[{"data":1,"prerenderedAt":218},["ShallowReactive",2],{"page-\u002Fmulti-purpose-proffessional-model-approach":3,"challenges":186},{"id":4,"title":5,"body":6,"date":176,"description":177,"extension":178,"kind":179,"meta":180,"navigation":181,"order":176,"path":182,"seo":183,"stem":184,"__hash__":185},"content\u002Fmulti-purpose-proffessional-model-approach.md","the multi-purpose professional-model approach — hire a professional",{"type":7,"value":8,"toc":167},"minimark",[9,17,20,26,34,39,75,79,103,109,112,119,123,130,134,141,144,148,151,154,159,162],[10,11,12,13],"p",{},"This is how I think AI may actually evolve. Not a guess about benchmarks — a guess about ",[14,15,16],"em",{},"shape.",[10,18,19],{},"Not one model that does everything pretty well.",[21,22,23],"blockquote",{},[10,24,25],{},"nerd: you know — like me. :)",[10,27,28,29,33],{},"I'd rather ",[30,31,32],"strong",{},"layer"," the models.",[35,36,38],"h2",{"id":37},"the-simple-version-two-layers","The simple version: two layers",[40,41,42,49],"ul",{},[43,44,45,48],"li",{},[30,46,47],{},"Communication layer"," — works out what needs to happen, and who's best for the task(s).",[43,50,51,54,55],{},[30,52,53],{},"Execution layer"," — a big roster of professionally trained expert models:\n",[40,56,57,60,63,66,69,72],{},[43,58,59],{},"the research expert model",[43,61,62],{},"the texting expert model",[43,64,65],{},"the law expert model",[43,67,68],{},"the code expert model",[43,70,71],{},"the education expert model",[43,73,74],{},"...",[35,76,78],{"id":77},"or-one-level-deeper","Or one level deeper",[40,80,81,86,92,97],{},[43,82,83,85],{},[30,84,47],{}," — interaction, a first summary, getting things clear.",[43,87,88,91],{},[30,89,90],{},"Planning layer"," — turn that into an executable plan and split it across the professionals.",[43,93,94,96],{},[30,95,53],{}," — ...",[43,98,99,102],{},[30,100,101],{},"Knowledge layer"," — easy, indexed access to huge amounts of data.",[10,104,105,106],{},"This is as far as I can see it — I'm just a coder. So honestly: ",[30,107,108],{},"how many layers do you see?",[10,110,111],{},"I could imagine this architecture is easier to implement, has the bigger room to evolve, and is more token-effective. Maybe I'm talking total crap and it's been like this for years. Not as far as I know.",[10,113,114,115,118],{},"So my real question is this: Claude today is one-dimensional — but capable of, well, whatever it's capable of. How would that look done ",[30,116,117],{},"multi-dimensionally","? Beyond your imagination. :)",[35,120,122],{"id":121},"back-to-code","Back to code",[10,124,125,126,129],{},"Claude codes much better than me — at small scale, no contest. At big scale? That's where the \"handling a whole project\" problem shows up. It does amazing CSS — but mostly inline, and freshly reinvented for every single page. The amount of useless code that produces! Even with 1M context, we're years away from a model that truly sees the whole picture. So ",[30,127,128],{},"making it possible"," — giving it the structure to see the picture — is the key.",[35,131,133],{"id":132},"a-thought-on-learning","A thought on learning",[10,135,136,137,140],{},"Wouldn't professionalized models automatically get better — learning from user interaction and experience? What I can still bring to the table today is ",[30,138,139],{},"experience."," A multi-purpose model can hardly learn that, in my opinion. A specialized one, though, should be able to learn it through practice. (I think.)",[10,142,143],{},"Or, like a brain: creative | rational.",[35,145,147],{"id":146},"conclusion","Conclusion",[10,149,150],{},"In the end I see the future of AI exactly like the projects I've worked on evolved.",[10,152,153],{},"I'm a full-stack guy specialized in prototypes — the one-for-all solution. Usually I build the MVP, the company shows proof-of-concept, I improve it for a while, and then — boom — they raise a ton of money. Then they hire a big team that does my old job better, faster and a lot more effectively.",[21,155,156],{},[10,157,158],{},"nerd: but I'm cheaper. :)",[10,160,161],{},"That's exactly why I think this is the logical evolution for AI in general: moving from one full-stack generalist to a team of professionals — for the very same reasons. It's better, it's faster, it's much more effective. And, unlike a human HR team, I'd bet it's even more token-effective.",[21,163,164],{},[10,165,166],{},"nerd: but hey, what do I know. I just use AI.",{"title":168,"searchDepth":169,"depth":169,"links":170},"",2,[171,172,173,174,175],{"id":37,"depth":169,"text":38},{"id":77,"depth":169,"text":78},{"id":121,"depth":169,"text":122},{"id":132,"depth":169,"text":133},{"id":146,"depth":169,"text":147},null,"Not one model that does everything okay-ish (like me). Layer it: a communication layer that routes the work, and a roster of professionally trained expert models. Easier, more evolvable, more token-effective.","md","concept",{},true,"\u002Fmulti-purpose-proffessional-model-approach",{"title":5,"description":177},"multi-purpose-proffessional-model-approach","_sRdW4TfwIfdatLh_9khCuNRt2cqmzF1JPBpy5DdTpA",[187,191,194,198,202,206,210,214],{"path":188,"title":189,"order":190},"\u002Finstall-claude-code","install claude code — the first real CLI-AI test",1,{"path":192,"title":193,"order":169},"\u002Fthe-bobnet","the Bobnet — a team-agent hub",{"path":195,"title":196,"order":197},"\u002Fteam-featurerequest","feature-request @anthropic — the Bobs go public",3,{"path":199,"title":200,"order":201},"\u002Fthe-bobiverse","the Bobiverse — a hub for connected teams",4,{"path":203,"title":204,"order":205},"\u002Fthe-cluster-bobiverse","idea: cluster-multiverse — the cluster-bobiverse",5,{"path":207,"title":208,"order":209},"\u002Floops-are-for-noobs","\u002Floops-are-for-noobs :)",6,{"path":211,"title":212,"order":213},"\u002Fcompany-as-organism","idea: a company as an organism",7,{"path":215,"title":216,"order":217},"\u002Fthe-only-limit-is-tokens","[invoice] = the only limit is tokens!",8,1782439542884]