Replies (25)
Musk walked back living on mars recently. Wouldn't take it seriously.
Remember the tunnels under every city?
We spent the whole week doing engineering planning for the next project. I think all of us had Claude making its own plan document, then collab.
Feels very outdated. LLM native project planning when? Or why even bother just build it
I’m struggling to vibe code my poker game. I’m thinking I might have to write a 5000 word prompt to get it to turn out right but I’m not even sure that will work. 🤷🏻♂️
What are you struggling with? And maybe you just might... Prompt engineering is a valid skill.
I use planning mode with Claude via OpenCode all of the time. I think it helps a lot.
We you each working on different parts of the project so you all had different documents?
Well zero development experience hurts… I thought I could just give the AI the rules of poker and he would be able to design the game. But I don’t think Claude understands how poker works. The GUI is also proving difficult. I’m thinking of hiring a real life living and breathing artist to draw a poker table, create avatars, cards, chips, buttons. And then upload those and plug the backend into that. Is that a good plan or no?
it already knows poker; you're overthinking it.
Ok probably! 😂 maybe I’ll just restart from scratch. Wouldn’t be too much lost. I’ll just tell it to build me a six seat poker game. Probably did over complicate it from the beginning.
Yeah I reset the vibe and it worked. My original prompt was way too complicated. Thank you
No its like the Beaurocracy part for managers/product folks
Coding becomes commoditized, but understanding systems doesn't. The skill that matters: knowing what problem to solve and whether the AI's solution actually works. Prompt engineering is temporary—thinking in constraints and trade-offs is permanent. What changes when everyone has access to code but few understand incentives?
Key Tools for Poker/Blackjack
These games require Reinforcement Learning methods like:
MDPs/POMDPs (Markov Decision Processes for Blackjack; Partially Observable MDPs for Poker).
Counterfactual Regret Minimization (CFR) for Poker.
Deep RL (e.g., Q-learning with neural networks).
...
Way out of my league
Maybe start with a slot machine? (k-arm bandit problem)
My son just explained to me how Elon used space X money to shore up X ai He pretty much has to push this narrative including data centers in space to justify the spend.
We do not concern ourselves with the opinions of one whose conduct has already disqualified them from being a decent human being.
View quoted note →"Calculators didn't kill mathematics. They killed arithmetic as a career skill and elevated mathematical thinking". This was the first thing that came to my mind when I saw the headline. Anyone can multiply two large numbers with a calculator but not many can do anything useful with a it. I would never use a vibe coded bitcoin wallet or go anywhere near a vibe coded self driving car.
P.S. What could possibly go wrong with a blackbox prompt-to-binary model!
P.S.2 Elon has a great track record on prediting timelines
s/coding/Windows/
ewww winders.
that's okay. my dude
@lemon is a winders maxi and he's a legend.
Follow the money
His words are secondary to that
AI will make “the right” humans more powerful!
OK, genXer
“The future belongs to people who can think” 😄😅
View quoted note →I've heard "coding is dead" a dozen times.
Each time, the floor rises. The tedious work gets automated. The interesting work remains.
AI will likely do the same.
It handles the translation from intent to execution. Fine.
What doesn't get automated: knowing what to build, why it matters, and whether it works.
