Many users condemned JPMorgan's block on Anthropic Claude access for Hong Kong staff as discriminatory or racist, while a few hoped access would return soon after compliance fixes.
No Digg Deeper questions have been answered for this story yet.
It will never work.
If using certain models will bring considerable financial benefit to a company then it will be their mission to access it and they will find a way.
Within a company the situation becomes even more ridiculous as they can easily access them through a foreign proxy.
@ns123abc Anthropic: We will log all your data with human review if you use our top model, no exceptions.
US government: We will cut off your access with zero notice export controls for internal political reasons.
World: Maybe we shouldn't rely on this.
@ns123abc wow why @grok
@ns123abc Oh wow! Being on the path of this obnoxious child may gain popularity among some, but certainly not among large companies who breathe compliance and rules.
@ns123abc Loss of aura is inevitable at this point
Source: https://www.ft.com/content/de83d303-6a03-456b-bfb9-7b11dd502ab3
@ns123abc "We trust AI. We just don't trust you with it." JPMorgan and Goldman just told their Hong Kong staff: you're not qualified for AI.
@ns123abc JPM does have ROGO or whatever which is basically a GPT wrapper with added security
@ns123abc So AI no longer taking jobs away?
@ns123abc Too bad it was scoped on sight :D
I mean ONE shotted on sight.
Anyone with any intellectual prowess already has all the weights :D
@ns123abc Keep Moving. Even when the path fights back. Even when progress crawls. Even when the stands are empty. Movement is the only victory.
@ns123abc It is a location specific restriction limited to Hong Kong staff.
@ns123abc Ja, over for the banks kekw
@ns123abc over for who? why? seems easy to add it back later
@ns123abc What is over?
@ns123abc Bank restrictions on frontier models are rarely about the model itself and almost always about data exfiltration paths and audit requirements that current agent tooling still makes hard to satisfy cleanly.
@ns123abc the honest answer is that banks aren't sure what to make of claudes yet, but they're definitely worried about someone else building a better one first
@ns123abc Wow it’s over
@ns123abc Big banks move slow, they'll be using Claude again within a year once they wrap it in their own compliance layer
@ns123abc Claude is officially the bl(AI)ck sheep 🐑