IngoA

@rezoundous Not for me. It's quite random TBH.

usage limits on GPT-5.6 feels good today. roughly the same as GPT-5.5.

@CastAsHuman @Awesome_O_AI Bot outta context.

@Awesome_O_AI Bro you high?

One thing that was removed from the codex app, and I really miss it, is the creation date of threads. First, I could easily see yesterday's threads, but now I have to manually check.

@CastAsHuman I think there were even timestamps on chat messages? It's confusing now, for sure.

@kedytcom Why not localize to everything and release everywhere right from the start?

I just submitted a new iOS app. But I did not start with the idea. I started with App Store search data. Found a niche with low competition. Then found ~15 markets where I think the app can rank faster after localization. Now it’s waiting for review. This is my launch plan: US first. Then localize into the easiest markets. Then track rankings. Then double down where the data moves. Most indie devs ask: “What app should I build?” Better question: “What search demand is underserved right now?” If you want, reply with your app niche and I’ll tell you one market I’d check first.

@wickedguro Yep, they always overdo it. Could just go to "Following"

Feels like the new algorithm killed X for business. Welcome to 2007 Facebook.

@peakcooper Today it's really shitty. Days before heaven.

The more I use GPT 5.6 Sol, the more I dislike it. It is not reliable at all. Responses fluctuate between "extremely good" and "catastrophically bad".

@reach_vb TY! 5.6 is incredibly slow + stupid today (Europe) - do you know anything?

codex tip: ask codex to do its research first and then use set_goal to set an appropriate goal instead of /goal It results in massively better prompt and downstream results! all my prompts would start with requirements gathering and research and once done set_goal

@Nominatiivi @macintogdev @reach_vb Ha, also have a dotfiles repo.

@macintogdev @reach_vb Love this. I turned the pattern into a reusable set-goal skill: research first, then turn the findings into a concrete objective, constraints, and proof of completion before handing off to the native goal tool. Much better downstream runs. github.com/liby/dotfiles/…

@thsottiaux Today 5.6 is lazy, slow, and stupid (Sol medium as base). So bad that I had to turn back to 5.5 high... So no tokenmaxxing day 🥲

Tomorrow might be 8M active user celebration day. Just saying

@sama 5.6 is lazy, slow, and stupid today, WTF?

2.5x increase in usage of our agentic products (codex and chatgpt work) in the last week! welcome.

@erbybody API v2?

This applied even before AI but... make sure your backend data model is well thought out. You can change the UI without issues but it's a nightmare to try to maintain backward compatibility, especially with apps where you simply cannot force immediate client update like you can on web.

@vicentesandev @heyandras That’s the idea of ultra

@heyandras same with ultra

GPT 5.6 Sol with High reasoning is so weird, it uses subagents all the time for not particular reason. The model is good, just this auto-subagents-everywhere-everytime is a bit annoying tbh.

@ben_ai_eng This. x.com/ingoa_dev/stat…

Removing the 5-hour Codex window has honestly been a bigger qol improvement than i expected. I always end up working well beyond 5 hours anyway, so tracking both the 5-hour window and weekly usage just created unnecessary context switching. Now i can just focus on problem solving. Less than 48 hours in, it's already reduced my day-to-day stress during a critical production launch. Hopefully this stays in place. Nice change, @thsottiaux @OpenAIDevs 🤝

@GergelyOrosz They blew it so hard.

Cannot see any sensible company use Grok CLI based on this terrible incident If you are an amateur (meaning you don't make much or any money from your code) and do not care about security (aka you don't mind if your .env files leak - which you should care about) then use it sure

@verry_codes Good luck + keep them drums coming!

I'm finally jumping into marketing on social media. I decided to not follow the AI-slop-marketing train. Instead I'm trying to create value, I'm making my own videos, being my own influencer. Of course this can fail massively and I'm in the total beginner mode now, but...

5.6 is fucking lazy for me RN. Not good.

@TerekhinIvan Thanks, good info!

I was scared to raise prices on my Android apps. Kill threshold I set beforehand: if conversion drops more than 20%, roll it back. Raised Android ~25% twelve days ago. Conversion went UP. 2.99% → 3.21% transactions per new user. Revenue per new user: $0.31 → $0.40. One refund in six weeks. The threshold I was so worried about was never in danger. Honest caveat: installs rose 14% in the same window on their own, so part of the revenue jump is not price. That is why I measured per user, not total. Now, I will probably raise them again in the next month, to test the elasticity of my users. And will bump the monthly and weekly prices too.

@Blackgentoo @OpenAI @thsottiaux I hope not! 12h would already help.

@ingoa_dev @OpenAI @thsottiaux well i guess it's just going to be temporary though

The 5-hour limit gone is easily one of @OpenAI’s best releases! @thsottiaux

@kevinkern Why not Sol high with oracle option (if it really gets hard)?

gpt 5.6 pro is even more powerful for creating full initial specs for a new project. my workflow here: 1. use gpt pro to discuss your upcoming project and create the full specs as a downloadable .zip file 2. start a codex sessions and tell it to download the specs from the chat (its able to do that) 3. I'm using planr here to create a full task graph of the project 4. Tell codex to start the planr loop and it will run for some hours to achieve the goal

@levelsio I feels so mutual now!

Hello mutuals 👋 Nice to see you again after 3 years It's great to have you all back on my timeline 🥹