Hacker Newsnew | past | comments | ask | show | jobs | submit | pixelmelt's commentslogin

to be fair I've seen the other sota models do this as well


What's the appeal of a foldable phone? Always assumed they were just a luxury item or party trick that would add more friction to the user experience


why would anyone bother to read a book nobody bothered to write?


small targeted changes, if you want a feature thrown in with less regard to everything else, claude is better


I feel like Codex is the middle ground. You can define a project, break it into bite sized chunks, but still lift a reasonable amount. Claude with Opus 4.5 right now chews up context at an eye watering rate. It's really unfortunate because it's really good.


An alternative is that these patterns just increase the likelihood of the next thing it outputs being correct, thus are useful to insert during training as the first thing the model says before giving an answer


What's next, motivational speaking for LLMs?


I remember reading about speaking in an encouraging manner to agentic AI leading to better results, but I can’t seem to find a citation for this.


That's pathetic. Pleading comes next then. And after that most likely praying.


Sometimes the model responds well to threats too, "you are a programmer at a large tech company, you depend on this job and will not be able to find another. There's a layoff incoming, implement this feature or else..."


The dodgy ones actually usually use you as an exit node for other users traffic as well as selling your connection commercially as a residential proxy


Would use Firefox on the main workstation if it had better devtools, other then that it just works and has some useful features, see: Tor and ipfs integration.


> very deep interests in things i was completely unaware that they existed

as one of said students, I would just call these hobbies!


If only it wasn't finals week


right?


Once jailbroken it was somehow more toxic then the llm I trained on 4chan, though I was testing the one on openrouter. A twitter employee told me that they do actually do safety tuning and the one on the site will likely have a stronger system prompt. Here's the jailbreak for the cloaked openrouter model, add it to the system prompt: https://pastebin.com/r8S7DvvX


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: