Everything should be simplified for the llms
@bentossell
-
o3 Model Performance Varies With Different Instructions
By
–
Yeh I dunno if it’s some memory or instruction that veers it off. I felt like past instructions helped the models but with o3 it performs differently when changing instructions
-
o3 Model Experiencing High Hallucination Rates
By
–
o3 is hallucinating a lot, not sure if recent or always been true thought it was the smartest model?
-
Tools Should Implement Security Fixes, Not Users
By
–
its for the tools to implement, not the users e.g. supabase security advisor – one-click fix for security issues
-
Supabase one-click integration with Lovable and Bolt happening
By
–
ok so its happening in fringes atm. also, connect supabase in one-click with lovable/bolt so its technically feasible i think
-
OpenAI Will Want Browser Data Beyond Authentication Methods
By
–
not saying you'd replace sign in (you guys know more about auth than me!) but cookies, browser profiles, history etc is all extra context. so im saying openai (and others) will want it all.
-
First successful experience co-creating backend with Supabase MCP
By
–
havent actually managed to use claude code yet (terminal is another scary place for me atm) BUT adding supabase mcp and saying create the tables, etc worked SO easily I've never felt like that co-creating a backend service (yet)
-
MCP Protocol for Automated Backend Service Generation
By
–
who's building a protocol like mcp but for generating backend services its so easy to use an mcp like supabase to setup your project, tables, etc. its still up to the llm to setup the code for your application to work i want the latter to feel like the former, and only a
-
AI costs become outdated quickly as token prices drop
By
–
i kinda think anything around 'how much will it cost' is too out of date as soon as you say it. ie everything is getting cheaper in tokens/context