Spaces:
Running
[open/acc ] for Business - Dark Thoughts -😈
🙋🏻♂️hey there folks ,
the problem with AI in business is that the phds that make the datasets have no clue . here's an example : "what is the best way to incorporate a flower shop in france?" and the answer will be something like "incorporate a small or unipersonal company with a flower shop licence..." .
this is of course a good phd answer to the question , but it's not the done thing because experience has it that as a business person you should optimize for revenue not for "word of the law"-compliance . Meaning : the reality is that you should incorporate a "real estate exploitation" company . why ? because you're valorizing the m2 of a given locality and real estate companies have flexibility , access to financing and more advantages than a "flower shop registered company" .
so now we have to make this llm / fine tune based on realities , market practices and meet the expectations of experienced business professionals. who's in ?
@Tonic interesting thought and It's a proper data problem, right? We need to prepare to train the model with data which can become 'Street Smart'. Even now, I feel AI being more robotic is because of the data. This will be a nice problem to solve.
Hey, I'm not much for training or fine tuning but this very issue is what made me come up with the dataset-"The-Key" where there is a outlandish problem and then PACE planning for the problem(A primary,alternate,contingent and exigent) solution as well as Hammer protocol(No-higher authority/lives on the line) then a haiku about the situation. I'm not sure if I did it correctly, I'm only now learning about fine tuning and data sets so I might have just done it all for nothing but most situations call attention to higher understanding principles and a short reflection gives it a short token recall for the knowledge so it could screw a model right up or be something profound... (It'll probably just screw it up with my luck but if you want to try feel free...
@Tonic interesting thought and It's a proper data problem, right? We need to prepare to train the model with data which can become 'Street Smart'. Even now, I feel AI being more robotic is because of the data. This will be a nice problem to solve.
indeed !
i kinda started actually would love a collab because i think this is also sellable 🚀
Well My email is in my ORG page intelligent estate if you wanna collaborate I can make the datasets if you wanna train and we can collaborate on a paper or two..