Aayan Rahman

Assalamualaikum. Stryke started four years ago as a small curated marketplace in Dhaka. What stayed with us wasn’t the products. It was the unglamorous work behind commerce — trust, onboarding, support, seller workflows, and where things actually break for small businesses. We’re now turning those lessons into practical AI tools. Not a flashy pivot. Just a clearer direction. Building from Bangladesh.

@rezoundous Will it be good enough or equivalently great if I use it through OpenRouter?

GLM 5.2 = Opus 4.8?

@LyalinDotCom Love this idea

I think the idea of Forward Deployed Engineer (FDE) already feels kind of out-of-date I think the future is Forward Deployed Staff

@sflorimm Very true. Well said @sflorimm

hot take: in 2026, you're not a founder until you figure out distribution.

@TimJayas I still feel that ChatGPT's image model is better. Or maybe @TimJayas , I am a bit old and my information has become stale since Google I/O. Should I give a try to Gemini?

I get it bro, Gemini has the BEST image generation tool But you don't need to create an image every time 😭

@argofowl @argofowl rn

i can't do this anymore chat how does anyone do any work with claude these days? it's been so unstable after the fable ban i need her back

@thsottiaux The real leverage is not “which model wins.” It is building workflows where models become interchangeable engines: frontier models for hard reasoning, open-source models for control, privacy, and cost. Codex becoming model-flexible is a big deal.

Reminder that you can use the Codex App, CLI and SDK with any open source model, not just with OpenAI models. #oss-mode-local-providers" target="_blank" rel="nofollow noopener">developers.openai.com/codex/config-a…

@piotrzkowski Appreciate it. I’m building Stryke — starting from Bangladesh, focused on making online commerce and AI-powered tools more convenient, reliable, and useful for builders, businesses, and customers. Still early, but building seriously: strykehq.com

@Aayan_Stryke Let’s go! What are you building?

I’m looking for serious builders, creators, AI learners, founders, operators, and people building from Bangladesh/the internet. Reply with what you’re building.

@alex_prompter I am happy with my GPT 5.5 and would be even happier if GPT 5.6 continued the trajectory of improvements of GPT 5.5

Honestly, I stopped caring about models since Opus 4.6 came out. It can do most of my tasks pretty well imo

@theohandsh I wouldn’t frame it as switching. I’d frame it as stacking. I’ve been mostly ChatGPT/Codex so far, and the unlock for me is the execution loop: turning intent into shipped work. If budget allows, use both. If not, pick the one that makes you ship faster.

I’m a Claude user. Should I switch to Codex or is it overhyped?

@KaiXCreator As long as one stays banned, it truly is pointless.

Hot take: The whole "Claude vs GPT, which is best" fight might be pointless

@egorkabantsov Sadly, this is what may possibly happen when the competition gets wiped out.

Since when does Codex just stop current task when you run out of limits? In the past, it kept going to complete your last prompt if you asked it before you reached limits. It was one of my favorite things about Codex :(

How do you approach and manage your usage rate quota on Codex? I am genuinely curious. Please help me out with some tips and recommendations.

@CaptainInsightX I won't because I agree with you. Anything that's available and gets the job done is good enough.

Hot take: ChatGPT is better than Claude. Change my mind.

@argofowl @sama Yeah! Ain't nothing like a healthy dose of healthy competition.

hey @sama go ahead and drop 5.6 today who cares about thursdays imagine how much this will upset dario - your #1 enemy in the whole wide world time to get back to mogging mr mogman

@_Smoke_pai_ ChatGPT

Be honest, Which subscription provides best value for money ?

@rezoundous Mostly because people tend to say positive things about those who are dead, and you very rarely see people appreciating those who are still alive.

Fable 5 is the first model I've seen that has near unanimous positive feedback on X

@haider1 Oh no

openai trying to survive june 23rd: release gpt-5.6, call it better than Mythos, and get banned immediately release gpt-5.6 and say Mythos is still better, then watch everyone roast it

Perfectionism is often just fear wearing expensive clothes. Founders waste too much time over-preparing for the perfect launch. I've fallen into this trap myself. Start before you feel ready. Post before it feels perfect. This is Day 1 of becoming harder to ignore.

@VraserX @yacineMTB I guess you won't have to worry or be sad about missing Fable that much if this is true.

GPT 5.5 already basically matches Anthropic’s Fable 5 coding performance at around 50% of the cost. That’s wild. If GPT 5.6 lands the way I think it will, Fable might become irrelevant almost overnight. The only real question now is will we all get access?