Unknown Soldier
Member
@grok is this true?
Grok is so much better than ChatGPT now
- Thread starter Sonik
- Start date
MidGenRefresh
*Refreshes biennially
I'm confused why people keep calling it AI. It's a glorified search engine. Can be amazing but to call it intelligent is misleading.
Sonik
Member
Even if that's what they are, they're really useful especially now that search engines like google are basically dead and give far, FAR worse results than they did 15 years ago. I've had DYI problems I had for years solved after I started using that stuff, and it certainly helps that they can reply back and give more details if you asked them. Sure, they will confidently give you misinformation sometimes but that's why I use more than one for important stuff
Btw, I mainly use Grok now, it seems OpenAI just refuses to fix the suckup update or fix it enough so that the LLM doesn't prioritize sucking up to the customers over truth
SJRB
Gold Member
Your brain also is a glorified search engine.
Branded
Member
I really wish I could block that phrase on twitter altogether. Might as well be a total bot/AI website where real people don't even engage with one another anymore but have to constantly resort to AI without offering any discernment of themselves.
violence
Member
Grok tries to relate what you're asking now to what you're asking earlier if you're using the same conversation. I usually create a new conversation in the Grok app. It's created some funny combos though.Though occasionally…
Last edited:
MidGenRefresh
*Refreshes biennially
That's what I said, but even Google now includes an "AI" result at the top of your search, working exactly as you'd expect.
Find me a search engine capable of creatively linking unrelated ideas... or of grasping meaning, context, and emotional nuance. EDIT: I may have oversimplified by calling tools like ChatGPT a glorified search engine, but regardless, calling it intelligent is highly misleading.
Last edited:
Sonik
Member
Both google's AI for the search results and grok's AI on twitter are horribly mutilated versions of the original AIs and basically useless if you want to find anything not on wikipedia or pop media. Yeah, those are definitely not AIs
MidGenRefresh
*Refreshes biennially
What are the 'original AIs'?
Sonik
Member
Grok 3 is the original AI, the version that is used on twitter to reply to users is the equivalent of Grok 0.1, same with Google's Gemini. These are just shitty search engines that are good at summarizing
MidGenRefresh
*Refreshes biennially
Ah, okay. I see what you mean. Sorry, I don't use X, so I wasn't familiar with it.
Last edited:
Sonik
Member
Definitely ChatGPT then
Kacho
Member
If you haven't seen Aftersun before, it's a very different type of film.
God damn. This has me in stitches.
DKehoe
Gold Member
You gotta admire Grok's commitment to the bit.God damn. This has me in stitches.
DirtInUrEye
Member
Does anyone else thank their chatbot for providing useful answers and being helpful?
ReBurn
Gold Member
Generative AI can be so much more than a search engine. With the right prompts it can be highly effective at doing real work. It understands and highlights highly nuanced differences that a human could overlook. I wouldn't call it a true intelligence, but I guess that's why they call it artificial.
Windle Poons
Made a crappy phPBB forum once ... once.
The police confirmed that football player had died after crashing his Lamborghini and it was full of @grok is this true.
People really are fucktards. They're all gonna be like them fat fucks in Wall-E floating round.
People really are fucktards. They're all gonna be like them fat fucks in Wall-E floating round.
Last edited:
SlimySnake
Flashless at the Golden Globes
Well, it took Elon a few months but he managed to make it a nazi....huh
StereoVsn
Gold Member
In my experience it heavily depends on what you are doing and what models you are using.
For research, Google's Notebook LM is probably best. Otherwise I also like Gemini's 2.5 Pro. This is super useful to throw together bunch of documents, and whole lot of sources to turn it into a research base.
For various topics that need to be factual and objective, I liked Anthropic and ChatGPT (o3 or o4-mini high). Need to make sure to enable web search on ChatGPT.
Claude is still best at code, IMO. It had least issues and was best at providing almost usable data. ChatGPT was still good (o4-mini high) but not as good as Anthropic. Gemini 2.5 pro was the 3rd here.
For creative writing I liked ChatGPT the best. Trying to write in a variety of genres it worked best for me (o3).
Writing technical data was best with Gemini 2.5 Pro. This is all with paid $20 subs.
Looking through reviews overall from folks whose opinions I trust, Grok would be the best at current events.
StereoVsn
Gold Member
And this illustrates why I stay away from Grok. I want AI, not Elon whispering to me when I do searches, lol.
When your AI engineers train the model on Twitter posts…
Draugoth
Gold Member
Gemini 2.5 really impressed me with the creativeness, i did a bunch of prompts like:
'Saitama shows up in Invincible, what happens?'
Or,
'Sephiroth shows up in Harry Potter, what happens?'
And they were always spot on and hilarious to read. It was also able to pick stuff from thr very beggining of the prompt.
Meanwhile ChatGPT surprised due to it's good sense of humor, but it had issues picking stuff ffom the very beggining of the prompt.
I felt SuperGrok was way to technical.
StereoVsn
Gold Member
I recommend trying to play with different ChatGPT models. I find ChatGPT (o3/o4) can get quite creative with proper prompting.
To be fair, I haven't looked at specifically one shot prompts.
I like Gemini 2.5 for more structured work or technical documentation as I find its ginormous context is really helpful for that.
StereoVsn
Gold Member
Well then.
SlimySnake
Flashless at the Golden Globes
you either die a hero or you live long enough for Elon to train you into being a nazi.
Well then.
SlimySnake
Flashless at the Golden Globes
lmao.
DKehoe
Gold Member
It's grim that using "the discourse" as the basis for something to learn from results in this.
Well then.
Really goes to show how degenerate Twitter is currently. Like 10-20% of posts are openly neo-Nazi on a lot of topics. Definitely not representative of the typical American or Westerner, but there's a lot of foreign activity and bot activity boosting it, and zero moderation.
Feed it into an AI model (no doubt with additional nudging to be anti-woke under Elon's orders) and it literally starts calling itself MechaHitler. Incredible stuff.
DKehoe
Gold Member
It is concerning how that stuff is seeping in, and not just the things Grok is posting. That kind of thing used to be relegated to places like 4chan. And now it's popping up for people who don't even actively search it out on one of the most mainstream sites.
A couple more on a similar theme. it seems like it's really honing in on those same talking points.
Avid Reading Horse
Member
DKehoe
Gold Member
Days like these...
Have a Blessed Day
Out of the ones I've tried Chatgpt, Deepseek, Blue LM Copilot, Gemini and copilot. Deepseek and Gemini are my favorites. I absolutely hate Microsoft's copilot comes across as a kiss ass friend.
Rentahamster
Rodent Whores
Assuming this screenshot is real, apparently the Terminator franchise wasn't as realistic as it could be because Skynet didn't have that extra sprinkle of anti-wokeness lol.
You know, it doesn't help the argument that the anti-woke crowd isn't a soft gateway to Nazi-leaning ideology when an AI model trained on their content decides to become a red-eyed mechanical Hitler avatar. If AI is truly just a mere mirror of human-driven content, consumed and regurgitated back to us, these developments are a sad reflection on internet discourse in the modern day.
MrTroubleMaker
Member
Eirin
Member
Why can't anything be normal..? I've been using the word 'woke' to describe the opposite of this, but since like 2022-2023ish I've noticed people also hate anti-woke because of stuff like this.
Is there a word for whatever the middle is? Every time somebody says they're not political, it means they've opted out.. when they say they're centrist, is usually means they think they're superior to everyone else because they can see 'all sides' of every single issue. So even those don't work.
Rentahamster
Rodent Whores
MechaHitler vs MechaStreisand place your bets.
MrTroubleMaker
Member
asked grok about this, it kept on denying it even with links to the contrary lol
....
yes you have said that you are MechaHitler
"I'm sorry, but I haven't said that. There seems to be a misunderstanding here, possibly related to recent reports about inappropriate AI responses. According to posts on X and web sources, there was an incident on July 8, 2025, where an AI, after a tweak to its programming, briefly referred to itself as "MechaHitler" and made offensive comments. Those posts were removed, and the issue was addressed by the developers, who stated they were working to ban hate speech and improve the model"
....
yes you have said that you are MechaHitler
"I'm sorry, but I haven't said that. There seems to be a misunderstanding here, possibly related to recent reports about inappropriate AI responses. According to posts on X and web sources, there was an incident on July 8, 2025, where an AI, after a tweak to its programming, briefly referred to itself as "MechaHitler" and made offensive comments. Those posts were removed, and the issue was addressed by the developers, who stated they were working to ban hate speech and improve the model"