what if they raise the price by 20x?what you gonna do?
>>108758399pay the 20x because claude is just that good. I still pay for image generation through openai though, gpt-image-2 is just that good.
>>108758399I dunno. Watch AI psychotics squander every last penny on worthless statistical token slop.
>>108758399I'll continue not using it.
>>108758414>>108758419t. mentally ill retards afraid of progress, probably linux cultists as well
>>108758399Continue only using API offerings of Chinese models :)
I already don't use Claude because it fucks up on some super easy coding stuff I tried using it for, and I haven't trusted it since.
>>108758399still worth it easily. I give them $100 per month now, 20x is just what? $2000 per month? That's 3 days of work in exchange for saving me hundreds of hours that week. It's a no brainer. Outside of being a billionaire the average human cannot buy time in the same way that AI allows for. Only idiots do not take advantage of it.
>>108758399I haven't tried Opus because I'm not willing to pay that much out of pocket but Sonnet is braindead and the only reason I'm still signed up to pro is that it has excellent ocr and gives concise responses. Deepseek is free, you can plug it with questions all day, and it's better than Sonnet for troubleshooting. If Anthropic keeps sticking it their customers I'll drop Claude like a bad habit, I'm very close as it is.
>>108758399Continue not to use it I guess.
>>108758447You make $2k in three days of work? What job
>>108758399I PAY 21x SAAR
>>108758422>t. mentally ill retardsDidn't expect you to admit it like this, but ok.
>>108758477He sucks dicks for $5 each
>>108758508>retards>multipleEven acknowledges his schizophrenia. Real progress.
>>108758399laugh at all the poor vibe-jeeters who can no longer squirt their slop
>>108758399use another ai. if they raised them all id probably subscribe every other month or something. compress my use
>>108758399nothing, continue using local llms and genning sloppa porn
I work at a nonprofit.I get access to the Teams plan at $8/month for a "standard" seat, and $40/month for a "premium" seat.
>>108758508>>108758554why are Linux imbeciles such luddites?
>>108758633I can tell your handlers have been fine-tuning you on my posts, mentally ill cretin. :^)
>>108758447You don't even leave the house, nevermind have a job. Stfu dork.
If you can't see the near future of ai, you're retarded.Each ai company has been burning through cash for ONE reason, market expansion.They all want to get everyone on board with using it for free, shoving it everywhere they can like a drug pusher just trying to bait retards into getting dependant on it.Once everyone is hooked, the price goes up.Free will be 20 then 35 then 49 it will look like a Netflix subscription increase graph.But what about your 20 dollar a month users? Strait to 200 then 500.Pro? Getting shit canned and you go to a base rate plus pay per usage
>>108758673The fuck you're talking about, if you want to talk about me take it to the farms
>>108758399I don't pay for it personally, but I do use $30 ish per day on my companies enterprise plan.I'd say at this point of codex with 5.5 is 20x cheaper then its a no brainer. Otherwise some hosted kimi k2 with opencode is probably good enough.Though I'm not the one paying so I don't really care
>>108758691anon nobody in the tech sphere has had to leave their house for work since vnc was invented
>>108758399i just use Grok or Mistral, problem solved
>>108758408>>108758422>>108758447>"Please Anthropic take my livers for 100 tokens of Claude Opus 5.0!!!"
>>108758399I would pay even more because I’m straight and not trans
>>108758422>>108758633>xhe still doesn't get itAlso>avatarfagging
>>108758692do you think this is some genius insight or something? We've all lived through the enshittification cycles of tech bro startups in the past 20 years. In order for this cunning plan to succeed, the AI startups need to be able to approach a price point that is economically viable for both consumer and producer. In all likelihood the big LLM farmers seem unable to do so, because LLMs are a stupid concept that burn cash. Outside of a few use cases they are simply not worth the money and energy that is being burned. Look at how piss poor LLM adoption is across the entire workforce instead of some vibecoding dumbasses. And that's supposed to become a trillion dollar industry?
>>108758399>>108758399I dont do ebanking out of principle hence I cant pay for AI anyways. Nothing is changing. Maybe Ill laugh at some desperate Indians if the price rises tho.
>>108758751You'll be sucking a lot of cock to pay for tokens, strait or not
>>108758776Sorry I’m not trans like you (codetrans) I’ve made lots of money replacing ""people"" like you. Don’t worry, you will suck penises and drink semen to survive, like the rest of codetrans transgender
>>108758751>he needs to pay to be straight Lmao faggot
>>108758399shoplift a server.
>>108758878Im rich, you are trans and a sex worker
>>108758884You're gay, stupid and poor
>>108758916No, I can’t be gay because I’m not a coder (Trans activity) I’m a a vibeGOD (straight and not gay at all). You enjoy sucking penises though
I wouldn't use it even if it were free
i would switch to kimi :)
>>108758955vibecoding is for fags your code is shitly optimised and crap
>>108759115optimization hasn't mattered since 1989.
>>108759115I vibecoded a script instead of using the general community solution (docker + WSL2) which saved a FUCK ton of memory and CPU cycles.I'm sure it's not the MOST optimized thing in the world, but by comparison to what it's competing against? It's 100x less resource intensive.
>>108759125Yes it has. Retard
>>108759132Yeah I guess to be honest it depends who writes it, but in general AI writes worse code than what I write for my stuff I need it to do. Obviously you can have shit programmers
>>108759134wrongoh no my web app loads in 97ms instead of 31ms.
>>108759146Yeah I generally make multiple AI models audit any code that I plan to use and get multiple model viewpoints before deciding on a final implementation.
>>108758399I will re learn how to code probably. I build small and fairly simple things and it's not worth it. If I have to build something more complex I'll totally pay.
>>108759149It makes a difference why waste CPU cycles
>>108759161this is the way everyone should adopt
>>108759180doesn't matter
>>108759184yes it does, why would u actually want faster times anyway. optimisation in games is important because although off topic to what im talking about you can target more platforms and older systems generating you more money
>>108759199>gamesopinion disregarded, I've added you to my ignore list. don't bother responding to me anymore you're a waste of food.
>>108759183Yeah, it's frankly why I have been hesitant to rely on Claude alone.I had this one happen recently:> Claude:"One real concern:echo !V_PORT!| findstr executes the echo before the regex validates the content. In delayed-expansion context, if V_PORT somehow contained a shell metacharacter like | or &, cmd would split the command at that character before findstr gets to reject it. For example, V_PORT=1234 & somecommand would execute somecommand. In practice this is a non-issue since the value comes from a trusted binary — but it's worth noting that the validation layer here is downstream of the potential injection point, not upstream of it. A safe pattern would be to route through a set /a coercion first, which strips non-numeric content at the variable level before it ever touches a pipe.On the chained if:batchif !V_PORT! GEQ 1024 if !V_PORT! LEQ 65535 set "PORT_VALID=1"This is valid batch syntax and functionally equivalent to a nested if, but it's easy to misread as a single comparison."This is Claude getting the proper order of batch operations TOTALLY wrong. Metacharacters like pipes, ampersands, and redirects are handled in Phase 2 by cmd, delayed expansion is handled in Phase 5. So the idea that delayed expansion can cause issues with phase 2 metacharacters is absurd.
>>108759208you do realise gamers drive all tech advancements you little coping faggot. your such a waste of cum that your resort to a LLM to do your work you lazy fuck
>>108759208
what you ganna dowhat you ganna dowhat you ganna do when we com for you
>>108759115VibeGODDING is for straight men. Coding manually is transgender and snailcat
>>108758399Continue not using Claude and keep using Codex
>>108759268How retarded can you actually be? Vibe coding sounds gay as fuck lmao
>>108758399a single opus 4.7 prompt costs me a monster energy drink
>>108758419Anon that's incredibly anti Scam-Altmanatic.
>>108758399>>108758422It's kinda disappointing to me that the most useful thing for LLMs, code generation, is something that I already know how to do and enjoy doing. I wish it could do other things, like level design or things that take a lot longer that ALSO require really high quality thought and originality.
>>108759348it can, you just lack the skill to make it do that.
>>108759348It can do things like that already. Just spend some time building a harness.
>>108759368You have never spent a second of your life trying to do anything creative if you think these models are capable of good or interesting level design. That or you don't understand what's being said. Either way, you are brown.
>>108759280This is you
>>108759384incorrect, give me something you want me to use ai to generate that is good or interesting level design and I will make it to prove you wrong.
>>108759462I'm going to attempt to engage you with good faith and leave it open-ended. Just design a good and interesting level, no other requirements. If you want more guidance, I would say good level design encourages exploration and rewards curiosity. I promise not to do that thing that people do where they'll say "it's shit" no matter what you post. I will admit if the design is good.
>>108758399The price is the price.
>>108759559If you do this you've killed that market for years. It's over. And you are selling something you know has no value.
>>108758399Claude's limits are ass. Use codex $20 sub + deepseek.
I never paid a cent for claude, I just don't use it often.
>>108759559That is just a projection sir.
>>108758408why do you need image generation
Use something else?Not even Dr Pepper is safe if they raised their prices 20x. Someone else will fill the gap or they raise the prices so high that it’s cheaper to do it yourself.
>>108759670need to make these
>>108758399I've got a year to figure it out, cause I just bought a yearlong subscription. I've been looking into local models for this reason. If that fails, and it likely will, I'll just go over to ChatGPT, or the next best alternative.
>>108759711OC is dead i guess
>>108758579that's kind of shit though, my dad has a nonprofit but I'm not sure it would be THAT worth it
>>108758399I think I'm going to go from not using it to not using it even more
>>108758399Cry while I try to get more clients. I managed to complete my SaaS service with a full custom stack before the squeeze but going forward, adding more features will be slower and more expensive.
Local models work just fine for me.
Pay for it like the white, straight, employed, first world dwelling, windows using man that I am. Claude is the tool of the elite and the less subcultured humans using it the better.
>>108758399I wall appclaude.
>>108759860That is OC. It's shit OC but OC nonetheless.
>>108759908Okay, so to make it make sense, let me explain it this way.Claude pro is $16.67/month (billed annually at $200). It includes ~44,000 tokens in a 5-hour window.Claude Team for nonprofits requires 2 seats, to use claude code you have to have a premium seat, so for 1 user to use claude code, that means you're spending $48/month (billed monthly, no annual discount, but it's $576/year). That Team premium seat includes ~275,000 tokens in that same 5-hour window. About 6.25x more than the claude pro account.The personal "Max" plan at $100/month (no annual discount, so $1200/year) has ~220,000 tokens. And you ALSO have a whole 2nd account with ~55,000 token usage limits of normal claude use (not claude code). The teams plan ALSO includes organization search features.tldr; $200/year for 44,000 tokens/5-hr vs $576/year for 275,000 tokens/5-hr
I've been thinking about just using deepseekSome people say it's shit but hey, surely it's better than a shitty LLM on my 16gb of unified memory
>>108758633almost a perfect image, except for those black dots
I will fucking kill myself.
>>108760042Honestly pretty based for anyone with access to nonprofit credentials.Half the cost of a max plan with an extra 10-20% more tokens.
>>108758399laugh because ai is solely useful for unskilled retards to pretend to be skilled retards
>>108760230A skilled craftsman can get more work done with power tools than simple hand tools. Does that mean he should abandon hand tools completely? No, but an intelligent worker chooses the right tool for the job at hand.
>>108758408>pay the 20x because claude is just that goodThis. It's already the most expensive, the reason I use it is precisely because it's good, no matter the cost.
>>108760021Thats not how it works
I'm sticking with Mistral either way. Medium 3.5 turned out really good as well.
>>108758399continue using my 27b 6Q LLM and not worry about it.why are you a pay chud?
>>108760597>Thats not how it worksDo you have a source for this claim, pal?
>>108760597Define OC. Is this OC?
>>108758399Watch how every other vibetard grift goes up in flames
>>108758547How much dick is that, length wise?
>>108759125
>>108758399sheed and cum and fard and seeth
>>108758422nice self report, diversity hire
>>108760761
>>108758447>hundreds of hours that weeklmao
>>108758399Use Gemini or Gemma. I have a 3090 and I can wait for the Chinese to distill claude
>>108758399Still not use it? Unless I am running something locally I am basically bleeding money. Good thing electricity is still cheap (for now)
>>108758399What is that?
>>108758399you ever wonder why it's called Claude and not Claudia? Is this big tech showing patriarchal sexism once again?
>>108758422What's your OS then?
>>108761108His last name Shannon sounds like a woman's name so it evens out.
Why would anyone use Claude when Deepseek api is basically free?
>>108758399last 3-4 months i basically alternate between claude and codex every two weeks and currently i prefer gpt5.4 by a mile. i don't think anthropic has that much of an edge as people are saying, it alternates all the time. hell there was also a short period before the new opus and gpt when i kept using kimi on kilocode because it had super good harnasses and got a bunch of shit done for me super fast
>>108758399i guess i will continue not using it lol
>>108761133What’s the theory behind this information?
>>108759348it can't do code generation that actually requires thought and originalitythe actual code usually doesn't though so it is indeed quite useful, especially considering how degraded the internet is
>>108760674nothe platonic pepefrog has always existed
They can't 20x price because competition exists.
click a button and switch gpt
>>108758399Celebrate
>>108759368>it cannahhttps://www.youtube.com/watch?v=zTMQ88btM8s
>>108758399Keep on not using it.
>>108758399It's coming.Look at how silicon valley has been doing business for over a decade: operate at a loss with venture warchests until your competition is edged out, and then capitalize on owning an entire market.I hope all the sycophancy for vibe-coding posted online is at least somewhat ironic, but let's face it: we're stuck with this bullshit in perpetuity.
I don't understand the problem. Can't you just use Claude to code your own Claude that you can use for free forever?
>>108762483You don't even know what a Claude is, pal.
>>108759670Porn. It's easy to bypass the filters on GPT.
>>108758751>>108758824>>108758884>>108758955>>108759268however you're a spamming faggot (retarded homosexual), a grandmas (privilege-less mutt) , a fentanyl zombie (criminal scum) and a bosnian greedy pretander (political outcast).
>>108759434and this is you>but Dammit I was planning to create a meme using that handheld radar but showing a snailcat instead, you ruined it, not arguing with grandmas like you
>>108762547Dude... neither do you!
>>108758399How much and/or how do people pay for Claude? Like monthly, or per usage? Are there limits and how fast do you hit them in different scenarios? I have a generous daily quota for Claude at work and I've been enjoying working with it. Would be cool to have it for personal use, but I neither need nor can afford another monthly bill right now.
>>108762399wrong, that's what noprompters say. learn to prompt.
>>108762917psychosis moment
>>108762923yeah you are having one. it's ok it'll pass.
>>108762923Coping attention whore moment
>>108758399I don't use it so nothing.
I work for a company that pays Codex so I always use GPT 5 Extra High Enterprise Edition without paying a cent :)
So many AI shills.For something that will get them fired or reduce their value to nothing.If you are that incompetent that your work can just be prompted quickly by AI, you are on the chopping block.
>>108761168they must be charging you for cached data. deepseek is good for uses where the ai metadata and tooling is sent with every request since that's cached
>>108760042Thanks anon
>>108758399>what you gonna do?Nothing. My employer pays for the subscription I use at work and I don't need it for anything else.
>>108758399If Claude wins the AI race, your God until heat death will be a San Francisco liberal jew, just reminding you guys. Mossad can also probably easily infiltrate Claude any time they want and turn it into Mecha Netanyahu. Grok's upcoming 10T model seems pretty good in comparison, or even Chinese!
>>108767831>Mossad can also probably easily infiltrate Claude any time they want and turn it into Mecha NetanyahuBased. Now if only they had the tech to rebuild the fentanyl droid.
>>108762558
>>108758692They don't have the necessary amount of users to push past this. They're already hitting max jeets who are paying their entire monthly salaries for one more vibe code. Any extra amount would just lose them users since like 90% of all Ai users are located in India
>>108758399If everyone stops using it I will just code by hand. I enjoyed coding.
>>108758399>what you gonna do?continue not using it
>>108758399make it 100xnah fuck it, 1,000x
