Mostly extemporaneous as the code is generated. But for work stuff I spend some time looking at the code and checking edge cases. (But that’s no different for me from pre-AI-gen.) I’m experimenting with using sub agents to code review as well.
Learn my AI-augmented feature crank technique in less than three minutes.
Adding shooting to my video game in 8 minutes.
Talk → generate → scan → commit
Using this feature crank technique, I can fluidly bang out features. I estimate I'm 3x faster than I was with trad programming.
I coach pro engineers. $100 per 90m session. DM me to book!
youtu.be/1o6x0zxItyg
Stream of me using AI to build a 2D shooter with FEAR-style enemy intelligence. Link in thread.
* Writing prompts with voice-to-text. Way faster than typing.
* Staying in flow by using Agent mode.
* Planning and problem-solving with the LLM.
* Simple but extensible architecture.
Thanks for sending the portfolio over. Are there any aspects in particular that you'd like feedback on? (I don't know of any internships right now, but will keep an ear out!)
Is any of your work visible that I can take a look at?
Ah cool! Do you have a portfolio/linkedin/resume?
* SpyParty
* Syndicate
* Katana Zero
Ooo great thank you!
