This is not a bad idea. I don't know why you're being downvoted so much. Lots of people hate corpo LLM, understandably. But then when someone try to use LLM for user benefits against the corpo? The people here... still hate it. I never understood why.
Personal Finance πΈ
US-centric community for content related to consumer retail banking, credit, & investment.
He said the no-no word and the brainless anti-AI masses swarm to downvote without engaging or discussing. Par for the course here unfortunately.
I think there is also widespread support for KYC-overachieving banks -- because there seems to be no pushback from consumers. I think most people believe only criminals need privacy and they are happy to let banks snoop without warrant or restraint to catch the bad guys. They give banks whatever data they ask for and answer their interrogations without hesitation.
So I really canβt be sure to what degree the contempt for my idea is anti-AI hysteria or pro-KYC boot licking.
Iβd bet money itβs way more blind anti-AI than pro-KYC.
Iβm definately in the corpo AI hating group myself because it really looks bleak for LLMs controlled by NGOs and tiny operations. It looks like we must choose between licking the boots of a tech giant to get chatgpt-like service, or be excluded.
There are exceptions though for some services. I run argos-translate to do offline language translations using FOSS. Itβs entirely free world stuff. I wonder if the answer is to promote argos-translate while spotlighting that itβs an LLM that you can self-host.
FWIW, as someone working in fintech in the EU, that "KYC over-achievement" is not as overzealous as you think it is. Each of those pieces of data are very useful at making fraud very expensive for the fraudsters. They need to burn a lot of capital to compromise people or fabricate personas.
And, at least at my place of employment, we take the PII protection very seriously because of GDPR.