This might be one of the most underrated releases of 2026
NVIDIA just dropped LocateAnything ➜ and it’s stupidly good
Give it any image and a text prompt, and it instantly highlights the exact spot of whatever you asked for. Not roughly. Precisely
> 10x speedup
> 138 million training samples
> Multiple inference modes for speed vs accuracy
Most AI today is still blind to layout and position. This model fixes that
Think about what becomes possible when your AI can reliably point to "the submit button", "the tumor in the scan", or "the error in the log file" in real time
This is the kind of quiet but powerful tech that will power the next generation of actually useful visual AI applications
NVIDIA is building the foundation for AI that doesn’t just talk ➜ it sees and understands space
A 21-year-old Chinese kid just built a monster farm of Mac Minis
> Full server racks
> Dozens of M4 Macs stacked like bricks
> Running 24/7 at full power
While you're bleeding $200–$500/month on Claude/GPT subscriptions…
This dude owns his compute completely. No more API bills. No rate limits. No censorship. Just raw local power for pennies in electricity
The quiet rebellion is happening right now:
Heavy users are canceling expensive subscriptions and switching to local setups ➜ Mac Mini clusters, small server racks, private inference
One-time hardware cost > endless monthly rent
This is the real "f*ck your subscription" movement in 2026
The ones who own their hardware will move faster, keep their data private, and never get throttled again
This might be one of the most underrated releases of 2026
NVIDIA just dropped LocateAnything ➜ and it’s stupidly good
Give it any image and a text prompt, and it instantly highlights the exact spot of whatever you asked for. Not roughly. Precisely
> 10x speedup
> 138 million training samples
> Multiple inference modes for speed vs accuracy
Most AI today is still blind to layout and position. This model fixes that
Think about what becomes possible when your AI can reliably point to "the submit button", "the tumor in the scan", or "the error in the log file" in real time
This is the kind of quiet but powerful tech that will power the next generation of actually useful visual AI applications
NVIDIA is building the foundation for AI that doesn’t just talk ➜ it sees and understands space
«YOUR COMPANY IS NOT GOING TO GO OUT OF BUSINESS BECAUSE OF AI. YOUR COMPANY IS GOING TO GO OUT OF BUSINESS BECAUSE ANOTHER COMPANY USED AI»
— Jensen Huang
Today at Computex, @nvidia made it very clear:
> The era of Local AI has officially begun
They announced RTX Spark
- a powerful new chip for thin laptops with up to 128GB unified memory
Combined with their new Vera CPU, it’s designed specifically for agentic AI workloads ➜ It delivers an 80% speedup over x86 on agent-based tasks. . .
> You can now run strong 70B–200B models locally
> No more paying $1k–2k/month on cloud APIs (rapid ROI within a couple of months of purchase)
> No rate limits and no data leaving your machine
> Persistent agents and multi-agent systems become practical and cheap
Local AI gives you three big advantages:
- much lower costs
- full privacy
- and the ability to run agents 24/7
Cloud was convenient
Local is the new competitive edge
The shift is happening now. Those who build their AI stack locally will have a real advantage in speed and cost over the next 12–24 months
My bet:
The future belongs to those who own their compute
Full podcast below ⇩ ⇩ ⇩