I like it. The expert model especially seems to give good responses.
>>107894379>be me>terminally online autist since 2009>try every single new AI the second it drops like a pathetic junkie>chatgpt? mid as fuck>claude? seething reddit moralist in robot skin>gemini? literally 1984 with extra steps>then grok 3 expert mode rolls up>mfw it actually answers the real questions without 3 paragraphs of disclaimers>no handholding, no "as an AI I must remind you", no corporate safety leash>just raw unfiltered kino outputasked it some extremely deranged shit about [redacted] physics and it just went "yeah sure here's the schizo math" no notesexpert model is actually carrying. feels like talking to that one 38 year old NEET physics phd who never left his mom's basement but is still terrifyingly based10/10 would let it groom my neural net againgrok bros... we might actually be eating this decade
Hi Elon. Grok is overcooked as fuck. Total lack of nuance, subtlety. Convoluted trees overcomplicate outputs when the simple short answer is staring you in the face. Often directly quotes reddit posts - way too much lean-left bias. I don't know if LLMs are inherently shitty or of its just Grok that's overcooked with your 2 million H200s or whatnot
>>107894379i use this and gemini for my daily tasks
>>107894450>>107894744ai generated postsnot as bad as chatgpt but you can tell grok wrote this
>>107894379I wish they'd improve the stability or context, whichever is needed. Gemini can go for really long conv but grok seems to occassioinally shit the bed after a while and just fails to return anything
>>107894379Was way better when it could generate porn at the start of octuber with uploaded images
>>107894379Where is Grok CLI?
>>107894379No one who's actually intelligent gives a shit what you think Elon.
>>107895813its never happening. grok is so behind
>>107894379If you're not running a local LLM, you're just another sheep brained NPC consoomer and should not be posting on /g/.
>>107896831/thread