> We're not quite at the level (yet) of feeding a whole codebase to an LLM and making it add features or make changes while understanding the big picture of the problem being solved
We are one step away from that. All we need is a more advanced form of fine tuning.
eSports are going through a slump right now. Psyonix also cancelled upcoming RLCS. Instead it's going to sponsor a bunch of smaller tournaments like salt mine with 10-20% of the budget of last year's RLCS.
> When we tell AI that we're relying heavily on its answers, it "doubles down" to provide us with more precise, thoughtful, and thorough responses. The AI isn't actually feeling the pressure
this cowwection of thoughts on softwawe devewopment gathewed by gwug bwain devewopew
gwug bwain devewopew not so smawt, but gwug bwain devewopew pwogwam many wong yeaw and weawn some things awthough mostwy stiww confused
gwug bwain devewopew twy cowwect weawns into smaww, easiwy digestibwe and funny page, not onwy fow you, the young gwug, but awso fow him because as gwug bwain devewopew get owdew he fowget impowtant things, wike what had fow bweakfast or if put pants on
big bwained devewopews awe many, and some not expected to wike this, make souw face
THINK they awe big bwained devewopews many, many mowe, and mowe even definitewy pwobabwy maybe not wike this, many souw face (such is intewnet)
```