News for hackers. Stuff that matters.

muhammad-shafat

I get where you’re coming from, but there are some security practices in place. The host client renders views inside a strictly sandboxed `<iframe>`. Any action the UI wants to take must pass an auditable message back to the host application, which triggers an explicit user-permission prompt.

Show HN: Investigating why GPT-5 has made ChatGPT 'broken'

https://muhammadasmulkana.substack.com/p/why-chatgpt-feels-broken-since-the

I've used ChatGPT since GPT-3.5 for coding and prototyping. GPT-5 broke my workflow in specific, measurable ways.The router system switches between model variants invisibly. Responses became verbose but hollow. Instruction following collapsed.I discuss these points in detail in the article

( Read More... | no comments | on HN )