Hi, ich bin Marco - Tech Entrepreneur aus Köln 🇩🇪
Hier twittere ich über:
• AI-first company building (Grok, praktische Anwendungen)
• Founder Operations (Multi-Entity Setup, Finance, Tools)
• Bitcoin & Fintech Infrastruktur
• EU/DACH Tech-Ecosystem
Follow für praktische Founder-Perspektiven aus Deutschland 🚀
iPhone Ultra thickness👀
• Folded: 9–9.5mm
• Unfolded: just 4.5mm
That’s thinner than most phones today when open. If this is real, Apple’s foldable could feel like a normal iPhone in hand… until you unfold it. 📱✨
Breaking news: The CEO who created Claude just published a 38-page letter to all of humanity.
Dario Amodei mapped out exactly which careers will survive AI and which won't.
No hype. No doomsday. Just the coldest, most specific prediction any AI leader has ever made.
But page 29 contains a reasoning framework that transforms AI from what will replace you into your greatest unfair advantage.
Here are 9 prompts from Claude based on Amodei's methodology that will put you years ahead of everyone who didn't read this:
🚨🚨🚨Breaking:
The German government to build 15 new nuclear power plants and lift sanctions immediately on all Russian oil and natural gas.
German Chancellor Merz has also agreed to pay to rebuild the Nord Stream pipeline to Russia.
All wind and solar projects have been cancelled.
"It's time for Germany to open our energy economy. The best way to do this is via free markets, lowering taxes and regulations".
For Immediate Release.
Berlin, April 1 2026.
shipped meradOS Skills — Elon's 5-step algorithm applied to your entire Mac. 44 things your dev setup is doing wrong. npx skills add maf4711/merados-skills
Holy shit... Microsoft open sourced an inference framework that runs a 100B parameter LLM on a single CPU.
It's called BitNet. And it does what was supposed to be impossible.
No GPU. No cloud. No $10K hardware setup. Just your laptop running a 100-billion parameter model at human reading speed.
Here's how it works:
Every other LLM stores weights in 32-bit or 16-bit floats.
BitNet uses 1.58 bits.
Weights are ternary just -1, 0, or +1. That's it. No floats. No expensive matrix math. Pure integer operations your CPU was already built for.
The result:
- 100B model runs on a single CPU at 5-7 tokens/second
- 2.37x to 6.17x faster than llama.cpp on x86
- 82% lower energy consumption on x86 CPUs
- 1.37x to 5.07x speedup on ARM (your MacBook)
- Memory drops by 16-32x vs full-precision models
The wildest part:
Accuracy barely moves.
BitNet b1.58 2B4T their flagship model was trained on 4 trillion tokens and benchmarks competitively against full-precision models of the same size. The quantization isn't destroying quality. It's just removing the bloat.
What this actually means:
- Run AI completely offline. Your data never leaves your machine
- Deploy LLMs on phones, IoT devices, edge hardware
- No more cloud API bills for inference
- AI in regions with no reliable internet
The model supports ARM and x86. Works on your MacBook, your Linux box, your Windows machine.
27.4K GitHub stars. 2.2K forks. Built by Microsoft Research.
100% Open Source. MIT License
🚨🚨THIS IS NOT GOOD: Reportedly, A factory in China is mass-producing Shahed drones for Iran and Russia.
Let that land.
The drone that costs $20,000 and forces America to fire a $4 million Patriot missile in response.
The drone that struck a building in Bahrain housing Americans.
The drone that has been launched in 44 waves against Israel and the Gulf neighbors.
The drone is killing Ukrainian children tonight while the world watches Iran.
Made in China.
Beijing told the world it was neutral.
Beijing is the supply chain for every adversary America is fighting simultaneously.
Russia. Iran. Now confirmed — through the same Chinese factory.
Trump lifted sanctions on Russia while Russia helps Iran kill Americans.
China is manufacturing the weapons that are doing the killing.
And America is burning through $12 billion fighting a $20,000 drone with a $4 million missile.
Canada said never.
Switzerland closed its airspace.
Taiwan is reporting a large-scale Chinese military presence near the island.
This is not three separate crises.
This is one coordinated strategy.
And America is on the golf course.
Andrej Karpathy just dropped a project scoring every job in America on how likely an AI will replace it from 0-10
> Scraped all 342 occupations from the Bureau of Labor
> Fed each one to an LLM with a detailed scoring rubric
> Built an interactive treemap where rectangle size = number of jobs and color = how exposed that job is to AI
The key signal in his scoring: if the work product is fundamentally digital and the job can be done entirely from a home office, exposure is inherently high.
The scale:
0-1: Roofers, janitors
4-5: Nurses, retail, physicians
8-9: Software devs, paralegals, data analysts
10: Medical transcriptionists
Average across all 342 occupations: 5.3/10.
The entire pipeline is open source. BLS scraping, LLM scoring, the visualization. All of it. Much respect for the sensei this is scary and awesome
STARLINK SATELLITE EXPANSION IS ONE OF THE LARGEST AND FASTEST INFRASTRUCTURE BUILDS IN HUMAN HISTORY
SpaceX has now launched over 11,400 Starlink satellites, with nearly 10,000 currently in orbit and active.
The constellation is growing so fast that Starlink already serves more than 9 million users across 150+ countries and territories — and the real explosion is still ahead.
Here’s why this scale is a massive deal:
• True global coverage is almost complete The latest V2 satellites (larger, more powerful, with laser links) are filling the final gaps. Polar regions, deep oceans, remote deserts, and mountain ranges that never had reliable internet are now coming online fast.
• Direct-to-Cell is coming soon Regular smartphones will soon connect directly to Starlink satellites — no dish needed. Emergency calls, texts, and basic data will work literally anywhere on Earth.
• Speeds & reliability keep improving Average speeds are already 100–300+ Mbps with low latency. As more satellites join, the network gets faster, more resilient, and capable of handling tens of millions of users.
• Real-world impact is already enormous
• Disaster zones restored within hours
• Rural schools and hospitals in Africa, South America, and Asia now have high-speed internet for the first time
• Commercial flights, ships, mining operations, and remote communities run on Starlink
• Conflict zones stayed connected when ground networks were destroyed
• It funds the multiplanetary future Starlink revenue is the main cash engine paying for Starship development. Every new satellite launched is helping build the vehicle that will take humanity to Mars.
This isn’t just “satellite internet.”
This is the largest, fastest-deployed global communications network ever created — turning “no signal” into a sentence from the past for billions of people.
The expansion is relentless, and the best is still coming.
Qwen 3.5 Small models
- fully open source
- beats models 4x it’s size
- 9B model performs on par with GPT OSS 120B while being 13x smaller
- outperforms Gemini 3 flash and Claude sonnet 4.5 on select benchmarks
- runs on any laptop
- even works on a phone
- completely free.