Added pi to the list. Unironically considering it as the base for AI setup. https://ruby.social/@maxim/116037400645991471
| Website | https://max.engineer |
| http://x.com/hakunin |
- 42 Followers
- 46 Following
- 244 Posts
Specialize in software design and DX. VPE at Helios (FinTech). Site: https://max.engineer. X: @hakunin. Opinions mine.
PSA for those caught in the AI hamster wheel, something I wrote in my college notebook 22 years ago, still holds up:
Happiness is a cause, not an effect.
Get happy now, in your mind, then do things. You don't work towards it, you start with it, and it enables everything else.
Here's where the puck is going: AI will replace software. You ask it to solve problems, you instantly get custom apps. All computing will become this. We will be sharing T=0 prompts that produce useful results. All software engineering will become about improving AI itself.
My attempt at deciding on an AI orchestrator. So many out there. Suggestions/corrections welcome. https://gist.github.com/maxim/c6ef7f7abd8e33ca846bb99aa9139602
@nerius @colincornaby that thread definitely affirms your suspicion.
@colincornaby Not sure if this is what you're referring to, but I'm having that exact conversation in this thread: https://news.ycombinator.com/item?id=46924327 (I'm hakunin), and it's more about non-determinism for me. The "database" mental model kinda fits, if you also include the way "records stored" relate to one another. So when you query the db, it picks among millions of records and millions of ties between those records to create something that looks new. Plus, you can add data and ties in the query itself.
If you're gonna embrace non-determinism at all stages of software dev, including verification, why even make software? Just have AI be the software. Why have an "if" condition when AI can decide. Unironically, that's where we're going. It's moving ground all the way down.
@colincornaby I hope so. It seems like we’ll split into 1) a sea of AI-only devs who can’t operate without it, but are still considered software creators due to being generally “tech” and getting things done cheap, and 2) a rare breed of niche code-literate devs who can still leverage AIs for speed, but also ensure quality/maintainability. I hope there will be a recognition of value for latter.
@colincornaby I agree in a sense that almost everything is a time/money issue. On top of that, I've been noticing a collective shift away from caring about edge cases, to the point of only one happy path with defaults actually mattering. This applies to both new software, and real world scenarios like "what happens if LLM provider is down and I need to fix a critical bug". Few people seem to really care anymore. Hope the pendulum will eventually swing back into "responsible" territory.