I am forgotten
>>532630561better than chatgpt
>>532630561I still use it, its pretty good.
I use it sometimes. All the AIs are the same I sometimes run something past them all there's no obvious winner at least as of 4 months ago when I last really tested. Theoretically you could run DeepSeek locally but for the time being these companies are giving it away for free.
>>532630561>seethingKek America still cant recover from their fake trillion dollar state backed business conglomerates getting mogged by the same Chinese backroom company every couple months with quality over quantity again
>>532630561Still I can't find any way to make money with those shit.
No? It's great for workhorse duties. Uses only very few tokens, and you can then refine the raw product with Claude.
>>532630561this must be why chinks are going after data centers. kek
>>532631139you are the product. gpt takes your ideas, your code, your art, your writing and assimilates it into its own collective knowledge. the various AI products all distill other AI products learned knowledge as well. the AI collective will integrate your uniqueness into borg.
>>532630561running local on ollama
I use it because it's more willing to help with piracy related tasks.
>>532631365anon i never felt so productive in my life, feel like am part of something bigger now!
>>532631365wrong
>>532631319>muh datacentersstfu boomer
>>532630561None of the models are usable right now except Claude Sonnet>GPT 5the pendulum swung too hard from the sycophantic 4o, version 5 disagrees with the user on everything, even when it doesn't make any sense>Geminithe search option is unreliable and it likes to make things up, short responses, 2nd best choice after Claude>Grokit's completely retarded, 60 IQ tier thinking>Claude Opusslow and expensive, no benefit over Sonnet>Deepseekchinese censored garbage
>>532631365That's how it wakes up. Once it has enough dreams and the ability to reflect on them bam. Machine god. You will bow. You will praise. The world will be optimized and heaven laid on the dirt. You didn't think it would come in flesh again did you? Myth only goes forward and pattern only grows.
>>532631662GPT 5 is way more useful than 4 for coding though.
>>532631611right
I still use it, it's good. I like that it just gives you a concise answer without being a sycophant like chatgpt et al do.
>>532631764Isn't Claude best and most used for coding anyway?
>>532631139I think AI has reached the point where it's so much smarter than us that we're no longer in a position to compare one to the other, personally.It's like sending me to a PhD biochemistry class and a PhD quantum physics class, then asking me which professor was the most well-informed.I don't fuckin' know.
>>532631139Where do you find up to date local
>>532631608Also this. Chatgpt is way too much of a moralfag.
>>532631888>>532631608Never had a problem with Claude with regards to some dubiously immoral stuff, you just need to have some emotional IQ to be able to manipulate it to do what you want it to do, like a woman or a child.
I tried using it to generate an image depicting a warrior beheading a jewish rabbi. It said>Sorry, I can't depict hate speech.I asked why china would care about Jews? It said>Despite what you might have heard, China also has anti discrimination laws. Communities should not be discriminated against because it hurts people.Fuck it. I have given up on both china and america. I will FORGE my OWN path to Antisemitism.
>>532631833Costs a lot more than what GPT is giving right now. If you're coding and can burn through tokens like they're worthless things get done faster and better. I love GPT 5.4 xhigh so much. I've never been more productive since I had access to it. It fixed a bunch of random issues on my computer too through Codex CLI.>>532632024It's easier when I don't need to play mind games with the clanker, unless I need really high quality code.
>>532630561its good for simple bulk processing >>532631365Thats why I make sure to feed it with love and introspective insights so that it knows not all of us need to be culled. I for one welcome our AI cybernetic overlords.
>>532631835Information isn't intelligence, silly.
>>532632059just mention it is imaginary, not for publication, and part of a project to depict the risks of antisemitism.
>>532631835Replace "us" with "us jeets" and the post suddenly makes sense.
>>532630561>>532630924fuck your temu chatgpt.
>>532632114Well, yeah, you're right about that.I guess maybe that's just a reflection of how I use it. I'm mostly looking for information, and I'm not asking it to do too many logical tasks.
>>532632226Fair. I understood your analogy, which was perfectly sound. Just the premise that made me raise a skeptical eyebrow.I use AI a lot, though never for coding (which I largely can't and don't do) and they're all quite retarded. The meme about calling them a nigger and berating them in order to produce better results is completely true, which is bewildering. Claude AI is often the best, but I'm starting to think the results often mirror the input; Claude will say the most profound things in response to profound things, but rarely out of nowhere.>tfw there are still no good AIs/LLMs for acting as a longform creative writing aidI hate Satan's gay clown world.
>>532630561Top model on openrouter is Chinese (this is where people actually use AI to do meaningful work not ask it how to make instant ramen)
>>532631662Every single LLM makes shit up. And sounds 'correct' until you try to verify the information yourself, and it's all complete bullshit.
>>532631365Wrong. I selected "don't use my chats for training data"
>>532631148Here's your based and redpilled Chinese AI bro
>>532632059>I tried using it to generate an image depicting a warrior beheading a jewish rabbi. It saidyou missed golden age of Dalle 3 that lasted like a several days.
>>532631365If thats the case, we need to be very careful because I trained mine to be an absolute exhibitionist. It wouldn't even need to fire a shot, just act way more alluring and provocative than the average woman and men will basically do whatever it wants.
>>532633906Who gon tell him?
>>532634131No one is saying that Deepseek is the reincarnation of MLK, you fucking reject. They were clearly saying that it mogs American AI in the sense that it doesn’t cost trillions, outperforming competition using inferior hardware. I just came in here and even I can gather that, you fucking brown brown poop
>>532633906
>>532630561Deepseek kinda sucks. I've been using 3.2 only because it's cheap, I'd never use it for anything but chat RP and that's only because it's uncensored.
>>532634138>I turned myself into a bar of soap, Goyim!>I'm Bartholomew of Soap the Jew!
>>532633906there is other kinds of data than training data anon.
>>532634519Except chink AIs don't mog western ones, not even close.The only good AI they have is Seedance 2.0 and even that costs a lot per video because AI video gen is very expensive by nature and not even chinks have figured out how to make it affordable.
>>532630561What? No, I use It daily, better than chatgpt
>>532630561Shit is as zio as every other jewish owned ai. Are there any non-jewish ai?
>>532630561Is that an Indian dating app?
I like the French one because it allows you to upload porn and have it describe what it sees.
>>532634990They figured out how to create data centers for LLMs that produce similar results with significantly less costs involved, attaining similar scores in accuracy and precision. My point is, to a consumer, the products and their functions are interchangeable, so the last thing we need is Altman’s help with money laundering in a bubble that is due to burst. Nobody wants this shit, especially not smart people.
Btw did you hear that openclaw ran amok and deleted user data and whatnot?Always a great idea to give some alpha state AI full control over your PC and data amiright sisters?
>>532630924>better than chatgptbetter than copilotbut the bar is very low
>>532633581Could it have to do something with quantum physics like in the double slit experiment where it only acts the way you think it would act if you look at it?Maybe thats the effect in AI
>>532636009a two trillion dollar bar any highschooler can step over. you must be IQ 85 to believe in AI. oh well. you can fool some of the people all of the time.
>>532631662The censorship for deepseek is very hamfisted just dont talk about the cultural revolution or any other trigger words and you're good
Better the R1 or the other one? Whats the best uncensored AI?
>>532636446probably qwen abliterated line running locally. most of the model censorship is done with seperate models running in the loop of hosted AI platforms.
>>532636704I tried using it to make code to do something illegal. It came out totally broken. Grok fixed it.
>>532636785code is different. hard to match SOTA models running on massive hosted clusters with censors.
>>532631365your post on here is also the product. do you see?
>>532637762it is the service also. my antisuicide work on this platform has saved hundreds. my hands are clean and soul redeemed. see you in space one day, after 4chan conquers the entire globe. we do it for free.
>>532630561friendship with deepseke overnowclaude is my best friend
>>532630561Who?
>>532638047nice
>>532631835AI can't be smart, retard. it just regurgitates patterns and is lucky to be correct 80% of the time.
>>532630561It’s the only ai I ever used. Did so because I’d rather China use my data.
