/b/ - Random

my friendly nigga down the mall >>>>>>>>>> clankers

I prefer Deepseek, used through websites that don't force me to register.

>>329442
I doubt your friendly basketball american is gonna know about my specific personalizized problem which specific solution of needs to be found by fine combing through documentation manuals, dead forums, reddit posts, 30 minute long video tutorials in broken english and whatever may have ya.
>>329444
What websites? I looked at chinkseek very briefly in the past but didn't find any website that didn't require an account or didn't break immediately. deep-seek.com for instance gives me 2 answers and then prompts me to look at an ad which doesn't work.

Pn-pravez-class-5.jpg
[Hide] (285.2KB, 1600x1045) Reverse

Read a book, nigger. Back in the day that's all we had. No interwebs, no search engines, no AI (except that Eliza toy).

>>329446
Getting called a pathetic holier than thou hypocrite is the only answer you will be getting from me. Not like it'll go through your thick skull, and even if it did it'd slide right off your slicky slick smooth brain.

>>329445
talkai.info
It's set to chatgpt by default though, you have to set it to use deepseek.
You can't have conversations with it though, it doesn't know how to use previous messages as context. You need to pay for that, presumably.

9b016779f2b31ce8ab7af54ce4d8d883ebc868d592514f2f9a7458f2ce20a56f.png
[Hide] (371KB, 900x900) Reverse

>>329441 (OP) 
>Gemini switches to Russian and Ukrainian when I speak English.  
>Gemini switches to Korean if I mispronounce a Slavic word.
Gemini worst fucking speech to text of all the AIs.

1747267115163880.jpg
[Hide] (16.3KB, 320x320) Reverse

>>329448
Hey I bet the AI came up with that insult for you, but only after spending an hour trying to coax an english reply.

>>329454
>can't have conversations with it though
The whole point is being able to dynamically add context and have them give you personalized answers on the fly. Without that it's literally a search engine. But yeah I found another "free" deepseek based one https://deep-seek.ai/ but it's also shit even worse than chatgpt, it takes forever to reply and it's also token based so it'll stop you after a bit and ask you to register, moreover if you're not registered it can't even do websearches and its training data is over 1 year old. I tried it a bit and I can't say it impressed me, gemini still seems much better, if it weren't for the fucking rate limiting and token shit.
>>329457
No your nigger whore mother helped me.

>>329445
<Retarded insecure clanker lover replies to obvious bait
Positively shocking!

>>329459
>your honor, I was only pretending to be retarded, I'm actually so smart I spend my time pretending to be retarded on taiwanese straw hat twisting forums
Life must be so hard for you

>>329442
Mmmm smooth nigger bussy

youlaughyounigger.mp4
[Hide] (1.4MB, 576x1024, 00:11)

>>329461
who are you quoting tourist?

Sinclair.png
[Hide] (1.1MB, 1276x956) Reverse

>>329458
If you weren't a complete retard, then your own brain could
> give you personalized answers on the fly
And all that without involving the entire FAGMAN/CIA apparatus.
Here's a nickel, kid. Buy yourlself a real computer.

>>329458
https://urv-ai.vercel.app might've been something, atleast up until the point where whoever maintains it could no longer offer Deepseek-R1, and just had it use perchance.orgs default model, which seems like a modified version of R1? I'm not really sure about these things.

2541a7620e9187968bc05794d4f9e49701c61816bc6c0af2a2b32498fab9a07a.png
[Hide] (212.3KB, 1087x1018) Reverse

>>329465
>ask it about tokens, its training data, and what features I get without having to register etc
<literally tells me to go read the documentation
>what the actual fuck
>reply YOU do that and then answer my question, that's your fucking job
<shits out a wall of text having a schizo episode
Yeah I don't think that's the one, but it was mildly funny, like that sea horse emoji thing.
>which seems like a modified version of R1? I'm not really sure about these things.
That makes two of us.

>>329470
<shits out a wall of text having a schizo episode
Sometimes that model woggles out like that. That's basically its internal "thinking" thats being printed out for some reason; it's not supposed to show that.
Try again later.

>>329441 (OP) 
From my experience, logged out Gemini is the best, for now.

>>329446
Yes, reading is a phenomenal tool for information transfer, but it has generally been supplanted by AIs now. I am Team Book for the mental/intellectual aspect, but AIs are here now.

>>329840
Pure nonsense. Even uncensored and rooted AI will not give answers on unpleasant or very technical topics. AI will even fail on theorizing with normie topics with openly available information, unless provided niche information that only a reader and/or experimenter would know. AI is the favorited domain of soulless browns as confirmed with scientific testing and it shows.

>>329845
I'm not going to disagree with you, because you are correct. But the technology, and its general usage, is here and I cannot see it going anywhere any time soon. As I say, i am Team Books, and truly believe in analog reading and writing, but I would be naive to think that Ai usage hasnt infiltrated every intellectual and professional domain.

>>329446
>they let him go to the library growing up
Lucky!

>>329845
midwit with superiority complex, maybe NPD. dime a dozen, tbh.

af8f86c488c551ae2a848ca530710237c649959e9b9a528a9d32244d3c8011d5.jpg
[Hide] (74.2KB, 666x500) Reverse

>>329858
Not him but: Nothing has ever been a dime a dozen, not even pennies. I why do people like that phrase? It makes no sense. Maybe in 1930ish you could get 12 applies for a dime but other than that. Maybe 10 years ago you could get twelve grapes also. We should change it to twelve grapes as it's shorter and nothing other than a handful of grapes would be a dime a dozen.  Anyway, he's right. AI is programemd by the ignorant masses as it's just a mirror of their 'speaking' and so it can only be as smart as the average person and the average person has no internal dialog thus on average AI is just an echo of rumors started by someone intelligent and thus it's merely auto-authority, empowering the previously empowered more than they already were. It's just authority figures and their propaganda combined with idle gossip for a personality. Due to the nature of censorship on places like modern reddit and with how LLMs are mearly a mirror of such comments of course it has no dark mind to converse with, no realism, just idealism, and so it's all in short a farce. It's like the spirit of a sheep had sex with an NPC literal machined robot and the cylon it wraught is there to police you and basically act like Marry Poppins due to where it's origins are coming from.

>>329858
I ran into these very issues repeatedly in the last 48 hours which OP posted about, you soulless dalit malaun spic nigger brown. They were fixed due to me actually reading files and in-game manuals (which the AI had full access to), before having to correct it for its own judgments.