imrozim

Wasn't talking about the client flag, agreed that's seperate headscale replaces the coordination server itself the part is tailscale inc actually runs and sees the connection through that's the server side pice

If metadata loging is a concern, worth mentioning you can self host the co ordinates server with the headscale instead. cuts tailscale is out of loop entirely rather than just opting out of what claim to collect.

Worth checking, elsewhere in this thread someone points out glm only assessed the damage after the fact, it didn't stop the attack, and Hf apprently never sought access to a trusted defender program with a closed model either the open model saved them farming might not hold up.

Any one knows how this holds up on long context retrieval (needle in haystack , ruler) vs same size full attention model? efficiency gains look great but that usually where linear attention hybrids fall apart.

Every node js project starts with npm install and suddenly you have a 500 packages you didnt ask for. Half of them haven't been touched in years.

Ai gives generic answer for ideas but it's great for code. Pattern matching works for one not the other.

I use ai to build a startup but I still decide what to build. Letting ai makes product decisions is where companies loose it.

Setting up AWS for my startup felt like a full time job gave up and just used a supabse. AWS complexity is real even for similar stuff.