@Pulko33@RadisIrradie@CleoreYTB T'inquiètes pas ça arrive 🤷♀️ Yoann vas apprécier 😁 lui qui a si peur de moi pour une raison qui m'échappe 😂 sûrement car il sait le nombre de dossiers que j'ai sur sa tronche où il admets avoir manipulé tout le monde notamment pour du fric aussi. Hm 🤔🤔
@Juan_RD@Warcraft yeah husband had this issue as well, but in few hours he managed to create one, so dunno, but they're kinda forced to do so, too many peeps playing hordes tbh
@kshgrvs@vennbrd hey, if you really would like that extension to be back, i could try and find a way of doing it, am a developer myself since over a decade so it's no big deal, especially if the code is already there and they agree to share the code freely for non-commercial use too. or remake it
@buildwithrajath if this can help, i can show you some of my stats and such, I got over a dozen billion tokens used so far only but have tasks going up to more than 20 hours etc. Just dm me about whatever you'd like to see.
@buildwithrajath Yes it does feels like, i just experienced with it, and i used to be around 15% when they did, then received 2 reset tickets, went up to 31% and now tasks are longer for less usage for real, one task that usually takes 5% barely takes 2ish now. (use it daily btw)
@2go_AI@rezoundous are you dead serious? uh? Is that it? I know many Chinese people that are using Codex heavily, no wonder why then.. That and Grok apparently is go-to models. Except if they can grab some third party apis of Claude they rarely do sub, I figured.
@ToBeOrWhat@rezoundous Never did for me, with the prompt engineering I used to do, over 4 months+ of x20, I also learnt a lot about its behavior and could fix that issue out, my main proj is like 300K+ files & 40K folders, 100GBs+, and it was fine as long as being careful with prompt & claude.md rules.
@ZenZoneOn@rezoundous Nah about this it tends to be fine, as long as you're not too restrictive on your prompt engineering or so. But it tends to not see the big picture time to time unless you really do ask for it out. But he's quick at auditing and fixing his own mistakes on-the-go, astra,a bit less
@itsmattblu@rezoundous I agree partly on design. Natively Codex does better, but if you were to use Claude's design tool to go further, it would kind of beat Codex unless you can afford a third party plugin ofc. Artifacts tho I agree, and when I switched from claude, SSH tunnels weren’t on point either
@ronbmp@rezoundous THAT! And more-so the gap between Fable and astra and more-so when astra had just gotten out, that was some insane gap. And another reason for me was how natively, Sol/Astra are able to land beautiful visuals/UIs without third parties, but I miss Claude's design tool still.
@2go_AI@rezoundous Dang, may I ask what you were working on? I worked on various stuffs even done security tasks and audits on friends projects and so on, never had issues, not even when I was fixing Discord's selfbot (that were not meant to grief/nuke servers but to automatize legit tasks though)
@rezoundous@JahsehLiu They only blocked some tasks with Fable 5.0 personally lol, and made it auto switch to a lesser strong model yet my project did not include any security related task, but it believed so, and I needed Fable cause I needed the best code quality and logic/math there out.
@rezoundous I done the other way around and the reason for me was how limited claude ended up feeling like even at x20 (wasn't about usage anywaays, rather, me feeling like I capped the models intelligence for my main project, which over 4 months in x5 then 3m on x20 barely improved it
@AdamHoltererer Felt like more. Still kinda like giving out 11 weeks, for free, in terms of usage, which means we basically got nearly 3 months of usage given for free, which is huge imo.
@denizumutdereli@thsottiaux Only specific models for specific tasks would beat their, like ling-fin that may be better for trading-related task or even / a bit better for 1/100 of cost So y idk. Sol 6.1 and more-so astra are huge stuff, better than claude atm, out of own experience ofc.
@denizumutdereli@thsottiaux But with all resets given, lately, to be fair, one month of pro is like, 3-4 months of actual usage at least for me it is. It's like I paid for a higher tier or something at this point. So it remains kind of fair. And I wonder what model you're even talking about tbh.