Jean Pierre

[Jean Pierre @JeanPierre50284]

@RayyanAamir @sevenracin @plevenansh @Sir_Franco_ @thsottiaux

[Jean Pierre @JeanPierre50284]

@RayyanAamir @sevenracin @plevenansh @Sir_Franco_ @thsottiaux https://t.co/BNG4f4C6Ft

Thats just facts, you can check it here. OAI lead in alignement...

[Jean Pierre @JeanPierre50284]

@thsottiaux Thanks

[Jean Pierre @JeanPierre50284]

@sevenracin @plevenansh @Sir_Franco_ @thsottiaux Claude is intended for enthusiasts and those who want to do things just for fun.
OAI is intended for professionals.
Thats what ppl forgot when comparing the limits of 2 models

[Jean Pierre @JeanPierre50284]

@thsottiaux I hope we gonna get a reset Today, editing a lot of text with my few percent since 2hours, seems to be enough for this task, but it will not last forever. Bless me St. Tibo

[Jean Pierre @JeanPierre50284]

@hatihakuti @LuthandoNondumo @thsottiaux I think the problem runs deeper than that. We have two different outlooks on life: you’ll never be satisfied with what you already have and will always push for something better, whereas I’m grateful for what I already have, and if something doesn’t suit me, I don’t blame others

[Jean Pierre @JeanPierre50284]

@hatihakuti @LuthandoNondumo @thsottiaux People forget how recent our freedoms are. Things can always improve, including here in France (my country), but complaining 24/7 makes us lose perspective. We can push for better while appreciating how far we’ve come. Living in peace shouldn’t be taken for granted.

[Jean Pierre @JeanPierre50284]

@hatihakuti @LuthandoNondumo @thsottiaux I’m not defending OpenAI’s leaders. I can still see their side. Making money isn’t wrong, work deserves pay. Criticize what’s wrong, but give credit where it’s due. They’re trying to improve things. Why pile on more pressure?

[Jean Pierre @JeanPierre50284]

@hatihakuti @LuthandoNondumo @thsottiaux You act like OAI are actually trying to fck Plus users, why they just do their best. How are ppl unable to understand that life is not a movie, you cant get free infinite computing power, you have to make choice as a company https://t.co/foGZdTrt8u

[Jean Pierre @JeanPierre50284]

@hatihakuti @LuthandoNondumo @thsottiaux Im attacking Plus users???? How? Its just annoying to read cry babies all the time while they do effort. Plus ok, is not as good as Pro deal, but still extremely good compared to the real cost, OAI actually spend more money, than they earn with subscriptions, put some respect..

[Jean Pierre @JeanPierre50284]

@Sae_Ria @TheWrightRavens @Aditya_cuz00 @thsottiaux 95% of plus users prob dont even know what Codex is, its like all the ppl subscribed to a gym who go like twice a month. If you're not happy you can just unsub from ChatGPT, it will not change anything for power user, it will even give us more speed

[Jean Pierre @JeanPierre50284]

@hatihakuti @LuthandoNondumo @thsottiaux Bro, thats normal?? You pay more upfront, you get a better deal per token. Ever heard of “3 for the price of 2”? Same principle. If you wanna pay for exactly the tokens you use, use the API. But that would end up costing you way more than a Plus sub. Plus is still a good deal...

[Jean Pierre @JeanPierre50284]

@hatihakuti @LuthandoNondumo @thsottiaux Its not because Pro have received increased limit, that it make Plus a bad deal. You already got access to CRAZY computing power with Plus. You would have paid 500dollars to be using current Plus 3 years ago

[Jean Pierre @JeanPierre50284]

@montasees My bad, wasnt aware of the subscription cost difference (from your screenshot). You're the first person I've seen back this up with actual numbers here. Most complaints I'd seen had no sources, and I was relying on API benchmarks. I jumped to conclusions.

[Jean Pierre @JeanPierre50284]

@TonyDingjxl @thsottiaux Thats just my opinion lol, how is tibo involved here?

[Jean Pierre @JeanPierre50284]

@montasees @thsottiaux Fair, you said Sol, not Astra. I shouldn't have assumed your setup. I still disagree that Sol can't compete.
I can be wrong but i hear a lot of ppl saying Claude codex is better than Codex, but never seen any proof of that, while benchmark does give me factual infos

[Jean Pierre @JeanPierre50284]

@TonyDingjxl @thsottiaux Then switch to Claude what you want me to say? If you use Luna/Sol for most of your task and a bit of Astra for complex one you should be fine. If you're not happy switch to claude, and let OAI cook for ppl who believe in them

[Jean Pierre @JeanPierre50284]

@montasees @thsottiaux It just that a lot of ppl complain when OAI actually do effort, its kinda unfair. Also TPS doesnt represent output quality or speed? It just mean the model is using less token to think. Well OAI have capacity issue that slow the computing for sure, thats a point to work on

[Jean Pierre @JeanPierre50284]

@montasees @thsottiaux Skill issue mate. Sure if you ask Astra for something Luna or Sol can do you will burn your token. Not the fault of OAI if you cant manage correctly your workflow and your usage