Manuel (@mindscrabble) header
Manuel (@mindscrabble) profile picture

CPO & Co-Founder @Susplustech | AI | Blockchain | Innovation | Entrepreneurship | Gaming

Joined August 2014
Photo by Manuel (@mindscrabble)Photo by Manuel (@mindscrabble)Photo by Manuel (@mindscrabble)Photo by Manuel (@mindscrabble)Photo by Manuel (@mindscrabble)Photo by Manuel (@mindscrabble)Photo by Manuel (@mindscrabble)Photo by Manuel (@mindscrabble)Photo by Manuel (@mindscrabble)Photo by Manuel (@mindscrabble)
Training compute for frontier models is doubling every 6 months. The headlines focus on GPT-5 or Gemini 3, but the real war is for power and chips. Datacenters are now measured in gigawatts, not square meters.
2
67
Whoever secures energy and silicon first will dictate the pace of intelligence itself. The infrastructure will decide the winner and will power breakthroughs across every industry.
1
6
1
46
10,000 home robots generate more training data than YouTube’s 20M daily uploads. Factory robots plateau after 40 hrs. Home robots never stop: every coffee spill, door handle, and routine is a lesson. Your living room is about to become the classroom for superintelligence.
1
32
GPT-5 is the most powerful model yet. So why does it feel… modest? The answer lies in 5 forces that shape how we perceive progress. 🧵
1
1
22
V. Gaps As systems improve, the missing pieces stand out: – no persistent memory – weak temporal reasoning – brittle adaptation Each breakthrough makes the gaps feel bigger.
1
1
2
98
GPT-5 feels modest because revolutions don’t announce themselves. Railroads, electricity, the internet all seemed incremental…until they redefined the world. The lesson: Modesty is the moment when a breakthrough turns into everyday reality.
2
11
Settlement delays? Gone. Mortgage paperwork? Gone. Brokers clipping fees on every step? Gone. Friction doesn’t survive in a programmable world.
2
1
14
When information moved online, it blindsided incumbents. Tokenization will do the same to finance. This time the stakes are trillions.
1
12
The internet finally gets its financial layer. And once it’s live, markets will look less like Wall Street and more like software. The question isn’t if. It’s who adapts fast enough to survive.
1
1
1
48
In the 90s we digitized information. That gave us Google, Facebook, and email. It also killed newspapers and rewrote the music industry. Now we’re about to digitize value itself. And the impact will be 10x bigger.
1
11
The changes are already underway: - Dubai is fractionalizing real estate - Robinhood & Kraken testing tokenized equities abroad - Coinbase lobbying to do it here in the US The dam is cracking.
1
1
37
Example: You own 1% of a Manhattan apartment. That token can: - Stream you 1% of rent every month - Serve as collateral for a loan - Unlock one night per year to actually stay there Ownership becomes programmable.
2
9
Once assets live on-chain, they stop behaving like assets and start behaving like software. That’s the shift no one is ready for.
1
1
35
Look at the real numbers: - $120T real estate - $100T equities - $13T treasuries - $12T gold Every asset class is on deck. Not billions. Trillions.
1
1
15
The White House just released a 168-page crypto strategy. Buried inside: $600B in assets tokenized by 2030. That’s laughably small.
1
10
AI isn’t peaking, it’s only beginning. Too many treat it as another tech wave. In reality, it’s the foundation of a new era: reshaping economies, accelerating science, and forcing us to rebuild infrastructure from the ground up. Here’s why the biggest opportunities are ahead 🧵
1
11
5/ The bold will outrun the cautious Most corporations are too risk-averse to go all in. Startups that are AI-first will define the next decade while incumbents slow-walk into irrelevance.
1
6
What feels advanced today will look primitive in five years. The question isn’t if AI transforms society, but whether you’ll be part of the wave or left behind. 🌊
7