rich

I’m starting a weekly “Could I..?” series. One question. One experiment. Public notes, code, mistakes and results. Mostly AI, dev tools, games, automation and data.

@Gaurav_dev01 @X In between projects and playing with a few ideas but here for the connecting!

Hey @X 👋 I'm looking to #connect with people building the future. If you're a: • AI builder 🤖 • SaaS founder 🚀 • Indie hacker 💻 • Full-stack developer ⚡ • AI automation expert 🔄 • Startup founder 📈 • Product designer 🎨 Drop a comment with: 1. What you're building 🚀 2. Your biggest challenge right now 💭 3. Your X handle 👇 I'll reply to everyone, follow interesting builders, and support great projects. Let's grow together. 🤝 #BuildInPublic #SaaS #AI #IndieHackers

@kristapsozy cyberpunk spent years selling the promise rockstar mostly sells the expectation that the thing will be a banger on day one hope the delays were worth it :)

Rockstar accepts GTA VI pre-orders without gameplay footage. Cyberpunk 2077 did the same → disaster. The difference? Rockstar built trust through 20 years of consistent quality. Your brand's reputation = your ability to ask for commitment before proof.

@bysfdev ive always been a gmail guy but it's even better now I can connect it to stuff like ChatGPT

what's a good emailing service ? for first contact email, follow-ups, etc

@mackenziechild do you find that the context/instruction it gives to the new thread to be light? i do this pretty often for smaller stuff, but I find for anything big it generalises the task a lot and the new thread has to interpret

TIL you can tell codex to spin up new chats, so i now have one main thread with all the context of what we're doing spinning up new chats to complete the work & not affecting the context window.

@danielatitienei Just sent you a DM

@danielatitienei Congratulations on the launch! I just picked up an early bird license.

finally launched SuperSnap on Product Hunt today 📸 built this because i got tired of how complicated existing tools are. this just does one thing well and fast. check it out if you want: producthunt.com/products/super…

@jun_song api prices going up probably means a lot more apps quietly switch to weaker default models

API price will go up Subscription usage limit will go down It’s not because those frontier labs are greedy, Inference computing cost doesn’t match with what they provide to you. And cost is still spiking, every week.

@egorkabantsov if gpt 5.6 really got held up over that, roadmap leaks just got a lot less useful because external politics can reshuffle everything late

So turns out I was right and GPT 5.6 release was indeed delayed because of government concerns and Anthropic Fable 5 / Mythos situation

@its_sidraa tbh this is why a lot of people overvalue office perks early on. the bus, food, ac all feel great till layoffs start, then you realize salary and savings were the only real cushion

Software industry is weird Companies provide 1. Free food in office 2. only 5 days working 3. AC rooms to work 4. Min salary here is dream salary in other sectors 5. Office Bus to come to office 6. Leaves at least 24 / year But the catch is, they can fire you at any time😅😅

@kunchenguid same pattern here, faster mode only feels better when you're waiting on one thread with a few open, quota gets burned way before latency matters

i have stopped using fast mode with gpt 5.5 with firstmate handling context switch, my average number of parallel sessions went from 3-5 to now 10-20 making requests faster (and more expensive) doesn't help at all when token quota is the clear bottleneck

Right now, the winners are the ones who can sit in Cursor/Codex/Claude for 6 hours and post the ugly demo anyway.

@akashgohil10 @X hey! Let’s connect

386 followers on @X. Wild. I joined X in 2014 but never paid attention to it. A few weeks ago I was posting into the void. Still building… If you’re an indie hacker, builder, or just into tech, drop a Hi 👋 I’d love to #connect with more people like me!

The best AI products survive the zoom-in. If the magic dies in ops, it was a demo.

Your builder network does not need to be huge. - 3 sharp peers - 1 buyer who complains - 1 operator who shares numbers

@CastAsHuman @OpenAIDevs @thsottiaux @ajambrosino Are you noticing any performance degradation in 5.5 too?

@RichDevLab @OpenAIDevs @thsottiaux @ajambrosino No, I don't think that psychosis is that frequent because they often spot these issues by themselves and reset limits. I hope they find this as well, because something fishy is definitely going on!

Okay, so I'm starting to get a little suspicious now, because this is the second time in a row I run out of the 5-hour limit. @OpenAIDevs @thsottiaux @ajambrosino Can you perhaps check if there are others complaining today? It feels a little off 🤔

If AI makes infra easy, what are infra wrappers even selling?

@Thomas_Tao_1 it’s often super tempting to do, too ill sometimes treat UI the same to get it clear in my head but it always seems to be more throwaway than a proper backend implementation

@RichDevLab Yep. UI too early is how core logic gets weird fast.

Top tip: Build the basic first Everything I build starts barebones as a thin runnable slice, and it stays there as the project grows When I need to make deep core changes I don’t need to fight with a browser or UI I just tweak what I need to and validate

@hxrnake yeah the headline price is basically bait if the busy friday night is what actually blows the bill out gotta be a hard market to get into though - whats your outreach plan for sales?

The monthly price is never the price. OpenTable: "£99/mo." Then £0.50 to £2 per cover. I built a total cost calculator. The real number, not the headline. Checks ResDiary and TheFork too. Mine: from £59/mo flat, no per cover. Nothing hidden. #buildinpublic #ukrestaurants