Breaking things is fun. Code is law! I no longer own a Dinosaur. NOT an Architect. Fundamentals are dead.
Gary, IN
Joined April 2011
- Tweets99.4K
- Following725
- Followers1.8K
- Likes285K
Theseus 😈 retweeted
im sorry bro this nigga is a mad genius its literally insane
🚨 BREAKING: SpaceX locked Anthropic into paying them $3.1 billion per MONTH for compute starting December
$37B a year
= $120.5 BILLION through 2029
Every month, ~80% of Anthropic's current revenue would be SpaceXSI compute bill
BIGGER than its Google and Amazon long-term deals
Anthropic warns short-term compute "carries a higher cost per unit" and will "compress our operating margin"
He has the infinite power of the internet in his hands to tell him how to do it, and still complains.
Non-construction bros are not gonna make it.
I'll have job security for a long time!
AI is NOT eliminating white collar and creative jobs anytime soon!
If you believe this, you either:
- Only working in tech
- Don’t work with AI enough
Go ahead and try to implement an agentic workflow in an SMB. Tell me how that goes.
What continues to be painfully obvious is the need for a “human in the loop”
Here’s an example:
We are rebuilding our home after a fire and using AI to help visualize each room before buying things like:
• tile
• flooring
• appliances
• furniture
But it takes DOZENS of iterations, doesn’t remember instructions and straight up hallucinates.
This is my exact process:
1. Upload floor plans and tell ChatGPT what I am trying to visualize
2. Ask it to read the documents and give me a plan for generating the image
3. Review the plan and approve/give feedback
4. Generate image
5. Give specific feedback, point back to the context docs and ask for a plan to fix (I will often ask it to audit its own fidelity to the plans)
6. Review the plan and approve/give feedback
7. Repeat steps 5 and 6 until I get it right
Why do I have to go through so many rounds??
Because it LITERALLY ignored explicit instructions.
You think you’re going to get mass market, enterprise-wide, corporate adoption like this???
No.
Now, you might say that I just don’t know what I’m doing and there is a simple fix to that (or something much meaner)
But that’s my point.
If someone like me who uses AI every day still has difficulty getting consistent and predictable results, why do we think there would suddenly be mass adoption?
We are still INCREDIBLY early.
