Send it all to zero jesus fucking christ
https://x.com/AnthropicAI/status/1925926102725202163

design @splits | /wireframe /latenightcrew /event-pass /feud

Burning playbooks @ Heretic.wtf Prev: CMO at Aleo, RabbitHole, Gaia + a bunch of agencies.

Currently, OpenAI has the best product experience by a wide margin IMO. I don't really see a world where Anthropic takes that position and also becomes the household name, unless OpenAI commits suicide.

And their focus on safety IMO is a complete distraction that says:
a) we're going to prescribe our morals onto you
b) we're going to waste time debating what is "safe" instead of building the best product

I also don't think their models being slightly better than the current SOTA for various windows of time is going to matter much in the end. Models are commoditizing.

I’ve been coming around to anthropic besides the surveillance state stuff

i dabble with ai using open identity data from crypto https://alexpaden.tech

Linguistically it’s better written which actually has a high impact on some engineering problems I work on (even just describing the architecture for example)

It’s far better imo at writing rap lyrics which is essentially also poetry

No I use Claude web to discuss problems and write lyrics and I sometimes use Claude in cursor/windsurf

I don’t understand the value of Claude desktop for someone like me because I’ve never consistently had the agent provide the right answer on first attempt

you use the Claude app to code? or do you mean you use the model in cursor/windsurf? (I also use sonnet 3.7 often in cursor)