That’s what an AI would say. 
2 Likes
If you want to connect then insinuating that I am not a real person will not help.
By the way: let’s keep this conversation on-topic.
1 Like
It was meant as a joke - sorry for offending you, I didn’t mean it. 
Also, good idea!
I think them releasing GPT-5 is unlikely, depending on what “GPT-5” even means to people.
If it just means “5% smarter than the previous model!” then yeah - sure we might get GPT-5.
Thinking about it logically, I don’t think it will ever come out seeing that OpenAI is now working on AIs that have a whole other logic programatically speaking (Thought processes of the o1-family).
Newer models will either expand on newer concepts of technology or introduce completely innovative and new concepts of AI all together.
Maybe we’re getting an AGI-1 model in the future! 
1 Like
Who knows?
Maybe the next model will be a “4.5o” instead of a plain “5.”
Mostly, I’m hoping for a better training cut-off. It would be a big and welcome surprise if a new state-of-the-art model drops this Christmas.
Here’s to wishing!
6 Likes
- gpt needs i mean REALY needs to be state of the art with the api of it self.
and second i believe we really can put a fixed date on when agi will apear and we will see gpt to be really smart.
its obvious that openai is on a tight calculation power limit.
they could perform much better in lot of tasks if they hat more server power.
you cant run sora AND o1 AND all the other stuff and expect that everything will work out on best performance.
They cant make the best of what they have.
so we will see much more better stuff in as altman Said thousands days as he says in bet like 3 years.
open ai is building a super computer
till 2030?
so as soon this super computer is ready to use we can expect big things.
till then we will see some minor enhancements and multi media and cross content stuff. like real video interpretation (what we got right now is absolutely not video understanding)
co working on office stuff it will be just get more capability but for real smart stuff you may remember the shark an whale compareson altman did in one presentation?
ai is young in historical context developed like a flash but now we need to wait till our biggest calculator in the world is completed.
even if its painful.
1 Like
The link and description for today’s event have been updated.
They’ve been placing a new date in ChatGPT. So far, this only helps the AI produce counter-factual hallucination (and prediction as a side-effect).
A “do you prefer” model that likes to write with lines between its list element blocks …
Looking like filler release today.
Are they not knowing how telephones are working?
It is impressive that they have made ChatGPT available toll-free in the United States.
thats actually not absolutely dumb if you’re in a region with instable and or super slow Internet.
1 Like
i already thought about doing something like that for myself but with sms not WhatsApp
Only real use I am seeing for this having chat on 3way call to help with customer support or translation.
1 Like
China → Amazon → You → 1-800-CHATGPT for 15 minutes per month (caller ID, maybe looking up your unchangeable account phone number)
Then pick up another handset on the line to record, with the analog interference. Then get the phone guy to put an analog card in your VoIP PBX system or what have you and provision for direct dial.
Do I get the bug bounty now for reporting using a T1 to set random ANIs for CLID as authentication and account probing attacks?
You use lot of technical terms but I am thinking you are not right on this. Automatic number identification is fake not easily.
It has a technical bar to understanding and possessing the services needed to spoof the higher bar of inbound toll-free numbers, why exposure is limited by giving a heads-up, but I can assure you, there is surface.
Proof is doing. I am thinking exists easier manner to be accessing user informations? Only vulnerability I’m seeing to users is if model is having access to memory of what user say to chat? Or maybe less use for them? ANI is not being same thing as CLID. Because its used for billing more accurate and services faking it can be shut off of network fast.