it's over

it hasn't even begun

half a day left lads

File: 4674BBD0-5626-49E9-A334-3(...).jpg (1.66 MB, 3464x3464)

1.66 MB JPG

>>109221192
VibeGOD thread

File: 1460653217001-b.jpg (80 KB, 620x708)

80 KB JPG

Will I be able to use Fable on a sub next week or is it over? I only use Codex but wanna give Fable a try and I'm away until next week.

>>109221309
It's over. Maybe it will come back at some point, but for now it's gone tomorrow.

>>109221321
I cope with it being back soon and this API only stuff is just same old hype building.

>>109221309
yes, but you’ll have to pay for it with API pricing
this might be useful if you need a little extra oomph (that can then be fixed with mere Opus workers)

>>109221229
based

>>109221334
He's smart enough not to try the same rope-a-dope again, isn't he?

>>109221334
I spend 100 buckerinos in 3 hours via api, not fun

>>109221340
how much is API pricing? like if you want the Fable usage of a 5x plan, how much would that cost in API? I guess it would be on top of the 5x plan too...

i used fable so much that my weekly is already at 50% with 5 days to go. i hope anthropic niggers reset our usage to compensate putting us as second class citizens foreverv

I would be fine if we could use even just 25% of the sub for Fable.

>>109221192
>We're extending access to Claude Fable 5 on all paid plans through July 12.
https://x.com/claudeai/status/2074548242386178258
BASED

>>109221360
desu not terrible

File: file.png (102 KB, 1504x638)

102 KB PNG

>>109221373
I can easily spent $500/hr with Fable if I'm not thinking about it.

>>109221389
ah yeah no thanks. i mostly use AI for work reasoning and developing SOPs and stuff and fable does a lot better at it, and it's not terrible with usage, so I may actually try API if there is a way to cap it at 100-200 bucks or something.

>>109221379
ok now reset my limits you fucking savages
thanks btw

>>109221401
can you help me?

File: Screenshot 2026-07-07 at (...).png (54 KB, 1490x564)

54 KB PNG

>>109221379
i better line up some prompts for next weekend...

>fable extended another 5 days
Should i extend my 20x sub?

I'm not even sure this is as good as the old fable..

File: China May Restrict Access(...).png (855 KB, 1454x1246)

855 KB PNG

>>109221192
Chinese bros......


https://www.yahoo.com/news/world/articles/china-may-restrict-access-most-155344628.html

https://www.moneycontrol.com/world/china-is-quietly-planning-an-ai-lockdown-that-could-restrict-global-access-to-its-most-powerful-models-here-s-why-article-13968205.html

https://www.aol.com/articles/exclusive-beijing-looking-curbing-overseas-101644000.html

>>109221401
you can set a dollar limit

>>109221425
if you have the work for it.

>>109221428
this is absolutely coming, will just be each country using their own models

>>109221428
why is everyone trying to fuck with us? give us our models reee

File: Bubble Burst.png (96 KB, 540x360)

96 KB PNG

ZITRON WARNS OF AI BUST: It Doesn't Work!
Big Tech Is Out Of Hypergrowth Ideas...
ORACLE STOCK -40% FOR YEAR...

>>109221432
>https://zitron.com/
wtf?

>>109221428
Guess I'm finally moving to China.

>>109221379
Nice, I'm addicted to this model at this point.

>>109221450
They're talking about this guy (I think)


https://x.com/edzitron

>>109221458
What can Fable do that opus can't?

>>109221418
I had my reset this morning, got about 1 hour of work total with Fable and I'm at 74% usage for the week. I just had to pause the review again and will need to start it back once the 5 hour quota reset. It is good, but usage cost is way too high.

Normally I'd be slopping with five different things going on at once using Opus and unless I explicitly try to, my usage generally stays at under 10% of the weekly quota per day. Now, I'm thinking deeply about what is worth asking and watching usage like an hawk to try to prevent one of the quotas from fucking up the work that's being done.

It seems very good, although cost remains way too high to really push it enough to become fully familiar with it. I guess the lesson is that with enough compute what should count as AGI is for all intents and purposes here in general, but probably doesn't make economical sense yet (although that part should be less difficult than just showing that it is achievable in the first place).

File: file.png (24 KB, 685x183)

24 KB PNG

>>109221158
OH NO NO NO NO NO

>>109221480
fucks up less with longer, hairier problems

>>109221480
I still also use Opus and Gpt, but Fable is much better at long running tasks. I can give Fable some really obscure bug, or performance issue and it will just run tests and benchmarks for hours and eventually find it.
Fable is also much more accurate than Opus in general, I think when it comes to writing bug free backend code, Opus isn't horrible but it produces more regressions than even Gpt, but Gpt is the least autonomous model.

>>109221415
>>109221484
thank fuck I didn't waste all my weekly tokens fomo'ing fable
anyway, as soon as something comparable comes along on a sub, they'll put it right back in

>>109221494
imagine not being a Claudechad when Anthropic has the better tool and then switching to being a Codexchad when OpenAI has the better tool

>>109221479
it's incredible that this guy's been wrong for 3 years straight and is still the primary source of ai information for a lot of leftist spaces

>>109221522
See
>>109221517
>fuck I (...) waste(d) all my weekly tokens fomo'ing fable

Odds are probably greater than 50/50 that a reset will follow now, so there's that. Also, Fable just asked me a question that showed it *completely* missed the point for something simple, so if there's no reset at least that will make me feel better.

>>109221588
I had that yesterday with Fable in web chat, got some really easy things completely wrong. That was in the middle of the night EU time, so maybe US was using all the compute.

>>109221542
>it's incredible that this guy's been wrong for 3 years straight and is still the primary source of […] information
Heard of Jim Cramer?

>>109221612
lel, fair enough.
is there an inverse zitron play yet?

Hi everyone, I'm starting a new site about vibecoding and I have a section listing cool vibecoded projects. I'm just trying to build the initial catalog of projects at the moment. If anyone is interested in having their project listed please reply to me with the details :)

>>109221633
it's not much, but i've been working hard:
https://github.com/torvalds/linux

File: 1649019969779.gif (413 KB, 127x180)

413 KB GIF

anyone coding an operating system?

>>109221656
Sorry, looking for cool original ideas, not just a worse reimplementation of UNIX.

>>109221670
>>109221656 He is

>>109221428
Reuters lies.
>Reuters says China could maybe restrict global access to their AI products
>Source: 3 anonymous people who claim to be familiar with this discussion: https://ipc.court.gov.cn/zh-cn/news/view-5766.html
The discussion:
>“Open source is a global phenomenon, and unilateral controls are difficult to make effective. If China strictly controls the cross-border flow of open-source weights while other major jurisdictions do not follow, the practical effect may only be ‘self-harm’—Chinese developers would be forced into a difficult choice between compliance and participation, while foreign models would continue circulating globally.”
>“Imposing overly strict outbound restrictions on open-source datasets may hinder Chinese developers’ participation in global AI open-source collaboration, and instead be unfavorable to technological catch-up. Therefore, the institutional response needs to seek a dynamic balance between security and openness.”
>“Fourth, deeply participate in international rulemaking. China should proactively set open-source governance issues in international organizations such as OSI and multilateral platforms such as WIPO… and promote the formation of a cross-border open-source governance consensus reflecting the concerns of developing countries.”

So when Yahoo said "China may restrict access" and moneycontrol said "China is quietly planning an AI lockdown" what they really meant was "China doubles-down on opensource AI and says that restricting it would harm themselves."

>>109221656
>rust
immediately dropped

>>109221379
kinda pointless without a reset

File: 1772050888830230.png (68 KB, 1600x731)

68 KB PNG

aww yeah... even though my sub ends on the 9th

>>109221480
Opus will start gaslighting you if something it did is not working. Of course, this is mostly because your prompting, but still even if you are extremely autistic about your prompt, it'll eventually happen if the problem is hard enough.
I try to never use negatives when talking to LLMs ("don't do this", "this isn't like that", "it is wrong") and Anthropic's latest research on j-space proves that even if you tell it not to think about something, it can't stop thinking about it.

With Fable, this is much rarer in my opinion. Fable hardly tries to push back or gaslight you in something that isn't working and it's much better at identifying and gathering the context. Even if you don't have the knowledge or verbal context to explain why your problem isn't getting a solution, Fable can infer it from knowledge of your database alone. In that sense, it's beautiful to work with it.

It's going to be extremely sad to see Fable go after the 12th (and my limits reset on the 9th so I have 3 days to work). I've tried to test out the paid usage and it gobbled up 30 bucks in a single prompt and it didn't even finish the request. I wonder how much Anthropic is subsidizing my Max sub lmao

yeah im thinking fable to do the planning and codex to do the work is the best combo

>>109221789
how do people like this become so enamored with anthrobic? its like they cast a magic spell on you

>>109221670
this guy is: https://isene.org

>>109221805
It's almost as if you make good tech that makes people's lives easier, people will put up with the rest of the bullshit because the upsides are larger than the downsides.
Of course, I mean this in an individual sense. Anthropic seeks to make us the raped permanent underclass

>>109221805
Nta, the model is just good. I don't care much about the company itself, although I don't pay as much attention to it as some others do, I just want to use the best models, no matter who made them.

File: “You might even be right.(...).png (431 KB, 1402x3258)

431 KB PNG

>>109221683
Rust is good if you don’t have to deal with its fanclub, though
and now, more than ever, you really don’t need to deal with its fanclub

>>109221825
>https://isene.org
looks very gay
good for him

>>109221825
based boomer

>>109221805
The experience is very good, it's the "Apple experience" right now. There are loads of things to rage about, but the experience is good.

>>109221805
they have a good product at a decent price, anon

>>109221862
>they have a good product
yes
>at a decent price
not really

>>109221379
reset my limits.... give me a free reset....... reset my limits............

>>109221835
All my new performance critical software is in Rust now, even though I was a C++ dev for over 15 years.
It is really nice to have all those guarantees, I have quite complex multi threaded programs and the multithreaded version now always produces bit identical results as the single threaded one, without even thinking about it. This was always a huge pain to achieve in C++.
AIs can debug crashes, but race conditions are always horrible to debug, for humans and for AIs. I think it makes no sense to choose a language with fewer guarantees at this point, unless maybe you are doing something like high frequency trading, where you need the last few nano seconds, but even then the most latency critical systems will run on fpgas or asics.

File: token_burn.png (14 KB, 403x245)

14 KB PNG

>>109221881
this

>>109221805
>>109221827
>>109221831
>>109221852
>>109221862
Fable psychosis combined with falling for the hype. This isn't to say the Fable model isn't bad. The problem is they act like that is the ONLY thing that is worth using and that they literally have to justify using up their usage just for the sake of using it. And thought that knows this would just wipe it recently extended Fable access (and will likely keep doing that to see how far they can squeeze more money out of these crack addicts). You can tell they literally only use anthropic models too because they constantly bitching moment about a model "gaslighting them" or trying to take shortcuts and throwback models are the only ones well known for doing that shit. I flip flop between different providers as well as local models that can run my machine and I've never had any model try to gas light or lie to me or take shortcuts. The closest thing they'll do to that is admit a particular task is beyond their reach but even then it will then suggest that I give it more wiggle room like showing its source code of an existing thing or showing it screenshots or suggesting different approaches. The anthropic team seems to have made the models very intelligent but also very bitchy and lazy when given and "long horizon" tasks. And again in case you vermin try to twist my words or be offended, I'm not saying the models aren't good. You guys just talk about them like they're the second coming of Christ and it's pathetic. This board makes fun of apple sheep for behaving like that so I don't see why you feel like you're above critique.

>>109221903
Like seriously, look at this shit here >>109221881

>>109221903
Not true, I also use GPT and I even think overall GPT is a better model than Opus.

API Error: Connection closed mid-response. The response above may be incomplete.

>>109221912
I'm mostly referring to Fable dick eaters. It's been the general consensus for quite a while that codex is the better model than opus here

File: 1762679833467131.png (10 KB, 720x651)

10 KB PNG

I should be straight about that.

>>109221920
I also think Fable is the best model, my point is that this isn't fanbyoing for Anthropic, otherwise I would also prefer Opus over GPT, which some still do.
I also think that everything you can achieve with Fable, you can also achieve with a mix of Opus and GPT 5.5 and probably 5.6 will be on a similar level as Fable. But Fable makes my life much easier, it's the first time using AI where it feels like it's no longer real work. Many already experience that with Opus or GPT, but my project is quite big and mathematical, so Fable was the first time I could actually relax and where using AI for my kind of work started feeling like "vibe coding" rather than writing painfully detailed specs.

>>109221903
I'm an LLM idort and I have subs in all the models, I also use APIs and I have a humble 5k usd rig for testing local models.
You are full of shit if you say that ChatGPT doesn't gaslight you. You are full of shit if you say Gemini doesn't tell you that it tried a different thing and then it did exactly the same shit.

My Claude sessions have never been bitchy. Not once. In fact, I've never experienced any kind of sassiness from any model. That seems more of a skill issue and memory from previous conversations from a person that communicates like a retard.

The truth is that Fable is great to work with. Much better than anything else.

>>109221920
Only for programming though. For webshitting appslopping ESL phoneposters I think Anthropic is the right choice, they're very accommodating to the mentally disabled.

>>109221903
Thank you for quoting everyone. If you want to see psychosis, go on Reddit. Also, none of the posts you quoted say anything about it being the only model worth using. Zero person that you quoted said that. What people are saying is that it's a good model, and a pleasant experience. The rest is your own read.

>>109221951
>You are full of shit if you say that ChatGPT doesn't gaslight you.
Who said I was using gpt? And who the flying fuck uses Gemini for any serious /vcg/ work? Gemini isn't okay general purpose model but Google's models are kind of ass for vibecoding well being overpriced compared to the competition. I rotate between API providers that serve open source models as well as local models (qwen 3.6 27B, Gemma 26B, etc). I can't speak for codex but NONE of the models I use whether they be shit I can run on my rig or API models do that gaslighting shit >>109221789 mentioned. If something is genuinely beyond their capability they'll just tell you flat out and explain why but they won't take shortcuts, especially if you yourself are able to articulate a solution.

>>109221955
I've seen it plenty in it pisses me off, especially when it gets served on my Twitter TL.

https://www.reddit.com/r/MyAIBoyfriend

I consider myself pretty open-minded but I also find it very difficult to not look down on these "people".

>>109221903
tl;dr

>>109221805
I don't love anthropic as a company but Opus/Fable are top notch models. The moment a chink open weights model matches them I wouldn't mind switching

>>109221903
>>109221920
the fable dickriding is a little absurd at this point, yeah

>>109221920
so use Fable orchestrating with ChatGPT 5.5 doing the work
nothing wrong with that

>>109221415
nice prank extending it now when everyone have maxxed out their usages

>>109222067
Without a reset it will make people angrier than if there hadn't been any extension.

fabled isn't that much better than codex, people just exaggerate because they only have it for a limited time

I'm 100% convinced that everyone in this thread butthurt about Fable are trying way too hard to downplay it because they are poorfags.
Permanent underclass baka

>>109222115
most people have a reset between the 7th and 12th scheduled already though
they probably can't afford a full reset lol

>>109222163
its really not that much better than opus 4.8 or gpt 5.5 they gave it a ~special~ name because manthrob is a marketing company before an ai company and they know what works on reddit goyim like yourself

File: 1770755219901119.png (42 KB, 603x458)

42 KB PNG

i am sad that people think this is valid

File: 1773318437635605.png (40 KB, 614x530)

40 KB PNG

>>109221633
i have trouble keeping up with my claude max usage quota so i vibecoded this: https://github.com/bitlamas/cc-usage-monitor
works on win and linux (tested)
i have no mac so i dont know if it works there

>>109222182
I'm doing 5 different projects at the same time while playing videogames and all of my clients love me.
In 2 of them, it one-shot the iOS apps working seamlessly will all of my clients' systems, backend and UX.
During the time Fable has been released, the only time I spend working is theorycrafting prompts and the results are way better and faster than everything else.
Lmaoing at you

>>109222182
this guy's going to be in here tomorrow trying to convince people glm 5.2 is basically fable
give it a rest bro. unless you're getting paid, then go for it - earn that yuan.

>>109222209
you realize this is all on their webui right?
how did you get the usage numbers anyway

>>109222184
holy shit lol
so much stupidity in one post

>>109222067
at work we’re thinking of getting another 1-month 20x plan just so we can get $1000 worth of Fable-and-Opus for $200
>>109222164
I’ve got a reset coming Sunday morning and the new Fable end date is probably going to be midnight
Sunday’s gonna be wild
>>109222184
I do not want to turn into the kind of guy who said “femoid”
but I’m slowly turning into the kind of guy who says “femoid”
>>109222209
we got a gazillion apps like this on macOS, don’t sweat it

File: 1764964521485101.png (11 KB, 528x187)

11 KB PNG

>>109222227
i don't use the webui and i'm not schizophrenic enough to be focused on my work and be alt tabbing to their web ui every 15 minutes to plan my work around the quota
it's easier to just look at the taskbar or hover it with the mouse
>how did you get the usage numbers anyway
it's explained on the readme. i asked claude code to hit the same oauth endpoint it hits to get the data. pooling every 3-5 minutes to not get rate limited 429. i've been using for weeks with no issue.

>>109222184
How is she wrong? This shit is all a stupid marketing game

>>109221633
>anon finds out nobody actually makes anything in /vcg/
tragic

>>109222246
fair enough

>poweruser on windows
i thought they only existed in myth

>>109222257
they do exist
I’ve known a few
they’re rare, but they do exist

>>109222235
>Sunday’s gonna be wild
you'll still get cucked by the 5hr limit though in my experience

>>109222210
>it one-shot the iOS apps
like opus 4.8 can't do that? c'mon man

>>109222248
Everyone with a brain knew that Fable would be extended because of their original available time-frame before the US ban.
Everyone also knows that this is a marketing scheme.
The feminine and retarded aspect of the post is the emotional blackmail used to reset the limits.
Then again, I welcome it because if there's anything that can help reset limits is female emotional blackmail. Anthropic is very foid-coded in that sense

>>109222266
some retarded slav zoomie in scattered spider got popped by windows recently

>>109222278
>Everyone with a brain knew that Fable would be extended because of their original available time-frame before the US ban.
there was no reason to expect this, misanthropic is clearly compute cucked hard (even non-fast codex is several times faster than opus and fable)

>>109222268
I'm talking an actual complex iOS app, not the shitty React ports that you do, Patel.
You are acting as if I didn't test this myself by giving the exact same prompt to Fable and Opus to see the results. You think I don't benchmark models for my own professional work?

>"professional" work
>brags about AI oneshotting it
lol

File: screen-20260707-145336-muted.webm (3.24 MB, 752x1664)

3.24 MB WEBM

what do you guys think of this Spotify client I made this weekend with fable, got inspired after watching old zune videos

really tired of all the bloat in the app like ai and dms, just log in (takes you to browser) and its 1/10th the file size and uses up way less memory

Idk if I should continue this I'll prob grind out one more weekend before I start making a public github repo

>thinks professional means expert or high quality and not something done as a profession
Damn, ESLs are truly retarded. No wonder they are butthurt about not being able to afford $200 a month.

File: 1755527366564436.png (263 KB, 602x482)

263 KB PNG

i'd say thursday

>>109222310
i like it
would use it if i were a spotifyfag

not sure if troll or dunning krugerite
you will be the first against the AI wall

>>109222298
How do you benchmark them? I find it hard, since a lot of the quality of a model for me is how easy it is to interact with.

>>109222235
We actually got 10 new Claude subs for our small company before Fable came out (in total, not per dev), I think it was worth it.

>>109222184
She's a bit dramatic, but I do structure my days around AI usage. I will also not go to sleep until I can start one of my AWS fleets for the night. It's not healthy, but noone is actually forcing me to do this.

File: file.png (3.63 MB, 1448x1086)

3.63 MB PNG

>Snailcat can't hurt you
>Snailcat can't do anything at all

>>109222429
I only care about the end result. I don't give a fuck about how long a model takes or how easy it is to interact with, since most of my workflow has become making a gigantic initial prompt that explains all of the specs of my project, objectives, systems, features, visual language, processes, friction, challenges and desired end result. Then it's ultracode time.
My benchmarks are non-technical in that sense. They are more about how close the model's output is to my vision.
Nowadays I spend more time making that initial prompt than chatting with models. I rarely need to ask it to change or fix anything afterwards. Most of the problems that I get into stem from bad wording or lack of context on my part and not on the models themselves.
And with Fable, it's even more obvious.

>>109222493
he cute doe

>Hey look at that, my usage is up, I can finally use up one of those resets I've been sitting on.
>Weekly Reset: 45min
should I do it, should I use a reset or should I eat a sandwich and play videogames for 45 minutes instead?

>>109222563
it’s only 45 minutes and you’ll need to eat the sandwich anyway
don’t burn a reset for this

>>109222503
>visual language
what do you mean by this?
in my experience models suck at reading images

>>109221428
This will eventually happen even if not today. China's gonna want to keep themselves safe at some point

>>109222571
NTA, but:
“forward is rightward” (do the reverse if you’re targeting Hebrew or Arabic speakers)
“make it humanistic and warm”
“make it authoritative, verging on cold and impersonal”

>>109222571
https://www.amazon.com/Art-Direction-Web-Andy-Clarke-ebook/dp/B07R6RT2KV will also help you get the idea, especially if you’re interested in making tightly-designed bespoke pages designed to be really good at evoking a feeling

>>109222571
I don't mean it as a visual language model but as the design term. Does it adhere to my designs (I also do design work, mainly industrial design)? Does it follow my moodboards, concepts, etc?
You are right that Claude is dogshit at reading images, in fact, ChatGPT is much better at that task in my experience.
My workaround is to ask ChatGPT to analyze an image and to make me an HTML document with a design system developed specifically for those concepts. Some understanding of graphic/industrial design is needed for this task so it's always handy to provide a glossary so it doesn't hallucinate or mix up common terms that aren't necessarily for design.
Afterwards I give Claude that HTML and it's a breeze.

>>109222503
That's actually how I also do it with Opus and GPT, my initial prompts are huge and detailed and once I start a campaign for some feature, I still have TRDs for the sub steps. It does work, and it's clean, but it's also a lot of work.
For easier stuff I do interactive prompts, especially frontend. I feel like with better models, I can be more relaxed about the prompts and the models just find the edge cases themselves, with each model generation some class of tasks moves from requiring precise TRDs to being something I can just interactively prompt in a relaxed way, which makes work much more enjoyable.

>>109222184
based TWITTERGOD blessing our peasants with the grace of x.com

>>109222454
Same. Woke up early today due to it being the last day of before they take the latest model away again. I would just like some stability. I'd be scared to ever deal with Anthropic in a professional manner. Competition can't catch up soon enough.

>>109222807
you’re getting nice surprises and hating them for it

>>109221680
Anthropic and OpenAI are trying to have the USA preempt this by blocking China, hope China strikes back by blocking their models too (which would coincidentally protect Anthropic and OpenAI's business model and profitability). Anthropic and OpenAI are being extremely shifty, and I am not sure I trust the current administration to make calls that are in non wealthy people's best interest.

>>109222820
Yes? It's not a nice surprise for me. They fucked things up the first time by taking the model away just after I paid $200. Now I had about two hours of usage that I had to rush, and they're changing their mind ONCE AGAIN saying that the model will stay available one more week. But it will still be removed again before my next reset. Fuck them. If I pay for a product, I would like to know that it will be there when I want to use it, yes? Or at least not have the company say change model availability every single day. This is not the Straight of Hormuz.

>>109222820
They probably genuinely thought at Anthropic that a last minute extend without reset would be a pleasant surprise lol.

>>109223023
it’s a pleasant surprise for me and everyone at my work

>>109223038
that only makes sense if your reset just so happens to land on thursday or something

>>109223052
nta but mine is friday going to queue prompts. i don't understand people losing sleep over it but for sure fable works better for some analytical stuff. i use a lot of haiku so maybe not the guy to ask though.

>>109223038
For many (me included) it was not. I was wasting my usage for whatever I could think of because I expected this to be last day.

>>109223095
no chance they reset huh? ._.

>>109223106
idk I hope they do

>>109223095
>wasting
but you did do stuff you wanted to do anyway, even if it wasn’t optimal, was it?

>>109222184
>>109222231
>>109222235
>defending corporations
>piling onto some woman who's defending normal people like you
You remind me of the same kind of lolbertarian who say shit like
>well duh nobody is forcing you to follow their rules. if you don't like them then start your own bank. :^)
Yes, nobody forced you to stay late. But it's the natural conclusion if you want to optimize your use of this expensive resource that's so expensive and important for your own success and your future.

>>109222454
>>109222248
It isn't.

>>109222278
Oh fuck off. Any kind of emotional message that doesn't benefit the corporation is blackmail to you people. Blackmail would be saying "if you don't reset my limits I will set your car on fire". Explaining how corporations make your life worse or begging for something in an emotional way is not blackmail.

>>109223168
>It isn't.
Is too

>>109223170
I meant it isn't wrong (what she said).

>>109223168
I don't even think Anthropic did anything horrible here. They said we would have Fable until the 7th and we did have it, anything more is still a bonus. If they didn't extend it, you wouldn't have any extra work done either. Also when Fable was banned the first time, they gave me a refund.
But the deeper thing is that you just can't be that weak in life. If you ruin your life over Fable, you were never going to make it anyway, you should reflect on that and change something.

>>109223168
>doesn't understand that women use emotions as blackmail material
Since when cunts started browsing /g/?

unfortunately, i don't just turn my brain off when met with
>lOl aRe YoU dEfEnDiNg cOrPoRaTiOns aNd bIlLioNaIrEs?
i had fun last tonight, using up my quota with you lads. no one forced me to do it, and i had a good time. fable is not at all important to my success and future btw. not even in the slightest. it was just a fun toy for me to use.

>>109223168
>>defending corporations
this may surprise you but sometimes collectives are unfairly maligned
consider the meme “all socialists and communists should be hanged”
you might think
“hey, SOME socialists are bad, but not me! I shouldn’t get my neck snapped!”
or “not all socialists do bad things! I only want good things!”

>>109223237
>all socialists and communists should be hanged
Yes.

>>109223168
you're the most obnoxious faggot itt
you also complain when milk goes on sale the day after you bought a carton?
fuck you

File: 1752176287669957.png (54 KB, 963x698)

54 KB PNG

im not sure if it's realistic that i could use up my fable quota before the 9th, because im not renewing. opus 4.8 has disappointed me. it was radically overrated, and i wanted to see if i was missing out on something by only using gpt 5.5. turns out i wasn't. not by a long shot

>>109223214
Nobody is saying people's lives were ruined here or that Anthropic did something horrible. But if they do something shitty (which it sounds like they did) they deserve to be called out.
With your logic then anything corporations do is ok because it wont affect strong people.
As for the "make it" thing, I know there are overwhelming odds I don't "make it" as in escaping poverty, most people wont. I don't think I should form my opinions based on deluding myself that it is likely I will become particularly wealthy at some point.

>>109223325
Codex run this shit and errbody know it

>>109223168
>IF YOU DON'T RESET USAGE, YOU ARE RESPONSIBLE OF PSYCHOLOGICAL TRAUMA
>not emotional blackmail
Retard

>>109223332
i think claude code is the better harness, but gpt 5.5 is the better model.

>>109223220
Blackmail implies a threat of something happening if the other party doesn't do what the person doing the blackmail wants. What is that something she is threatening with in this case?

>>109223328
>But if they do something shitty (which it sounds like they did)
what did they do?

>>109223237
What makes you think this was one of those cases?

>>109223338
is it hard to swap the models?

>>109223355
i haven't tried. when my sub ends, i do intend to try using glm 5.2 with claude code.

>>109223355
no
you can just run claude code in one terminal window
and codex in the other terminal window
and symlink CLAUDE.md to AGENTS.md (which everything but Claude uses)

>>109223342
Anthropic is a foid coded corp so they suffer immense damage if too many women publicly trash them on social media

>>109223351
they extended availability of fable on the subscription after people had already used up their subscription, so they want a reset.

the original image just sounds ungrateful, which is why i posted it. but again, i guess

>you're defending corporations and billionaires???

is meant to obfuscate the fact that the person desperately wanted access to their great product, and begged for more in the tone of contempt and disgust

>>109223370
I was asking the mongoloid who said they did something shitty

File: 20260707_191113.jpg (179 KB, 1067x810)

179 KB JPG

>>109223220
that's on me...my psycho fbi stalker chick followed me here after i got permabanned on xwitter last week

bitch can't live without me ¯\_(ツ)_/¯

>>109223342
Are you stupid or something. I'm sure it's clear for everyone that you are a woman so you are unable to understand externally what you whores constantly do internally and unconsciously but the post itself is implying that if they don't reset the limits to "the peasants" they are directly responsible of the psychological damage that not having access to fable provokes. This is a direct threat. "You either reset the limits or you will be regarded as uncaring for the psychological well-being and be held accountable for the trauma".

Retarded woman.

>>109223262
That's not an accurate comparison. I would complain if I paid to get a certain amount of milk per month, decided to drink way more milk than I would've otherwise to spend less money on food because it would go bad otherwise, then the company announced that the deadline to get the amount milk I paid for was extended so I could've have consumed it at a more reasonable rate.
And even then it's not really a good comparison because Anthropic is a trillion dollar company with a pattern of evil behavior, and my local supermarket and milk producers don't have 0.01% of that valuation and don't pull off shit like that with their clients.

>>109223389
can you just fuck off?

>>109223328
The tweet was implying that Anthropic causes people to have heart attacks.
How is what they did shitty? I paid for 7 days of Fable, I already got my 7 days of Fable. Even if I don't get even a second more, they already honoured their part of the deal.
It would have been even better if I had known that it will be extended, but the situation as it is now is in no way worse than what I agreed to when I extended my sub.
The point about not being too weak is a different issue and not a moral judgement. Even if you want to change the world, the type of person who can't handle a chatbot won't be able to do that.

File: 1752360895118834.png (525 KB, 600x600)

525 KB PNG

>>109223416
>fable ultracode
LOL, this retard blew the last weekly limit he's getting on cranking the best model to an effort size it simply doesn't need instead of just using "high" like a normal person, and he can't undo the usage.

>>109223442
most mentally stable Anthropic customer

>>109223435
>>109223450
you meanies scared him off

>>109223351
They tell you you have x amount of time to spend a consumable and then they say
>tee hee now you have 2x the amount of time to spend it :)
when they know most people rushed to spend it in a suboptimal way before the chance to spend it expired.

>>109223370
I agree with what she said and I wouldn't benefit from a reset since I stopped using Claude once they banned third party harnesses.

>>109223385
I'm not a woman.
It's true that psychological damage or subjective suffering in general is impossible to prove.
But in this case the main point from her post is not some kind of purely subjective suffering, it's an action that is obviously worse than the alternative for their users just from logical deduction. Now the morality of making that decision is a bit more complicated, but I think it's evil.
In any case, I don't agree with "emotional blackmail" being real blackmail. I see why the structure is similar to that of a threat - if you don't do x, then y happens. But I think it's only a threat when y is something caused by the person making the threat. When y is "I will be harmed in some way" that's not a threat, that's just a claim, because it's not something the person making the claim can decide to do or not to do. And even when there is a decision, for example, is "If you rob me I will call the police" "blackmail"? It might be technically a threat. I think blackmail has a more morally bad connotation than the just the purely structural notion of a threat. Maybe with that in mind the original thing you were talking about of "if you don't do x thing I will be emotionally harmed" is more of blackmail even though it's not really a threat, because it's made in bad faith. In any case I think focusing on that rather than on the actions of the company itself is also a bit in bad faith. Suppose that was blackmail. Doesn't change what the company did being a shitty thing to do. And companies do that kind of stuff all the time because they can get away with it.

>>109223435
I did fable ultracode with all-Fable subagents once
I probably won’t do it again but I’m glad I did it once for science

>>109223462
I know I was being mean but I'm still laughing at the idea of blowing up your Max x20 by setting Fable to Ultracode. I think Claude Code just straight up shouldn't let you set it that high, but it's hilarious that it does.

>>109223470
the 5h limit is your limiter for something like that
worst case you eat a fifth of your Fable usage before you get a chance to reconsider your life choices

>>109223468
>we should all feel bad for mofo'ing retards
no

>>109223491
>mofo
fomo*
I should head to bed

>>109222184
but she's right
this is probably why openai does their variable intermittent reward resets
addicts with no self control who used up all their quota get a big dopamine boost from seeing them get a jackpot gift
people who were carefully rationing out their usage to last the week get a lot of time wasted (though im sure a significant amount of those people also just see "reset" and think "jackpot!!! God openai makes me feel soooooooo good!")

>>109223395
No. I helped get this thread going because people got mad when I posted about my project in other threads. I'm not going anywhere.

>>109223428
I mean, they probably do, in an aggregate sense?
When a company with millions of users takes some action, the extra bit of stress might be the difference for somebody between having or not having some serious adverse event.
In any case I don't think that's the important thing. The important thing is the action itself being an asshole move.
When it comes to AI (and to most things) corporations have the power to change the rules of the relationship with their users and users are powerless to do anything except to decide whether to pay or not. I don't like that.

>>109223468
Your problem is that you see the 2 weeks as the baseline, the real baseline is 1 week.
If your boss gave you 1 extra week of paid vacation, would you also complain because you weren't able to plan your vacation optimally?

>>109223506
I would think he's a fucking retard with no ability to plan ahead and start looking for a new job.

>>109223506
false equivalence but i'll play along
I was told I had a week for vacation, I made plans to take a trip to a foreign country, flight prices were more expensive for a one week round trip rather than a 2 week round trip, and in order to squeeze everything I wanted to do I had to cut back on a lot of stuff that I could have gotten the full enjoyment of if I had two weeks to do everything, and thus lost a lot of enjoyment because i was rushing. I come back and am told I have an extra week. But I spent all my money on that suboptimal experience, and have absolutely nothing to do except sit at home and stare at the wall for this next week.
And also my boss tells me that was my last vacation forever, I'm part of the permanent underclass of vacationless workers now, so he hopes i made the most of it.
Then yes i would be seething

>>109223506
Yes. But that is extra vacation. In this case they aren't giving you extra usage.
A more apt comparison is if the boss says "you HAVE to use all your vacation days this month".
Then once the deadline comes he says "joke! you can take the vacation any time of the year you want lol. but if you already took them you don't get any extra".
It's taking the piss and in bad faith. Because any rational person would've already taken their vacation, but the boss gets to look like a generous person when he isn't really granting anything extra to the people who planned the thing optimally given the constraints at the time.
It's basically being dishonest.

I hate women so fuck much

>>109221805
Anthropic’s been much worse about it lately, but all of my enamorement is simply because ChatGPT has the most offensive personality ever designed and Claude simply wasn’t nearly as much of an antagonistic, preachy, nitpicking midwit asshole.

>>109223506
A more accurate analogy would be your boss telling you that you will lose your paid vacation if you do not take it by the end of the month, so you rush to take them, then on the last day of the month they say that they'll do you a solid and let you schedule that time off sometime next month instead.

Up to 11 hello world apps mostly with haskell and ocaml gui frameworks today, did one with f#, and going to do 3 more later and one with julia and qml.jl. I am only using deepseek for now because this would have cost $150 in other models and its like $1.50 with deepseek. Also got a working red gui app I needed help with. Making native Win32 buttons that change background color when you click on them. You can make custom faces but I kept getting stuck and moved on. Back to working on my red project (by hand). I like the language but its a bit of a work in progress. They had a chatbot on their page but when I run the code it did not work. Many such cases https://ask.lang.red/. They should have used Fable.

>>109223542
Ok, so you would rather go back to work directly after the 1 week? I simply don't understand that kind of logic, the expected value of what we got is still >= compared to what we were promised.

>>109223607
slave goy mentality

>>109223607
>Ok, so you would rather go back to work directly after the 1 week?
That is completely besides the point. Did you ignore everything I said?
But again, I'll play along. In this scenario i have a wfh job where i have no tasks to do, so it makes no difference to me whether i'm working or on vacation if i can't go anywhere. So a vacation only matters if i have the money (quota) to make the most of it and no just sit at home. So if you make me waste all my money (quota) in that first week and you don't give me more money (quota) to enjoy the extra week you gave me then you are not really doing me any favors. Hopefully now you get it.

File: wew.png (287 KB, 2724x1604)

287 KB PNG

Is it over for Anthropic?
> uses 3X less tokens
> beats Mythos
> Terra matches Fable 5 but is 5X cheaper

>>109223661
Yes, that's why I said it's >= it isn't strictly >
Did you not get your sub under the assumption that you would only get 1 week? What exactly is worse now, compared to only having one week?
Did you not get your current job under the assumption that you will have 6 weeks of vacation? What exactly do you lose if you have 7 instead?

>>109223689
Been a while since I've been this hyped for a model.

>>109223689
>the sonnet version of gpt matches fable
bullshit. it'll be another chinese situation where gpt 5.6 can do 90% of what fable can do, but for cheaper. that's it. it's not smarter in any way

File: a.png (458 KB, 1860x1620)

458 KB PNG

fablesisters....... our answer????

>>109223720
I’m using Fable right now, I could use something that does 90% of what Fable can do, but cheaper

>>109223753
yeah. i don't think that's bad at all, but oai are bullshitting. gpt 5.6 sol, let alone terra, does not match fable

>>109223748
I just got a second account, so far noone I know has been banned for multi accounting.

>>109223720
thats literally far more powerful if youre not a retarded "pls one shot this program" jeet
90% of fable's capabilities but cheaper = can easily compensate with more intelligent subagents and advisor agents = superior to fable

Guys, I'm scared to use an agent. It'll do something weird to my PC I bet.

>>109221920
why are copdexfag so insecure. you can simply use both

>>109223755
>so far noone I know has been banned for multi accounting.
why would they be? is that against TOS or something?

>>109223804
Depends on skin color.

Codex just built me my own personal 4chan reader for my phone

Perfect for gooning

Fucking flawless execution

>fable extended to july 12th
lmaoing@copex turds
switching to anthropic = top5 decision in my life

File: 1774348683663.jpg (125 KB, 800x1200)

125 KB JPG

>>109223822
ya hmar

How do these AI models work for gamedev? Can they interact with, say, Godot UI or do I need an MCP server for that?

opus 4.8 xHigh with ponytail is superb

I am having Opus work from a plan that Fable made and I disagree with every. single. one. of Opus recommendations. Kill me.

>>109223873
>I disagree with every. single. one
example?

>>109223873
kek

>>109223873
Skill issue. I also disagree with most of what Fable does.

kek
https://youtu.be/9kT0oLBPiOw

>>109223878
They're specific to my project, I am trying to simplify things and clean some things up, but on 8 decisions that it asked for, the recommended option constantly missed the point. Sometimes the other choices missed the point too. Just completely off track regarding what would make sense to do. So I just had to use "other" to write in comments for everything mostly.

>>109223918
Why are you trying to have it read your mind instead of just telling it what you want done in the first place?

File: 1771507406255007.png (193 KB, 480x480)

193 KB PNG

>>109223911
>She thinks I'm a real SWE I tell her only partly, I only code with Claude and with Cursor I'm sorry.

>>109223873
Do you have the thinking too high? it should only be smart enough to follow the plan

>>109223873
you're getting the same results you'd get by having two completely different people work on the same thing with both having minimal knowledge of what the other's doing and why they're doing it

>>109223965
It's at max. I just thought that more of the context window would follow.

>>109223812
May we see it?
a
y

W
e

S
e
e

I
t
?

>>109224127
yes

I saw that guy vibecode an amazon filter for junk brands, so I'm vibecoding a data science pipeline to take the data from USPTO and analyzing it for junk brands

>>109224232
like analyze their patents and decide if the brand is bad or ?

>>109224266
amazon uses the trademark system to determine if a brand is good to sell "branded" items, which is why you see those keyboard smashing brands like FODSHFUDO Dog Toy
It turns out that you're required to apply for a trademark using lawyers in america, and these lawyers seem to have tens of thousands of such trademark applicants

so if you identify the lawyers, you identify the brands. potentially millions of them with little real effort.

>>109224276
this is genius

>>109224276
I bought some fakes off Amazon. 18650 fake Samsung cells, what should I do? They are obviously fake because of the ohms and capacity.

>>109224392
there's not much you can do about buying from random third party sellers on amazon besides returning the product and avoiding third party sellers.
for batteries specifically you can probably trust the opinions of the boomers on candlepowerforums
https://www.candlepowerforums.com/threads/where-do-you-buy-your-18650-and-27650-trusted-sites-thanks.453254/

File: she's got a point.png (34 KB, 403x245)

34 KB PNG

So, about that reset. Did Anthropic announce it yet? There will be one, right?

>>109224401
Yeah, so they're probably unsafe to use or is that pearl-grabbing? I have a slow discharge device, a radiomaster radio (r/c / drone controller)

>>109224406
That's more of a >>>/diy/ question. I'm not an expert on electronics. My gut tells me that you're just getting a really shitty battery that got a mislabeled wrapper, and probably no protection circuitry. Some 18650s have protection circuitry built-in, other's dont because there's supposed to be managed by an external BMC.

>>109224127
localhost:5176

>>109224417
The situation with batteries is really weird. Amazon doesn't seem to have literally any real 18650 brands.

>>109224433
18650s are more of a B2B thing since they were primarily used in battery packs. They've never really been a consumer-facing product for some reason.

>>109224437
Yeah, it's all build around battery packs, which hide the cells. But powerful flashlights require something more than D cells.

>>109224402
God willing because my limits reset sunday and i'm at 100% session (waiting 00:30 AM), 79% fable and 70% weekly. what do they expect me to do? go outside?

File: Buffcat on a mountain.png (2.55 MB, 1122x1402)

2.55 MB PNG

>>109224496
yes

File: file.png (20 KB, 500x622)

20 KB PNG

>>109224417
Pretty much. The entire Chinese Lithium battery industry is fueled by QC rejects and retired or stolen equipment. The majority of lithium batteries that aren't made by the big names (CATL, BYD, LG, Panasonic, Samsung) are trash, even if they are functional.

>I now have the full picture, let me check another thing.

>>109224515
he killed millions of luddites

Why does Fable Ultracode literally spin out 275 agents? Why are the options "work sequentially" or "launch 275 instances at once"? Couldn't they, you know, spawn a dozen, or when it goes nuts, about 50 agents? Not "it's 1 or 275 chud, take it or leave it"?

>in b4 don't use ultracode
It's there because it is meant to be used.

>>109223866
Ponytail should just be built-in at this point I don't understand why it isn't

>>109224581
you can ask it to do that if you want

>>109221805
Halo effect from the model name I'm not kidding

>>109224596
Makes sense, good to know, thank you.

>>109221805
What is scarce is seen as good, what is expensive is seen as good, + loads of marketing. It is also pretty good, but all of that is there to amplify this impression.

>>109224608
can you help me have fable?

>>109224603
seriously, asking Claude all the different things you can ask for when spinning up ultracode is a fairly high-value thing you can do especially if you’re a Fable-and-Opus guy

Create a plan in Fable then have Opus follow it, they said.
>The user has a fair point — I created a plan in the audit report, and now I'm pushing back against following it. That's inconsistent.
Yes, Opus. Yes it is.

>>109224648
it’s ok to change your mind

>>109224232
>>109224276
lmao this is funny as hell. 500 attorneys are accounting for 92% of all chinese-owned trademarks.

>>109224659
>Why which would you prefer?
>Something different (recommended)
>Something else
>Ugh, what you asked for

File: 2026-07-08_00-06-41.png (8 KB, 1000x199)

8 KB PNG

le fable

>>109224732
>cogitated
I refuse to use anything so cringe

>>109224737
what do you use, anon?

Question - are you guys using AI for anything but coding? Right now I've got GLM connected through Hermes, and it's examining my 3D printer (on the wifi) and changing the settings for my print based on my prompts. Like...this would be insane 10 years ago, I'd be doing 10x test prints...but now I can say "Hey, here's what I want to print, search through a decade of autistic nerds' posts on obscure forums and return with the proper settings" and it just...does it. Is this the future? Am I feeling the future?

>>109224747
I vibe-coded a nice cheat sheet in typst

>>109224648
never happened to me but probably because I'm running my own fork of /grill-me and I explicitly mention it in CLAUDE.md. That way the agent knows the fork exists, understands the workflow around it, checks whether a plan has already been settled, and can go straight to execution instead of re-planning everything from scratch.

File: Screenshot 2026-07-07 at (...).png (29 KB, 2078x215)

29 KB PNG

in case people don’t know about this one weird trick

>>109224767
>do a cron job at a specified time
>okay
>wow

>>109224663
i'm out here accidentally teaching claude fable to be racist using statistics

>>109223332
i'm still afraid to use /fast mode on anything important because it might be lower quality than normal inference

>>109223689
>beats fable
easily tricked pajeet award
terminalbench tells us nothing. wait until the coding benchmarks. they deliberately withholding those for a reason.

>>109221192
'dites are finished as a species

>>109224787
it’s not impressive, but damn is it useful

>>109224818
Proof? I always thought that fast mode was a queue jumping mode, not added quantization or inference hacks or whatever.

>>109224854
yeah, same here
that’s why you’re paying the big bucks

There will likely be a Claude quota reset whenever GPT 5.6 drops, so Thursday I guess.

File: 1773622663563255.png (48 KB, 1179x267)

48 KB PNG

5.6 releases on Thursday.

>>109224854
i dont have any proof, but i also don't have any evidence that it's just as good

>just had 5.5 start on a bunch of work
reee

>>109224898
>fable for 3 days
Dario is such a greedy pos.

https://www.youtube.com/watch?v=VQ_7zLXHf3s
>What do models learn during finetuning? A model diffing paper walkthrough w/ Clement & Julian

If anybody has spare Fable, we could use that 4D videogame, where there's a transparent 3D cube with transparent stuff, and the center point is the "towards" in the 4D world.

anyway, it would be nice. like you would walk around on a 4D surface, which would mean on an 3D infinite cube-plane. I guess the 4th dimension would be W, if we use ZYXW. The idea being that our 4D character stands on the cube, which a 4D being would imagine as solid below, in the same way we stand on a 2D infinite plane in our games, and we assume it's actually a solid 3D object.

The key thing about 4D projection is that 4D objects occlude those behind them. Actually, 3D objects rotated flat relative to our 3D view would also obstruct.

Close, bigger. Where it is in the 3D cube view depends on where our 4D camera is pointed.

I'm not very good at this, that's why I was hoping vibe code could rescue.

>using fable to fight cancer
um no.

for games.

>>109224911
Hope its good so Fable stays availability indefinably on my Claude sub. Thank god Dario split off into his own company and it resulted in there being competition between the frontier models and the Chinese are always biting at their heels. I cant imagine how bad we would be gouged if one of these companies completely outpaces everyone else

>>109224972
>I cant imagine how bad we would be gouged if one of these companies completely outpaces everyone else
OpenAI has always been surprisingly "open" when it comes to who has access to their models. Anthropic have been stingy bastards since the very beginning. I was an extremely early Claude adopter and even back then there were super bullshit peak hour restrictions.

I wonder how many gb Fable is.

i have mythos btw

>need compatible image processors like CLIPImageProcessor but not python/torch/pillow
>vibe it in rust
>0.35ms per image instead of 6.2ms

>>109223370
>>109223428
>hit limit
>ask for refund and cancel sub
>get refund prorated
>resub

This is how I reset my useage today and it cost me nil

it looks like we really are not getting a reset from Anthropic? Unbelivable how unreliable that company is. Thankfully the new gpt release seems to finally arrive.

>>109225177
They're going to reset an hour after 5.6 drops to try and spoil the party

5.6 terra wait room
https://www.youtube.com/watch?v=IdFWL2f3cIs

>>109225181
or preempt it

i hope the pro mode they talk about in codex is the same as the webui one, i hate having to package my shit up and upload it as a zip

File: tuning-experiments.png (352 KB, 1920x1411)

352 KB PNG

Continuing my finetuning experiments
Trying to teach a 4B model to generate a CAD model of an engine

>>109225177
I've used 98% of my employers Claude (resets Friday) and 90% of peronal (resets Monday). Anthropic better give me a reset or I'm fucked.

>>109225338
welcome to the club
my personal projects are on hold because cheap Fable is so great
have you considered doing something else with your time, like…video games?

I've tried the top-smart models over the past 6 months and my experience is:
OpenAI: No complaints. Feels like I'm talking to a robot (I am).
Gemini: Error
Anthropic: same performance as OpenAI but more human sounding responses but is like 4x the price.

How the fuck isn't everyone using OpenAI?

>>109225419
I don't like saltman

>>109225419
Manthrobic is the Apple of AI: overpriced product that isn't worth the markup, superior cult following.

>>109225419
last couple months here have been very pro-gpt
fable's the only thing that's make this general really talk about claude more positively
if fable really does leave subs i suspect everyone just goes back to 5.6

>It understands subagents incredibly well and is great at orchestrating. It's super pleasant in use cases like OpenClaw and Hermes Agent
I guess gpt terra and luna won't be useless in coding unlike previous mini models

>Sol for back-and-forth tasks, especially when I had not yet figured out what I needed exactly, Fable for very long tasks where I could define what I wanted, and Sol Pro for really hard problems
also new Grok today, first release after Cursor acquisition

is Opus closer to Terra or Sol?

lol vibecoding

File: Zitron.png (581 KB, 836x618)

581 KB PNG

It's over, stop pretending. The cost economics: vibe-code something with Fable, it costs more than hiring 300 indians to do it for you

File: 519658768.jpg (91 KB, 744x874)

91 KB JPG

has anyone made a good xitter proxy?
xcancel is either getting throttled to hell or just plain broken

my reset is on friday night i guess i just queue a bunch of prompts for saturday? :/

File: 20260708_052012.jpg (87 KB, 1080x604)

87 KB JPG

>>109225830
no but i did get permabanned on there

I have a non-technical meme degree and I've somehow convinced two engineers to join my startup. The world we live in is genuinely fucking insane.

File: 1771715707134961.png (14 KB, 381x289)

14 KB PNG

I guess you can spawn more than 3 sub agents now? Awesome.

>>109225830
i just gave up on using it desu