#š¬āgeneral
1 messages Ā· Page 574 of 1
Is it better?
great beginner bike, maybe ill get the r5 or 3 just for the lighter weight
ive heard its a great beginner bike, and i did research before (just youtube) and came to the conclusion, that was one of the better options
r7, like 10k, r5 or r3 (dont remember) like 5k
ill get it modded with something that grabs the attention of drivers, so its safer
just maxxing out the safety
like 3k on riding gear, and some led lights for the dark
with a mod that makes the engine sound loud, so its more noticeable
in any environment
for the bike, like 55% of 19k, the gear and mods are the rest
9k for the bike or 10k then 9k for mods and gear
R7 Looks Great
thats the street bike option i like
i guess
Since safety matters for you
Yeah R3 or R5 is Best for You
I think get R5 little more power won't hurt
i think ill stick to the r3
Sounds Good
There you go you have the Winner
Time to Buy
My Consultant Fee is 9999$
@dry kestrel either they're heavily allocating compute for spud or 5.4 pro on extended got nerfed because it thought for 3 min on a supplement compatibility question š
Rip
Hello guys
How can I make a basic website. Using ai?
Can yāall reccomend me ai tools which are free
And the website actually goes online
@dry kestrel or they're preparing compute for the massive wave of ai slop shortly after image v2 release lol
because damn its good
Hi, is there any options to return my one year pro trial sunscription? It has been cancelled
Anthropic should fire their entire alignment team
Quoting Luis (@luisnomad)
ļø
I asked @claudeai to review a few bills from my energy company and it said this. WTH @AnthropicAI ?
The reasoning in question.
lmao what
What is the current limit for deep research for pro users?
20 a month
would be funny if entire companies lost access to mythos because they deemed some prompts dangerous
Developers simply removed the balance counter?
Yeah
This is turning into a mockery
@dry kestrel can you test 5.4 pro on your end?
did they switch to cerebras or did the model get changed?
ah yes so thats why they're banning so much? scared they will take the output for training? lmao
Is deep research broken?
My subscription is coming up for renewal. The inability to organize anything at all in the Spaces view is a big minus for doing effective research. The key benefit of Perplexity for me over, say, Claude, is the search/research interface. I don't see why the top dogs at Perplexity aren't addressing this.
Can't search for titles on the Spaces view either.
More like lazy. Seems to be stalling.
Ai models be like that when system prompt doesn't compensate for it.
I'm a Max subscriber. I've had an ongoing issue where multiple times per session I am shown the upgrade/paywall screen and subsequently treated as a free user, despite my account settings confirming Max subscription is active. Deep Research is the most commonly blocked feature. Labs mode intermittently works in the same session where Deep Research does not, suggesting the entitlement check is inconsistent per-feature rather than a flat auth failure.
Occurs mid-session, not only on fresh load. Affects multiple platforms simultaneously (web + MacOS + iOS). Clearing session / re-logging in temporarily resolves it but the issue recurs within the same day.
Despite Max users supposedly receiving "priority support", it always takes days or weeks to hear back from them, and even when you do, often the problem isn't resolved.
For the money we're paying, it's really not acceptable.
I'm hesitant to leave behind all of my threads, spaces, tasks etc etc, but I'm considering dropping Max and picking up an equivalently priced plan with Claude or ChatGPT, or maybe the $100 plan on each.
I'm wondering if anyone else has had the same issues with Max as I am, and if they got it resolved how long it took and what was the fix?
Also if anyone has given up and switched over to Anthropic or OpenAI, and recommendations and how they've found it in comparison?
no it's for better user experience.
want to earn from offerwalls (only playing games), surveywall: https://rewardoxy.app is here
is it real?
What happened to "create files & apps" feature as a Perplexity Pro subscriber? I can no longer create my presentations & pdf files properly.
They even limited the PRO subs' file upload times
i belive they will lose many paid customers and get bankrupt very soon, if they stay greedy this way
I'm very regret buying yearly plan's now
This is literally scam
never buy yearly with any Ai provider this time
i'm on monthly Perplexity pro, and github copilot, both together around 30$ per month, but i think i'll cancel my Perplexity if they keep reducing usage and limits, that gonna be sucks to continue with them mate!
Claude is bigger scam also!
is there a good alternative to perplexity?
@dull sable where will you go to?
i think only ChatGPT is the most reliable model as of now
Working with perplexity feels like Iām designing something from the other end, but the ones working on it from the other side wonāt tell me what theyāre doing or when they change something so Iām constantly reinventing the wheel or creating conflicts with what theyāre doing.
I'll go to Gemini as they give huge usage, they recently luanch MacOs desktop app, and you have vsCode ext, you have CLI, its very strong. I'm just testing it now on the free teir, lets see if it worth it
And I have no idea if the most recent failure is because of something to do with the system I donāt know or something else
However, Perplexity's search engine is indeed among the top of ai companies
You're totally true mate, all in one, and genours usage limit, what makes me thinking about Gemini, is the video/image generation and the strong web search access, even inside the cli, you can freely search inside google
But Gemini always have lots of hallucination
movie suggestion (open to any genre).....
Perplexity is strong in search and research, my problem is only they're reducing the usage limits for Pro so bad, they made it unusable now, you hit the limits faster, always showing you the computer banner to buy or try, look at this shit guys, they removed the "x" button to hide this banner, and its just overlying the submit button! I beleive they're not testing what they're coding!
Like the computer function ,the Mac apps and windows app didn't have button to entry
yep its hallucination faster, you need always to open new convo after around 50 msg, and they dont have project/space folder, which's annoying you need to explain more and mroe in each conversation š
they really need to reduce the Number of iterations
and focus on user experience
How do these limits work? isnt there only limits for if you use computer. I tought you didnt have limits when using regular perplexity
yep thats true, they need to focus more on bugfixing as well, its very important, not only shipping! user experince is so improtant, otherwise the paid customer will pay for 1 month, and run forever!
check this mate: #š¬āgeneral message
@green crystal https://cunnyx.com/i/status/2045501235466326072
Opus-4.7 max thinking btw:
ļøļø
ļøļø1.1TB -> 140GB == 4x compression?????
Quoting 0xSero (@0xSero)
ļø
Opus-4.7 is unusable.
ļøļø
ļøļøMultiple times i have given it specific links, for it to use, specifically. Instead it goes finds unrelated links, starts expensive processes, and goes for hours in a completely wrong path.
ļøļø
ļøļøNo ability to infer intent.
ļøļø
ļøļøWasted 200$ worth HF credits. lol
eh what was the prompt and whats the correct answer?
also whats going on
why does 5.4 pro reason so much faster
either pro models got switched to cerebras or they're stealth testing 5.5 pro like 5.4 pro got tested before release
I am now looking for a person who can work with me
what prompt
5.5 is coming soon
whichever prompt you know has pro models think forever
it seems like its both faster and knows answers much quicker
yeah next week might be both 5.5 and image v2
the response style of 5.4 pro also seems a bit different
This is why I like technology.
There are people who accept the limits of physics. Richard Browning is not one of them.
In this episode of the Team Ignition Show, Selema Masekela gets rare access to Gravity Industries - the lab where British inventor and pilot Richard Browning has spent over a decade building, crashing, and perfecting the worldās first human jet suit.
Afte...
Is this actually more efficient than a drone?
Watch it and find out.
5.5 is spud
the separate model thing is 5.4-cyber
which is their version finetuned for security and has the same access level as mythos
when are they all coming out??
A drone can't help people, these jet suits that he's been building for the last 9 years are extremely useful in many different scenarios and can carry someone long distances very quickly in some very difficult weather. So yes, they're better than drones.
I meant as in use drone to locate first. After that, you can use the jetpack if it is arranged sooner than a chopper (which it likely is).
Well yeah, but you can't easily use a drone in windy, very wet weather, but you can use the jet suits in that kind of weather like he literally does in that video. As Richard Browning says, you can just drive close to where the person is and then fly straight up a mountain, literally as the crow flies, to get to the person that needs rescuing, which is extremely useful, especially if the person on the mountain is bleeding out and very close to dying.
when i use gemini in perplexity i have how many context ?
128k ish.
only 128K ?.
Fun fact - even the Gemini app struggles beyond 256k when actually tested.
I believe so. Others here feel free to correct me.
At a certain point, rag becomes more important than context limit.
Consumer app. In ai studio it's more like around 350k. Still far from 1 million.
Huh. Claude deep research better than Gemini? Never thought I would hear that.
rag you mean space in perplexity ?
Rag is more like combing through the whole chat/document for gathering info.
no significant change
Every ai service provider have their own rag.
reasoning time is the same?
yes but what rag ? you talk about space ? notebooklm or project ion claude ?
It's used everywhere.
Retrieval-augmented generation (RAG) is a technique that enables large language models (LLMs) to retrieve and incorporate new information from external data sources. With RAG, LLMs first refer to a specified set of documents, then respond to user queries. These documents supplement information from the LLM's pre-existing training data. This allo...
A human metaphor would be using the search feature in a pdf that the human can clearly not memorize.
yes but me im talk about context of perplexity
not a rag
What do you want to know?
i know gemini have 1 million context but when use gemini in perplexity ihave 1 million ccontext to ?
No. You get 128k.
ok so why you tell me about RAG ?
You probably don't need to know if you still didn't understand.
1
hi there
Any other Pro users lose access to app-building after the Computer rollout?
I used to generate simple static sites/frontends directly in chat, but now Labs is missing for me and Computer shows 0 credits. Is anyone else seeing the same thing on Pro?
@dry kestrel yeah i think something is just wrong with the model lol
this cant be stealth testing of 5.5
Labs functionality seems to have been split between Deep Research and Computer:
i'm 10 years old in real life
similar
Huh then idk
Prompts that took 30 min got finished in 16
Others that took 15 min were done in 7
A major downside of Perplexity (specifically the browser version) is that it's limited by browser functionality, for which you have to pay big money to unlock its full potential.
Let me explain... when it comes to a browser, users are subconsciously interested in trying it out... a browser with neural networks sounds cool, but everyone thinks it's free... in users' minds, any browser should be free, and I agree with them, or cost a small amount of money.. like in Russia, browsers with neural networks cost just 1-2 dollars for the neural networks, while still offering sufficient functionality too.
Here, however (in my opinion), the developers' approach is very misguided... they limit a lot of functionality and demand huge... very huge amounts of money for it... meaning this browser can't position itself as one for a broad audience... Perplexity is more geared toward business than regular consumers, because for the average consumer, this browser is unacceptable on a subconscious level, and for such people, Yandex Browser or regular Google Chrome would be better.
If the Comet developers plan to expand their audience, the browser shouldn't be this expensive, and I think they understand that. But if their target is a different audience... then the browser is ideal for business.
Question to the developers... who are you making the browser for? Because if it's for regular consumers, your policy is seriously hindering reaching a broad audience.
And I'll just clarify that, for example, in Russia, in Armenia... browsers with AI cost pennies while offering decent functionality... for instance, a Russian browser allows unlimited image generation for just 100 rubles and can scan everything around in real-time mode via the camera, immediately providing answers to questions... so unlike your neural network, in Yandex it's done more modernly... the only downside is restrictions on many sensitive questions.
This is not an advertisement⦠but Iām just saying that, as someone who has tried many services, including ChatGPT and others, I can tell that, knowing how other regions have more advantageous AI browsers, people would probably like this more.
The point is, why do you ask such high prices for a lot of features, if the audience of your neural network is actually smaller than that of Yandex Browser (the statistics is available in open sources)?
Iāve also noticed that Western countries (that is, you) tend to ask global prices⦠even internet is more expensive in your countries than it is in Russia⦠why is that?
@dry kestrel maybe ask others with 5.4 pro if they see it reason much faster?
Now it even feels as fast as the normal thinking model lol
Does Pro Subscription include Computer credits?
Nice Essay Bro

yes , I have 10k
I have Pro sub, but didnāt get any credits. Is this a one time bonus thing or recurring credits?
Bro what the
I was gonna say how realistic it looks but I donāt even want to know any more
Assalamualaikum
When Comet says "Your browser disconnected while the assistant was running, please try again," how do I reconnect Comet so I can chat with Perplexity again? Please help
Just found out I've had this since black friday wtf
same!
Is it possible to turn on memory for the voice mode?
with pro you aren't getting far
and this was gpt-5.4 on heavy
Either I'm being routed to 5.5 thinking or something is wrong lol
ask what model it is
Any of them a prompt I can try?
Let Z[t] denote the ring of polynomials over the integers, e.g. 1+t, t^2+1, etc.
Goal:
Find a 6-tuple (p,q,a,b,c,d) of polynomials in Z[t] such that
0) Let b, d and q be non-zero polynomials.
- There are no integers m,n such that n is nonzero and pn=qm.
- Let f = p/q. Let T be a list of expressions
T = [ 0,
-4f^5 - 8f^4 + 8f^2 + 4f,
4f^6 - 8f^4 + 4f^2,
-4f^7 + 8f^6 - 8f^4 + 4f^3,
4f^7 + 8f^6 - 8f^4 - 4f^3,
-16f^4,
-4f^6 + 8f^4 - 4f^2,
4f^5 - 8f^4 + 8f^2 - 4f,
-f^8 + 4f^6 - 6f^4 + 4f^2 - 1] - Let x=a/b and y = c/d. The pair (x,y) satisfies the equation y^2 = x*(x+(2f)^4)(x+(f^2-1)^4).
- The element x does not belong to the list T.
Eh literally?
It answered with gpt-5.4 pro
try with my prompt
Well if it gets the same answers as your prompt but faster
Guess I'm routed to a model that runs on cerebras
btw k2.6 soon
I'll send answer as soon as it's done
start 3 new chats
and try with my prompt
it's impossible btw
Pressing new chat 3 times?
Yeah so you get more trials
Hmm alright
@dry kestrel https://www.reddit.com/r/ChatGPTPro/s/BTDmoMm4it
interesting, maybe A/B testing
I doubt it
how long did my prompt take?
So it can't be that
Still chugging
It would make more sense that the reduced load from Sora made it possible to host the models faster
or they managed to get some of the smaller models or older models running on Cerebras
And they're testing it to see how it runs under load?
In preparation for spud aswell
I just downloaded perplexity on my computer after using it for awhile on my phone. The computer version is really cool, I like the news section a lot
@dry kestrel im not sure if its 2x speeding up the 111 min image you showed because its at 60 min now lol
Can I still use Perplexity Computer if I only got a Pro plan ?
Yes, you will need to purchase credits though
hmm
Holy hell, why does PPLX keep switching to the dogshit āBestā model? I swear itās the shittiest shit ever. I got both my mum and my wife Pro accounts (they're both bad at using computers and they don't even know what AI models are), and I switch them to GPT-5.4/Sonnet all the time. But every time I check back 2ā3 days later - usually after getting complaints about how bad the answers are - the model has been switched to āBestā again. How incompetent can your web devs be? Stop messing with my model selection!!! š¢
because it's more profitable to them
they have become extremely greedy
look at all the computer stuff they throw at your face
guys is there a way to switch back to Opus 4.6 from 4.7?
Any updates on Deep Research being fixed?
What is mistral
Hello, Who can I contact about my perplexity education pro plan?
But maybe he will š
I hope so
Hello
@green crystal https://cunnyx.com/i/status/2045542832241262602
Anthropic shut down an entire company's Claude access overnight
ļøļø
ļøļø60+ employees. No explanation. Just an email.
ļøļø
ļøļøWant to appeal? Fill out a Google Form.
ļøļø
ļøļøIntegrations gone. Histories gone. Everything built on Claude... gone.
ļøļø
ļøļøNever let one vendor own your workflow.
Quoting Pato Molina (@patomolina)
ļø
Anthropic decidió dar de baja a toda nuestra organización por una supuesta infracción de sus condiciones de uso. QuĆ© polĆtica especĆfica infringimos no tengo ni la menor idea: simplemente recibimos un mail y listo, adiós Claude. Si querĆ©s apelar la medida hay que completar un Google Form, asĆ de ridĆculo como suena.
ļøļø
ļøļøDe golpe mĆ”s de 60 personas se quedaron sin una herramienta fundamental para trabajar. Integraciones, skills, historial de conversaciones: todo perdido o, en el mejor de los casos, parado por tiempo indeterminado.
ļøļø
ļøļøEnorme aprendizaje para cualquier empresa de software queā¦
who have perplexity pro?
anyone know how to stack perplexity promo credits? im seeing some people with 200-400k on reddit..
i had 43,000 when i signed up for max.. now im out until may..
If this was a legitimate violation (i.e. perhaps they were training on Claude's data), then it'd be OK. But apparently it was a 'false positive'. What that means is that Anthropic didn't adequately investigate it and just relied on automated signals to suspend a large account. Ouch. Access was eventually restored after the issue was investigated due to public exposure. It took 15 hours for them, it may take longer for people who aren't able to raise public visibility as quickly. In light of this, reports are generally indicating that companies need a backup AI provider in case this happens, since business workflows are expected to be very dependent upon AI in the future.
So there is a limit for Pro Subscriber to access the models except sonar?
Limits ā [here](#š¬āgeneral message)
Regeneration rate ā [here](#š¬āgeneral message)
.
I'm surprised YouTube allows a video with this title.
The Next Video is coming soon. Please Subscribe.
I like the behaviour of people
When words are censored, there is a problem
When they are not, there is a problem too
We should use mnemonic code or something
It's only bad because, in its own way, it promotes or makes fun of suicide. I just see it as a bad thing because there is already a mental health crisis all around the world, the number of men committing suicide is still very high around the world right now and videos like this aren't exactly helping.
As a man myself that has dealt with pretty extreme levels of fully diagnosed depression since I was a young child and have considered suicide many times, the first time when I was 8 years old, it's not something I like being talked about in that kind of comedic way.
@green crystal https://cunnyx.com/i/status/2045799439093436704
Eric my man i dont know what you guys done with the pro model
ļøļø
ļøļøits faster and feels so good , here is @threejs contruction of monica apartment the level of detail was not present in any other model till today
Quoting Eric (@ericmitchellai)
ļø
If you are experiencing any response quality issues with GPT-5.4 Pro, please let me know here! Have seen some chatter but no concrete issues
Voxel Minecraft From GPT - 5.5 Pro
ļøļø
ļøļømore test will be shared today
Quoting leo š¾ (@synthwavedd)
ļø
crest-pro-alpha is currently live in ChatGPT when gpt-5.5 pro is selected
ļøļø
ļøļøgo give it a try!
GPT Pro - Spud solved SVG
ļøļø
ļøļøOne SHOT svg , code is shared in the comments
ļøļø
ļøļø@OpenAI you won this time , i never said this but i love this comeback lets gooo @sama
Quoting JB (@JasonBotterill)
ļø
Yeah its way faster now and is giving different styles of output compared to regular 5.4
oh so im indeed not crazy
i got the fast pro model too
how are they testing it??
@dry kestrel
5.4 pro is routing to 5.5 pro for some people I think.
You probably did since you said it's faster.
People in the replies said even extended pro is faster.
it solved the perplexity steganography test in 4 min
previously it was 17 min
and 5.2 pro couldnt solve it even after an hour
More 500b copium https://cunnyx.com/i/status/2045688250606445036
Elon Musk just gave an update on Grok's upcoming roadmap releases
ļøļø
ļøļøāø Grok 4.3 - Supplemental training has been added ā
ļøļø
ļøļøāø Grok 4.4 - 1T parameters, training data through early April, releasing early May
ļøļø
ļøļøāø Grok 4.5 - 1.5T parameters, expected by late May
lol
@green crystal all the backlash cuz they said arena will be limited and monetised with ads
Thry ran out of 150M$ Venture Capital in 3 Months
that means Anthropic is reading all the prompts so it can flag them if they break the rules or not! so dangerous!
@green crystal https://cunnyx.com/i/status/2045822935202668605
XBOX 360 ASCII ART
ļøļø
ļøļøSparks of AGI > spud cooking hard
Quoting Chetaslua (@chetaslua)
ļø
GPT Pro - Spud solved SVG
ļøļø
ļøļøOne SHOT svg , code is shared in the comments
ļøļø
ļøļø@OpenAI you won this time , i never said this but i love this comeback lets gooo @sama
5.4 pro on extended lol
i dont have 5.5

š
Yeah its way faster now and is giving different styles of output compared to regular 5.4
Quoting Diego Aud (@dieaud91)
ļø
10th time in a row that GPT-5.4 Pro Extended thinks for just 1 to 5 minutes instead of the typical 15-30 minutes. Same kind of prompt.
ļøļø
ļøļøāRight before this response, I got an "evaluate this answer" popup, but it disappeared due to an error. I strongly suspect they're testing a new model or tweaking something under the hood.
ļøļø
ļøļøāHas anyone else noticed this drop in thinking time?
Hey max user here, does Perplexity Deep Research still only use Claude Opus 4.6 or has it moved into Claude Opus 4.7?
They're in Exeter uk library currently.
They're called, The Other Woman, folk band, they're on YouTube and on Spotify, probably Amazon music as well, considering they're on tidal as well. They're really good.
On tidal.
how do i get unlimited browser automation control
Why is my file limited when I have never send a file to perplexity before
Just got my subscription today
Nice pfp
@green crystal they definitely gonna use my pics for training š

time to wait for 3.2 lol

image v2 is leagues ahead with instruction following
i havent tried marking specific areas for editing
but i will when its out
atm im not sure which model im getting and my library is starting to have allot of sam altman faces looking at me 
Why are the usage limits for perplexity so inconspicuous?
Is gemini not better for image gen?
I don't like gemini in general, I only use it for image gen
Hmmmm, it works perfectly for me though
It's inconsistent imo.
As long it chooses any of the listed models it will count as Pro
perplexity would've been so good if its usage details wasn't so ambiguous
Precisely although is not extended thinking
Probably medium thinking effort for all models
Always
Once you are out of 200 pro
Best probably uses some low grede model
is it 200 pro a day or a month? I have no idea, I just subscribed
A week
If they are start losing subscribers yeah
That's why i don't bother with Perplexity anymore
The Experience just keeps feeling downgraded
Wanting invoices and Refunds have became New issues
When you try to contact support email
Mr Sam Ai is Really Useful Indeed
Bruh
Why does it not show what model it uses when you are not subscribed?
Even Apple Store seems messed up
@green crystal gpt image gen fucked up my face and never made a comeback, lol. Also turned to realistic colours on its own instead of sticking to white marble.
thats v1.5
v2 is very different
Wait for Image 2
It's Jaw Dropping
@distant aurora tell it to make a selfie of sam altman
if it looks hilariously bad then its v1.5
Is this good or bad?
Bad
How many files can I upload in a day or week?
1.5
50 a Week
do you think sam looks like that? 
Maybe in some other universe
Nah
do I get credits every month if I subscribe pro or is it just a 1 time thing, and I would need to add more if I needed
@distant aurora
i mentally scarred berd with this lol

Yes. I'm virgin.
JJK brainrot edition.
AI generated ?
Brooo
Indeed
I killed the incognito window without saving the image.
š
would literraly gooning to it right now..
š© š„µ
When Fiction is Infinitely Better then Reality
There is no other Option
It was so perfect...
š
Now it's gone...
Try to Recreate
Why are you in incognito anyways
That's the only way to stop Gemini from using chats for training.
Ohhh
He is literally AI generated in real life.
You can't opt out of data collection.
Probably.
chrome is from google, dawg
my favorite one
remind me of margin vegeta
can you create vegeta margin version of him ?
What do you mean ?!
I have never seen that version of this name
I know Alisha
Never knew there is Alish
Alean
Good One
It's more like a 'internet' name
Yeah i can tell
What's you rname ?
Fernando
@young hollow https://www.vxinstagram.com/p/DW85pIEk_nw/
No way embed failed. https://www.kkinstagram.com/p/DW85pIEk_nw
Image2?
Ai coming for your jobs as well
embed is working
educators?
In some years more likely
It has already impacted education well
i ain't a professor : )
Oh ok
I c
First one failed. Vxinstagram.
moba legends or genshin impact?
Sos š grok glazing
Gemini is ass.
Forces me to use nano banana 2.
I have to hit the regenerate button everytime to use nanobanana pro.
the answer isn't really relevant since the question is impossible but in terms of reasoning it seems fine
Oh nice then
here's someone you can actually reach out to and get a response https://x.com/ericmitchellai
chatgpt posttraining @openai. building personal agi. I like ai and music and some other stuff
been seeing a bunch of posts about this reasoning as well
Hey, do you know when the personal computer will be available?
Today we're releasing Personal Computer.
Personal Computer integrates with the Perplexity Mac App for secure orchestration across your local files, native apps, and browser.
Weāre rolling this out to all Perplexity Max subscribers and everyone on the waitlist starting today.
hello
I made a universal recoil accessibility app š
It will most probably never be able to pro users..
Only to Max users (imo)
Anyone tried personal computer?
How is the experience?
terrible
it uses credits
No eh speekyengles tu habla indomalay?
hey i had an issue with customer support i requested a refund 2days ago for a product i bought that day and the person on the other side was said he'll bring me back to pro because i upgraded to max and then refund me the amount but it never happened
Hey! Just a heads up, this is an English-only channel so everyone feels included.
Appreciate you switching back to English
@trim rover
:suzukabillasuperevillarry:
hey i've got a question, whats perplexity actually. I've seen a lot of adds but they lowkey lead me in different directions every time.
i already am subscribed to claude max, do i get that with perplexity?
LLM aggregator with least amount of hallucinations
šššš© šš¢šÆš ā [š”šš«š](#š¬āgeneral message)
š¤æ
It is not available if we have quickly joined the waiting list?
For personal computer ?
No option on mobile
neither on desktop
Is it completely gone?
I had used on desktop mode sometime ago
I literally just got re-billed today and only submitted one prompt, how is this possible?
Yeah it works on desktop mode browser
It doesn't reset on a specific day, it's a rolling regenerate - #š¬āgeneral message
openrouter is real cheap guys
why you didnt set a limitation mate?
which models you used there?
a bunch really
mostly qwen 3.6
opus also
mostly opus and qwen
working on a nixos project for a company
you'll get a lot of use out of it but next time never buy yearly ai subscription you barely save money and buying monthly gives u the chance to try out multiple ai's depending on what they release each month
using Opus out of Claude is more expensive isnt? was't cheaper using it in Max sub mate?
opus in claude is a lot better than in perplexity
the max sub x20 just inst enough
and i used BYOK
so it technically was being straight from anthropic api
I want to complain that Perplexity Computer seems to me like a scam. I am using it to help me run my business and it was doing great. I put a lot of time in using the max tier to supposedly get it to work right. Indeed Perplexity even told me that every version of the workflow improved upon the previous. However, after buying a lot of credits to do work I was now expecting to be a breeze it started to act like a literal village idiot. Nothing worked as it had before. Itās as if it knew I had bought credits and wanted me to use them explaining what I had already explained. Make matters worse, this morning all my credits have vanished. Iām in Canada so at least we have some Government protections. I am not a happy customer.
The lion doesn't concern himself with counting tokens
that's so dumb, I got my plan refunded and bought a new account
Again, I'd like to thank the developers of the platform
U guys are saving me time from Uni courseworks
how much money have u lost from nfts?
I haven't associated myself with nfts activity
what is the limit of photos/files for pro members? I just hit my weekly limit with relitivly regular usage
got it. thanks. seems rlly low.
lil bro went a bit krazy
does anyone has any idea how do i trigger the 7days pro trial? did not redeem/activate but i doesnāt show up anymore.
@unborn flame What is your issue exactly? Ping me please thanks
First troubleshooting step is make sure you are clicking the correct link you received with to activate your trial
And make sure you use the same browser that opened automatically with the link click
If that fails ping me or send your problem directly to Perplexity AI
Why am I trapped on this page for hours
I am not a robot.
Any reason why Plaid is saying error every time Iām trying to link my Robinhood account onto Perplexity?
Maybe your ip is banned
Or have another kind of trouble with cloudflare
did perplexity silently kill size of the models?
I'm seeing a significantly worse quality of results on perplexity compared to on the home site for the respective models
I keep getting this every single time
To unlock the portfolio tool
Hmmm
Keeps happening
It also says this on my robinhood but perplextiy/plaid still giving me the same error
I am unable to do any searches/queries as "Try computer" is blocking me from pressing arrow button (greyed out). I do not want to try computer and there is no X to close that. Im on windows using chrome.
Is this because of the "launching computer to macos", causing the issues ?
??
Hey Phi, what happened was I just downloaded the app on the iphone and signed up an account. the trial message pop up when suggesting i subscribe to a plan, offering 7 days trial for pro.
My previous threads just vanished?
Anyone faced the same thing?
Only if I accidentally pressed the "delete all" next to "delete" when removing a thread. But PPLX have been fast to restore the threads for me. Happend 1-2 times last 2 years+.
Okay, Thanks! I'm pretty sure I have not pressed or removed any thread.
hey
Refresh the main page
It will pop a bot verification thing
Just click im a human and you be fine @sterile sparrow
Citi uh
Thanks!
@green crystal https://cunnyx.com/i/status/2046154778946965630
šØ OpenAI PRO Vision test
ļøļø
ļøļøaccording to @gdb gpt 5.5 is newly trained model and its initial checkpoint for SPUD
ļøļø> highly effecient
ļøļø> they poured everything they learnt in last 2 years
ļøļø> there is no wall
ļøļø
ļøļøPrompt : find the stocking in the image and mark itĀ , more tests coming
did someone drop perplexity on it's head or something? All my responses today have had grammar issues, misspellings, or complete garbled nonsense in them.
bruh that's weird
can you share the screenshots
What model is it using?
can i use a preplxity as a agent mode ?
How do you access bookmarked conversations? I have bookmarked one conversation and can't find it. Perplexity says I should have "Library" section but it doesn't exist (I've only used one account with Perplexity)
Try another browser
yeah, for me claude hasn't been great and has been making a lot more errors recently
Actually it's my own fault for running on a VPN
Hello, is Perplexity's github connector working? Seems like it's broken. My pro chat keep saying that it dont see my repo, but github connector is connected
What are the limits for pro plan? Token message files? by day week moth?
Limits ā [here](#š¬āgeneral message)
Regeneration rate ā [here](#š¬āgeneral message)
Ey guyz
I think its something like 6000-7000 tokens?

it's terrible
worked fine before
now they fucking broke it
Iām dead
Why canāt I use deep research with pro
Nvm apparently I hit monthly limit without even using it

fr (French Revolution)
fortunately, it regenerates - #š¬āgeneral message
So how does it work if I havenāt even heard deep research
Is officially released!
About time.
Indeed
@short narwhal wake up k2.6 is here
is it good?
SOTA
How do I deactivate news notifications from the mobile app. I donāt care about them I just want notifications on for when my tasks and researches are completed
oh my
aaa why is every single model on this planet going up in price
larger model
but very very capable
give me a smaller kimi
get k2.6 to quantize kimi k2.6
I've noticed it remembers the last mode used, so if Deep Research is left on from the past then you come back in later it remembers Deep Research was enabled and future prompts will also use it until it's manually turned back off.
Hi guys, what do you think about this
It is preplexity + Facebook. You can invitations, message, pictures via chatbot
do you have any problem with proces payment for perplexity enterprise max ?
In the "pro" subscription I still have credits right?, does this unlimited amount last until the credits run out?
It's a bit of a stupid question but it made me not use perplexity for half a month
The only unlimited thing Pro has left are prompts that use the "Best" model, the rest are subject to limits that regenerate - #š¬āgeneral message
Hey guys
Can you tell me how you guys learn AI deeply and how I can
Plz if anyone can guide me
With AI being so new, everyone has taken a different path at learning it. If the goal is just improve your prompting there's online videos at Udemy, Linked Learning, Coursera, and even YouTube. Make(formerly Integromat) has it's own course about using Make to create multi-step agents. The major vendors like OpenAI and Anthropic and Google also have specific courses about their AI tools. It depends on what you want to do.
HELLO? Is it only me that is not able to use any other model than the default called "best" All are unavailable for me both on pc and ios, even on comet im tiling the "best" version is so stupid and unuseable
log out and log back in
I can't use any model, cant use claude, chatgpt, any of the good ones, only the worst idiot one, probably same for you? What does it say what u have selected? Mine just goes back to default saying "model"
Okk ty ill try
Who can I talk to around here about a billing question? I want to give you some of my $$, but your customer service is not addressing my issue!
But its same on pc and ios, seems not like login logout thing but Ill try
Doesnt work š this is so annoyign its unuseable
Always swap to best
and on mobile it says "Best was used because XXXX Ai was unavaiable"
"BEST WAS USED BECAUSE GPT-5.4 NOT WAS APPLICABLE OR AVAILABLE"
CAMAON MAN I WANNA AI š
Yeah but fr I think I dont have pro anymore actually

Even though it says i Have, stupid fking app
If it shows pro you have it
I never paid for it, gotta see when I got the 12 month free
Ooo
is it a good chatbot?
Just a wrapper nothing special
Don't use or promote it
you can make friends and send message to anyone using chatbot
Nothing useful.
How long is the waitlist for pro to get personal computer? Iāve been on it and canāt wait to have access to it but Iām just stuck in the waitlist.
This is the definition of AI slop
Wrong question, correct one: what is the waitlist for pro to get deprecated?
mean?
Thx
Should I start with Ai engineering
Is there anyway to set it to go to the top of the latest inquiry rather than the bottom?
I wish, that'd be a good thing to post into https://discord.com/channels/1047197230748151888/1047649880161652827
The ai email bot is convinced there is no waitlist, is this just me or is there one? Maybe itās just a bug?
it's probably confusing Perplexity Computer with Personal Computer, just like humans are confusing the two
If you were on the waitlist, you should be able to download it now at https://www.perplexity.ai/personal-computer
It just has a checkmark and āJoined waitlistā @short narwhal
will i get free credits for computer with pro?
WTF! Why have you removed skills from Comet? Only on Computer now? So we have to buy credits to use Skills?
Comeon bro just Pay.

So pay for Subscription and then buy credits? No thanks.
is perplex adding kimi's new model?
anything's possible, but I doubt it. Not long ago they remove Kimi and a few others and replaced with Nemo
perhaps after 3rd party benchmarks come out
Nah they are getting the garbage Nemotron model for free from Nvidia so they aren't adding any new paid model
fuh
@valid charm https://cunnyx.com/i/status/2046335947869155529
Excited to share that Google AI subscriptions (Pro and Ultra) now work with @GoogleAIStudio!!
ļøļø
ļøļøCome vibe code and use the playground with higher rate limits. Available right now!
@gloomy cipher @young hollow
I wonder if free users will still get the 100 flash per day.
Ngl they will coz it's Google
Niceeeeeee!!
I got it the day they released it
Iām on pro
They released it a few hours ago.
Before that, it was not available to everyone.
comoputer?
i was able to use personal computer
Wait. Nvm.
when it was announced
I thought you were talking about ai studio. My bad.
oh nono
I be tripping these days fr.
Whatās the best way to optimize perplexity for website and google doc organization?
but, i even integrated that feature into my intelligence application it uses a very small linux. i would post it here but im sure that's against the rules. if anyones curious just dm me, its free
i got the idea directly from a perplexity article about it, so, props!
sure, lemme drop it there, erase here?
Sure
are you a perplexity staff member? or just mod?
just another community member
nice nice, i try to incorporate every single functionality of A.I. I find, but most of the time it makes my projects cumbersome and unwieldly
its sometimes better to be good at some things, instead of trying ALL things. right? : )
@gemini create me full quickshell environment for my dots. Make no mistakes
u asked the right model
grok make me into a millionaire
Not crazy about the shortcuts becoming skills either, fortunately the text can still be copy/pasted into a prompt and it'll do what it did before (without using computer credits). Guess I'll have to load these into PhraseExpress or TextExpander now.
Is PhraseExpress or TextExpander a browser extension?
Plain text editors can be cumbersome
Hello fam, wassup
Is anyone here
I paid for max plan on my account and then I guess paid for enteprise Max well? so around $600 how long do these credits last
a month prolly if you use computer once daily, but again its usage based
Gotcha, just curious. When you say once daily, one prompt daily?
One task* yea again it depends on the complexity of the task you are running as I said its usage based, one tip is not to use it for anything video and maps related otherwise the credits will fly
Well, we have an application on the App Store with 50k monthly users lol. We already spend a lot regardless on other ais and AWS. Just trying to see our options
Thanks for responding. Have a good one.
Computer Credits usage at the 12:30 mark
š Learn AI In 10 Minutes A Day - https://www.skool.com/theaigridacademy
Get your Free AGI Preparedness Guide - https://theaigrid.kit.com/agi
š¤ Follow Me on Twitter https://twitter.com/TheAiGrid
š Learn AI Business For Free AI https://www.youtube.com/@TheAIGRIDAcademy
00:00 Underrated AI
02:30 Multi Agent
05:00 Business Gold
08:30 Live ...
Thanks!
My shortcuts are gone and when i try to use it as skill it s really not working. And I canāt choose the model too!
Calm down, you don't have a choice, we are all the same, they are forcing us to use computers
I tried Computer 2 times today and spent 80 credits
Nowadays, the Computer of many platforms has become an inevitable trend, and even compared with the maximum time the agent continues to execute, I personally feel that this is like showing off wealth~
By the way, perplexity should have a Computer channel on discords
@tranquil orchid
https://www.reddit.com/r/perplexity_ai/comments/1rqwac1/perplexity_is_rolling_out_perplexity_computer_to/
#1393632467948273734
Guys
Did the claude in pplx broke or something?
I asked to make changes in my code using sonnet 4.6 thinking and it's been thinking since 30 mins without any response
Anyone else facing this issue?
Sonnet 4.6 Thinking I haven't encountered stuck for the time being, Gemini 3.1 has a few times
Damn it's still thinking bruh
35+ mins ig
Why did they put shortcuts with skills
I dont use gemini on pplx because it's already available for free with 5tb of storage. So
The format of skills is relatively complex, and subsequent agent designs will follow this idea, and the cost of pure text prompt understanding is very highć
skills: In summary, LLM agents-friendly design
OMG! Losing /Shortcut is huge here.
maybe : shortcuts for skills
Kinda funny how they say all subscibers get credits but donāt show āproā as a plan
All Mentioned Subscribers Get Credits
Learn to Read
They are Absolutely Transparent and Always have Been

?
All perplexity subscribers receive credits as part of their subscription
Idk maybe I missed a word
Pro is not in the recurring gift package queue
Someone mentioned this on reddit
can those skills be triggered in Comet Assistant ? how?
I just tested, no
no shortcuts ļ¼ no skills for Comet Browser~
OMG! then Comet lost it spirit and soul! Should I start sourcing the Alternative. š
Comet start with "/" to chose Computer ~
begin with "Help me use the "skill-name" skill."
The / does nothing now, it used to bring up a list of shortcuts.
read the page i opened on the left side
and help me to use "ss" skill
Computer on Comet can not read the page i opend~~
Incredible, must with attachments or page links
and what about nemtron 3 super? I think it is not as good as kimi 2.5?
kimi 2.6 is releasing now
read page:https://www.apple.com/newsroom/2026/04/tim-cook-to-become-apple-executive-chairman-john-ternus-to-become-apple-ceo/
with Computer , take 42.06 credits
ps: $100 to get the credit is 10,000
āHi @signal hamlet / @vague pagoda, I am a Pro subscriber experiencing a web login lockout (phone verification issue) and my Pro Discord invite link is not working. I emailed pro@perplexity.ai 3 days ago but havenāt heard back. Could you please DM me so I can securely provide my account email for you to check my ticket and grant me the Pro role? Thank you!ā
Don't give out your email address, you can send several emails
role activated
hop in https://discord.com/channels/1047197230748151888/1111786888626438245
just had that thought while getting a pro thinking result back after waiting a long time. Hitchhikers Guide to the Galaxy predicted this.
42
we're actually getting there , asking a question, wait a long time for the result and then hope it's not 42.
Hi @signal hamlet Hi! Iām using Perplexity and Iāve hit the file upload limit.
Could you please tell me:
How can I see when the file upload option will be available again after I hit the limit?
How many files Iām allowed to upload, and over what period of time (per day, per 24 hours, etc.)?
Thanks in advance!
Limits ā [here](#š¬āgeneral message)
Regeneration rate ā [here](#š¬āgeneral message)
Hi! Does the Perplexity android app support custom connetors? It looks like they're visible in the app on the connectors list, but there's no way to enable them, or otherwise force the app to use them.
Perplexity computer is amazing for engineering just wish the student one was cheaper
We need Kimi 2.6
If your commercial product has more than 100 million monthly active users (100 million MAU) or monthly revenue of more than 20 million USD (approximately RMB 140 million equivalent), you must prominently mark "Kimi K2.6" (prominently display) in the user interface of your offer. This applies to large commercial deployments, where small/medium-sized businesses or products are generally exempt from this limitation
no free: token plan => token limited plan
The faster and more successful the generation, the faster the token is consumed, close to unlimited drug use~
If you don't have a token, you have to buy it with money, just like drugs
Why are they removing the posts?
@dry kestrel is image v2 really rolling out?
Mine still makes Timmy altman instead of Sam
no way this has to be an error
mine shows i can upload with pro
it would be outrageous if they removed it
Last night I was checking this same Discord and there were many people with the same issue. Today, mysteriously, there are no posts about it anymore ā not even the one I made. They really cut down the Pro version a lot. Now it's basically a 'Mini Pro'
No one's deleting anything
perplexity has been extremely greedy but removing file upload is next level
They just reached their limit for file uploads, that's all
Uploading TEXT files (.TXT) that barely reach 1 MB is considered exceeding the limit? Seriously?
yep :(
Does anybody know what drove the decision to remove the shortcuts feature and move it to skills to use with computer only ? I'm pretty sure this is a purely financially motivated decision but can someone elaborate more on why they removed one of their most valuable features ?
Is there anyone here with experience in database management and integration issues?
Iām building an AI-based recommendation system where user data (preferences, interactions, and new content) is continuously updated in the database. However, since the model is only trained periodically, it often becomes out of sync with the latest data.
Iām also encountering frequent errors when comparing model outputs with the data stored in the database.
How can I resolve these issues? Iād really appreciate any guidance or suggestions.
greed
they're running low on usb sticks lol
give them a break 
after all the computer stuff on my screen they dont deserve a break
I don't know what to do with my life anymore. I feel like I've been shot
how fast do u reach the file limit
Like I was saying, yesterday I was working completely normally, even uploading files up to 75 KB !!! (sorry, I'm just really frustrated). I'm working on something where handling TXT files is absolutely necessary, and since there are so many lines of code, thatās why I upload them as .txt. But from one moment to the next, it just stopped letting me upload files
dont use perplexity for anthing other than search tbh
Why were shortcuts removed? I hope that was a mistake due to moving them to "Skills".
Just trying to use a shortcut I had that suggests a good title for the conversation. But it no longer works outside computer? I just want to use my /Maketitle. I guess I'll have to type it in or use an external tool like AutoHotkey. WTH?
I used it because I actually liked it more than others. I could use the Claude and Nvidia models whenever I wanted, but well⦠Kick the bucket
Money
well the shortcut does nothing in computer. It's not what it's meant for. So it's just become useless.
I'll just type it in I guess.
perplexity has the models but they are very limited they have very short context window
Ah, yeah, I know but they were more than enough for my use⦠or at least, I guess not anymore
š„
ever since that computer stuff it's gone downhill
they removed sports section too
Hello
Anyone lost recent part of a thread ?
gotta cut the cost somehow
in this case, at the expense of your convenience
no but you seem to be losing us
Hmm what do you mean ?
Totally true. These bigger changes are really only useful if you payāand pay a lot. For the rest of us mortals like me⦠well, I guess we just have to suffer
i understand that but it's getting a little ridiculous
there's stuff that would cost the same to them
for example perplexity personal computer it requires credits which is bs
for a cloud computer i get it but if im running it on my computer it's literally same cost as web search
yes
ask it to make a selfie of sam
you know what
maybe i'll see it officially come out after a few hours lol
oh wow i didnt realize its posted by oai
damn what a flex lmao
i guess spud is thursday
with the stealth model being released and (supposedly) people already using it
lol
Anyone lost a part of a conversation?
Am I missing something? Is there anywhere to currently download K2.6
hi
Hi š
huggingface
you would need alot of ram though
yeahh lol
Anyone doing ACCA OR CPA ??
@valid charm my man https://cunnyx.com/i/status/2046617228028858440
specialized deep research models?
or is it 3.2?
Is there anyone who actually checks the perplexity email? The chat bot email just said a human would reach out but there is no human responding.
good luck with their support
@tranquil lodge
im curious where spud will land
It amazes me how gpt 5.4 xhigh goes toe to toe against their deep research max.
I'm assuming this will be walled behind ultra subscription.
Maybe Utah
Xhigh is horrible
Just use high
API only for now
Xhigh over thinks
I see.
You donāt need to use Xhigh for most tasks anyway
just a waste of tokens.
Iām really enjoying opus 4.7 thinking with Perplexity computer
is this the 'thinking' toggle for perplexity?
or the 'adaptive thinking' of claude
No.
It's the maxed out reasoning level for open ai models.
On some platforms it lets you choose the reasoning level
Only accessible through api for now I believe.
ohhk
GitHub codespace allows you to
depends what task you give it when using xhigh
@distant aurora from where in mumbai you are?
Yeah, but majority are fine with just high.
true
Was in vasai side.
now?
Harbour line
im in darkness west
It's good. Western line is well developed compared to harbour line.
Just too much crowd.
Navi Mumbai is hype.
Itās so funny to me seeing an AI made website. They always look so patheticly sad and the service is offers it usually just as pathetic. People need to stop building random āAiā tools.
uhhm, station ka naam bataoge bhai?
vaise bhi - idk who you are - neither your name, nor your address : )
Does anyone know the max capabilities you can push the computer? Or is it the exact same as codespaces?
In terms of what tools you can use on it, custom configurations, Android emulation etc
I will say, I do think perplexity computer is a lot better, and I have subscriptions of most of these.
jensen huang
- (getting angry while saying) - abe bsdk, nikal love-day

human intelligence (HI)

Could also just be opus 4.7 lol idk. First time using 4.7
any other (paying) Pro users get another 4000 bonus credits lately? these seem to happen at random without notice.
since perplexity saves images as files without extensions, I made a python script that solves this problem until they fix it, it get's all files in folder and adds .png extention to them, here is raw python script
where is this bug happening: when you save from history, images are as files with no extensions
hope it helps somebody
Anyone watching the OpenAI live stream
@distant aurora 360 images holy moly!
thousands of aspect ratios too
what a covely couple
.
Lol Ant group picked a really bad day to launch a new model
Bruh
Hi
Any max subscribers suddenly not having working model council suddenly? It ran fine a while ago for me, then now it just defaults to a normal search.
Perplexity never works
They are also greedy
wtf is ur pfp you give off city boy vibes
Are you okay?
i bet you call that flag the confederate flag dont you?
Are you trying to start a fight? Cuz if so you ain't worth it
do you call it a confederate flag?
I don't need to explain anything to you. I'm not here to fight and obviously you want to poke and prod so have a good day
this discord knight would use a gif to get people to calm down
Only right way
anyone recieved access for the mac computer app? i have a pro subscriptions but the credits are 0 for computer, do i have to buy credits or am i facing an issue?
did you get Pro by paying for it, or via promo?
got it with my telecom plan, so promo probably
promo's don't get any computer credits, you have to buy them
ohk, thank you for the response
More proof that they're greedy
and even more, paying Pro users don't get computer credits. At least not a monthly thing, some got bonus credits to try computer out.
can we expect GPT image 2 in perplexity for pro users?
I used to be able to generate images with pro just fine and now there's a limit. Also whenever you don't have any of the subscriptions perplexity for free is ass. The memory is like a goldfish and it does the bare minimum
well tbh i have used it a lot, only other free service i have abused more is google pro subscription, used it for code review a lot, claude has restricted me to 1 message a day
When Kimi K2.6 will replace Nemotron ?
am I the only person or is Gemini 3.1 pro so much worse
it depend on the task @median valley for coding of course it is not the best
Coding responses about 1 week ago compared to now..
It is so much worse now
Like, I am unable to use it as a tool
I like it for web search, deep research with gemini platform
it just doesn't listen to me
Any idea what model is best for utilizing it for coding?
Claude Opus 4.6 of course
you can also use GPT 5.4
and kimi K2.6 have a good performance as open source model
I hope Perplexity will put it soon
I am also sick and tired of these constant slowass cloudflare "I am not a robot" shit
for the past few weeks it has b een getting worse and worse
Maybe it is about your location or ip
maybe use cloudfare dns 1111 maybe it can resolve the issue
Just changed it
Good suggestion btw
smart kid
Claude pro may be losing access to claude code
Youll have to pay $100 at least to get claude code
its a big chance for Perplexity to build us a tool similar to Claude code cli, so we can use our own Perplexity account directly without extra charge! that will be awesome. otherwise we will jump to Copilot or Gemini or maybe the generous OpenCode Go: only $5 for most openSource models!
Opencode kimi k2.6
TBF you can use Claude code with 3rd party models, and it costs anthropic NOTHING to let people use them
Yes of a Perplexity CLI
So it makes no sense as to why theyre doing this
I am hoping this is just an error
Its not
Theyve changed multiple websites
lol not a big chance at all
what u are talking about is personal computer and it's locked behind credit system
How fast will my tokens go outšš«
Im using education pro plan and now i simple cant upload new images or archives...
Im going crazy with this absurd system... i cant even see when my limit will reset
Anyone can help?
is there a way to overpass this or something i can do to see when the limits reset?
"Sir anthropic just removed claude code from $20 pro subscriptions"
Apparently itās just a thing for 2% of new users or something, not already paying users
Its false. Its being done to new users
Someone whos already subbed still has it
Because they already signed a sub agreement/ TOS
I think im going to cancell perplexity and go with claude pro
Claude pro is losing claude code access lol
For clarity, we're running a small test on ~2% of new prosumer signups. Existing Pro and Max subscribers aren't affected.
Im not interested in coding, but the limite of 10 archives or images at the pro plan is absurd
Yeah fake asf
girl what
Claude doesnt have image gen
Im saying theyre lying to save their ass
Its not generation, but anexing images and archives
like when you send a image at the ia chat
Yes, and read and analise documents and etc..
Thats a horrible use
Just use a real OCR
Theyre specifically made for that
Are you trying to index and save/ translate documents or are you trying to find information/ summarize the documents?
i work produciong documents and presentation for academics... So i need to anallise information, writing stile, read articles... and so on
Also if they were just doing it to 2% of people they would have edited the clause, not remove pro support entirely in their documents, or websites
So i annex these documents at the a.i chat and ask it to sumarise and produce a resume, presentation, etc...
I feel like you can just do all that using the free tiers of multiple LLMs
I can at some point, but i work with medicine and dentistry studants... so at most of the time i need a "smarter" analisis
Waiting to see if they actually take it away from existing subscribers as right now I still have it
They wont start doing that till the start of the next subscription cycle
Anyone knows something to do abt it?
They wont nor can take it from you since you already accepted the TOS that dictates you have access to it
So if you achired the one year plan, they will just remove next year?
Why would they wait until the next cycle? They release and remove models and features all the time inbetween cycles? Everyone's subscription cycle is different so that makes no sense lol
You reched the upload limit
Are you dumb or?
I know, but how can i see when it resets?
They cant just change what you have access to mid subscription
That would violate the user agreement
They legally cannot do that.
Really, then how are they removing access to Opus 4.5 or Sonnet previous models
Meaning the only time they can is when your subscription and IE your user agreement ends.
They most likely dont dictate what exact models you have access to in the TOS
As they update all the time
Meaning they would have to update their user agreement everytime
Most likely they generalize it by saying "Anthropic models" or "latest models from anthropic"
So you think Claude Code is the exception to this?
Legally that would be silly. But hey if you can point to where thats the case in TOS I will believe you.
Once again I have to ask are you dumb? Like why are you talking out of your ass?
I mean making personal attacks doesn't help your case.
Its silly for a company to make a new feature, resulting in them having to add that new feature to the user agreement?... Like do you hear yourself.
Youre deflecting
read the tos yourself š
They can if they want to. Would lead to backlash, but nothingās stopping them
And no current paying user is losing access to it as said already no matter what you might think