Max
1K posts


IF DUBAI REAL ESTATE IS CRASHING...
SOMEBODY FORGOT TO TELL @emaardubai .
A $55 billion project just got announced.
Oh...
And buying a property can now help you secure residency in the UAE.
Dubai Update 🇦🇪
English

Max रीट्वीट किया

The biggest shift AI has created for me is collaboration, not automation.
The mental model that works best isn’t “delegate work to the machine.” It’s “collaborate with a junior developer.” I often know when an answer is wrong long before I know the right answer, and that feedback loop is where the value comes from.
AI performs best in deterministic environments where success can be clearly measured. Cryptography, state transitions, and defined outputs create systems where you can iterate until the result matches expectations. If you don’t understand what success looks like, you can’t effectively guide or correct the model.
That is why I tell people not to use AI for things they don’t already understand. Build the hello-world version first. Learn the architecture. Understand the tradeoffs. Human judgment still matters when deciding what to build. AI is increasingly powerful at helping execute it.
Model selection matters too. Different models have different strengths. I tend to use codex for backend engineering and deterministic systems because they are fast and consistent. For UI and design work, I often prefer models that are more creative like claude. The best results come from matching the model to the problem.
The workflow matters as much as the model. AI gets isolated development environments, git worktrees, and production-like test clusters. Humans stay in the loop for commits and pull requests. AI can experiment freely, but deployment decisions still require explicit review.
One underrated use case is learning. Instead of acting as an AI operator, become an AI student. Use voice mode. Ask questions. Challenge assumptions. Let the model explain architectures and alternatives. Some of the biggest gains come from accelerating understanding, not just generating code.
Looking ahead, I suspect prompt engineering becomes less important over the next 18 to 24 months. Models are rapidly improving at understanding abstract intent. The bottleneck will not be writing clever prompts. It will be identifying worthwhile problems.
That is where humans still have the advantage. AI is exceptional at execution. Humans remain better at identifying opportunities, framing problems, exercising taste, and creating things people actually care about.
The future is not humans versus AI. It is humans who know how to collaborate with AI and those who do not.
English

Bitcoin Bob Day is Coming.
Bitcoin Bob USDC@RealBitcoinBob
The video that started it all. 21 November 2010. "What is a Bitcoin?" #bitcoinbob
Sale, Victoria 🇦🇺 English

Day 3 of $bob being criminally undervalued
See you at BOBILLIONS
7DWRAaAecR2bhPRUjg7VN6XqoXpRYCRqsq7W6t2Vpump

TEDDDY@tedddywins
Day 2 $bob still larping under 100k Holding until bobillions 7DWRAaAecR2bhPRUjg7VN6XqoXpRYCRqsq7W6t2Vpump
English


















