Rejoice, if Qwen doesn't release any new local model, it's a blessing in disguise

Do you remember the times when we only had lama2 released? a bunch of finetunes were released and some of them had real values . MoE are harder to finetune, but it's probably because the effort to try and find the solutions aren't woth it. by the time it takes someone or some group to finetune a model 4 or 5 newer and better ones are already out there , it's not worth the effort. I believe that once the open weights release slow down, fine tunes will begin rolling out again. and we have so much knowledg accessible (web and llms) that it is easier for more people to try new things. Maybe with time we will even find some p2p way to train true opensource model. People with skills will pave a new way, and others will lend their gpu capacity to finally break through the slavery of the big corporation. because yes we are slave to qwen, gemini and deepseek, once they stop releasing, we are stuck. So yes let's pray that Qwen stop releasing their models, and I'm sure people will start to find solutions 😄