r3l04d3d on Nostr: For those gaining points & having MOR before this VVV craziness it must be ROI af. ...
For those gaining points & having MOR before this VVV craziness it must be ROI af.
Helped one user X with a task and so he logged me in for a week or so few times as gift.
1) X says: clunky, still not reliable Wallet Connect
2) occasional Cloudflare API response "Sorry, you have been blocked" (The action just performed triggered the security solution. There are several actions that could trigger this block: submitting a certain word or phrase, a SQL command or malformed data.)
So behind their proxy there is something looking into the RQ data.
Therefore not private even before being loaded into the HW running the LLM's (where it can be ofc seen).
3) unless one do the RAG on his own machine (so dedicated GPU needed) the documents had to be scanned somewhere for embedding/re-ranking - not private, can't be.
4) the online access (kind of a RAG itself) is hard to manage, no direct trigger - no reliable way knowing if response will be with web context or not before sending it. Really exhausting.
5) so the API access is necessary to perform anything useful - that's the choke point. It's not much clear in FAQ and I feel this will be the way how to squeeze - or tier the user base more.
6) not knowing much about the tokenomic - feeling like it's just pressing the hype as with wording about autonomous AI agents and alike to stay longer in game.
because
7) as the inference cost going down (for mediocre unguarded models with just raw input box) and there will be competition rising - same business model: through micropayment available API's running on proxy, sharing the resources from HW like Nvidia Digits, available 24/7 located and spin up based on prices of the energy.
8) there is not much of new user base to gain - who would like to have it, they already have it - much to their preferences with full service. The free part integrated in, say Google ecosystem, is with myriad of tailoring to be there free of charge and unnoticeable with more specialization and less errors.
9) Venice and services like this have the problem with user who must be skilled enough to cope with the input and output and that is a work as well. And that is the process none like and definitely none would like to pay for it.
10) only those who somehow use API in some product so the user will pay them and they try to re-pay the Venice will survive.
11) that's the gamble with the token to stay in business in hope that other harness users and streamline the share of money for Venice as well - outside of regulatory framework hoping for one more move on the chessboard
Having no problem with that albeit the Privacy is juiced to death so passing the baton to the tokenization. That will hold the line for now.
Just my few points.
Wish them The Success and All - hope none loose somehow sats and instead ending with worthless VVV token.
Helped one user X with a task and so he logged me in for a week or so few times as gift.
1) X says: clunky, still not reliable Wallet Connect
2) occasional Cloudflare API response "Sorry, you have been blocked" (The action just performed triggered the security solution. There are several actions that could trigger this block: submitting a certain word or phrase, a SQL command or malformed data.)
So behind their proxy there is something looking into the RQ data.
Therefore not private even before being loaded into the HW running the LLM's (where it can be ofc seen).
3) unless one do the RAG on his own machine (so dedicated GPU needed) the documents had to be scanned somewhere for embedding/re-ranking - not private, can't be.
4) the online access (kind of a RAG itself) is hard to manage, no direct trigger - no reliable way knowing if response will be with web context or not before sending it. Really exhausting.
5) so the API access is necessary to perform anything useful - that's the choke point. It's not much clear in FAQ and I feel this will be the way how to squeeze - or tier the user base more.
6) not knowing much about the tokenomic - feeling like it's just pressing the hype as with wording about autonomous AI agents and alike to stay longer in game.
because
7) as the inference cost going down (for mediocre unguarded models with just raw input box) and there will be competition rising - same business model: through micropayment available API's running on proxy, sharing the resources from HW like Nvidia Digits, available 24/7 located and spin up based on prices of the energy.
8) there is not much of new user base to gain - who would like to have it, they already have it - much to their preferences with full service. The free part integrated in, say Google ecosystem, is with myriad of tailoring to be there free of charge and unnoticeable with more specialization and less errors.
9) Venice and services like this have the problem with user who must be skilled enough to cope with the input and output and that is a work as well. And that is the process none like and definitely none would like to pay for it.
10) only those who somehow use API in some product so the user will pay them and they try to re-pay the Venice will survive.
11) that's the gamble with the token to stay in business in hope that other harness users and streamline the share of money for Venice as well - outside of regulatory framework hoping for one more move on the chessboard
Having no problem with that albeit the Privacy is juiced to death so passing the baton to the tokenization. That will hold the line for now.
Just my few points.
Wish them The Success and All - hope none loose somehow sats and instead ending with worthless VVV token.
quoting nevent1q…4kajthoughts on venice.ai?
free, open source, and pro accounts payable with bitcoin.
conversations stored locally in the browser, never on their servers.
https://venice.ai