' folder. Create "CLAUDE COWORK" on your computer. Inside: About me, Projects, Templates, Output files. To download mine, go here: how-to-ai.guide. Don't pay anything. It's free in the welcome email (The most important step, don't skip it) → 10-25 min: Write your .md files. about-me .md = who you are, how you work. anti-ai-style .md = every word you'd never use. To download my files, go to how-to-ai.guide. (These two files replace 500-word prompts forever) → 25-35 min: Set Global Instructions. Go to Settings → Cowork → Edit Global Instructions. Folder rules. Naming conventions. What Claude must read before every task. (You write this once. It runs every time) → 35-45 min: Create your first Cowork Project. Cowork tab → Projects → +. Pick a task you do every week. Scoped memory = it remember what it did last week (You stop re-explaining yourself) → 45-55 min: Run your first real task. Prompt: "I want to [task]. Ask me questions first." Claude generates clickable forms to prompt you. It creates a real .docx in your folder. (Stop prompting. Start directing) → 55-60 min: Schedule a task. "Every Monday at 7am, create my weekly briefing." You wake up to a finished doc. That's the endgame. Pro tip: Always add "Use AskUserQuestion" to every prompt. Claude prompts you instead. To download all of my other Claude infographics: Step 1. Go to how-to-ai.guide. Step 2. Subscribe for free. Don't pay anything. Step 3. Open my welcome email (most skip this). Step 4. Hit the automatic reply button inside. Step 5. Download my infographics from my Notion. Bonus. Enjoy my best copy-paste prompts, too.

@simplifyinAI Fascinating. Fiber turned into a silent listener. One endpoint access, and walls whisper secrets. Privacy just became fiber-thin. Stay vigilant.

🚨 BREAKING: Your internet fiber cable is secretly listening to you right now. Researchers from hong kong just dropped a paper at NDSS 2026 showing how they can spy on your conversations through the fiber optics in your walls. They successfully turned ordinary Fiber-to-the-Home (FTTH) cables into hidden, long-range microphones. No laser bugs. No physical implants. No drilling through walls. Just the broadband cable that is already sitting in your living room or office. By connecting a commercially available Distributed Acoustic Sensing (DAS) system to one end of the fiber, they can measure microscopic vibrations caused by sound waves in the room. Then, they use AI to reconstruct those vibrations into crystal-clear speech. Through walls. From adjacent rooms. From up to 50 meters away. It was tested on actually deployed infrastructure. The attack cost is dropping. Commercial gear is all that is required if an attacker has access to the other end of the fiber connection. Millions of homes and offices have FTTH installed. And every single one is potentially exposed.

🚨 Running Google's Gemma 4 on my iPhone... without internet No data plan. No cloud. No monthly fee. Gemma 4 runs completely offline, handles 128K context, and fits in my pocket. Here's how I set it up in under 1 minute:

@Samaytwt Supabase

As a developer, which authentication provider do you prefer?

@vandanasoc28779 X and personal blog

Which platform is best to write your content? 1.) Medium 2.) X 3.) Personal Blog 4.) Others

@Still_learner Lets go

I'm blowing small accounts If you're under 41K reply and I’ll boost you 🚀

@ShrishtiSNagar @hridoyreh Instagram is for brain rots

@hridoyreh Instagram??

Places to get your first 100 users: 1. Reddit 2. Indie Hackers 3. Hacker News 4. Quora 5. Product Hunt 6. BetaList 7. AlternativeTo 8. SaaSHub 9. Side Project Ideas 10. Launching Next 11. StartupBase 12. Uneed 13. Facebook Groups 14. X (Twitter) 15. LinkedIn 16. TikTok 17. Medium

@creepydotorg What i read just now🙃

A 46-year-old woman accidentally snorted 55 mg of pure LSD because she thought it was cocaine, roughly 550 times the normal recreational dose. She vomited for 12 hours and remained under the effects of the substance for about 34 hours. But when she woke up, the chronic foot pain she had suffered for over 20 years, attributed to Lyme disease, was reportedly gone. She stopped taking morphine for five days and then began microdosing LSD every three days. Three years later, she had completely quit morphine, with no withdrawal symptoms or relapse reported. Researchers described it as an extraordinary case of non-fatal high-dose LSD ingestion with long-term pain relief.

@Mistonchain Hey hi

Follow 4 Follow If you reply, I’m following you

@sflorimm 🤣

Not sure how i feel about this.

@deredleritt3r 😀

Anthropic revenue (annualized): - January 2025: $1B - May: $3B - June: $4B - August: $5B - October: $7B - December: $8B to $10B -February 2026: $14B -March 2026: $19B -April 2026: $30B (WTF???)