Grok-1.5
x.ai
x.ai
I can’t see why I would want ever want to build on Grok especially when it’s not the best model nor is it fully uncensored, all with the special sauce of Elon rugging your access based on how he feels or troving through your conversation/api history in an act of petty vengeance.
Deploying it in production seems like nothing but a desire for pain.
Im an appsec engineer so by the nature of my trade im generally using it to look up things related to exploits, osint, or threat intel.
I actually canceled my ChatGPT subscription because when they implemented some automated security controls to limit the use by APTs it started to refuse to do anything i asked it too. It was different than lazy. Like refused to analyze anything for me. I don’t have proof this was the cause but I was routinely asking it explain blog post from google project zero. Something OpenAI said was a behavior exhibited by APTs using their platform. None of the devs at my job reported similar response behavior.
I tried gemini pro and it’s responses were on par with GPT but it was short winded and felt slightly better then their generative search.
I’ve been using Claude, and it’s been by far the best in terms of understanding my half baked fringe train of thoughts.
I saw were Elon said it was more lax in terms of what it would respond to but he’s trying to make money and I doubt it.
Does grok shine in any areas? If not response accuracy maybe app features make it worth it?
I do run local LLMs but I have a 2060RTX and it’s a bit like playing with an old dog.