Steven

@MV33Racing Not happening. That’s why Toto said they’ll talk to him after since he said if we race that way then so be it. Toto will not accept that.

Kimi is absolutely fuming. He now sees the kind of sneaky tricks George pulls off. Next time he should throw the elbows out as well and give George a taste of his own medicine. Apart from the opening laps, it was unfortunately a pretty boring Sprint race.

@JJdoesTech @rezoundous This is what I learned from one project and trying to build the UX design first for the other 2 projects so I can go faster with me being a little less involved beyond some QA

@rezoundous You go super hard in concept phase and then sketch out exactly what you want to do. Then you let an agent handle everything after that and only have it return when fully finished. Bonus if that agent is the boss agent, sends out subagents to then implement slices, etc

Running 3 parallel Codex sessions and I'm already feeling unproductive with the context switching. How is anyone running more parallel sessions?

@thsottiaux Codex 😅

How do you refer to it

@CameralisX @hxxwhite This. Makes no sense to pay something that is simple to do locally.

@hxxwhite Why pay when you have all the resources you need locally?

Serious question for other mobile devs: Why waste compute running Xcode sims locally, when you can just stream them from the cloud?

@kaaaash____ I have to say this has to be a lie. I personally don’t want to do any kind of farming work or lawn work. “Every” is misleading 🤷‍♂️

Be honest, Why is every software engineer’s backup plan always farming?

@Kyle99338388 @nikolasbydesign @husain_j53 @theo In my 10 years of experience the lowest I’ve gone personally was learning C. I have never seen anyone recommend learning or looking at asm (Assembly Language) 😅 I agree with your 90% and would possibly push it to 99%. Fortune 100 company devs, I never saw anyone doing it

@CoderSteven @nikolasbydesign @husain_j53 @theo He missed the point entirely - when I said under the hood I was referring to the asm that actually is the "code" not the high level abstraction. 90% of devs (prob more) do not review the asm which their code compiles into.

You guys still use IDEs?

@Crytonewb1 @theo Prompt -> you are a top UX Designer Built UX -> how app will look Codex/Claude code -> actually implement it Review -> minor changes but should have used your UX designs as reference to get nearly 100% one shot

@CoderSteven @theo No...not true ...when you actually are a front end designer, there is a lot of "maybe 1px, maybe 2 maybe 4, let's try 1 again" prompting this is very slow....fine tuning is slow by prompt.

@Pant_Siddharth @theo Never knew that was Cursor own ai until now. I’ve saw it but just never paid any mind to it. Use Codex/Claude models then realized it was pointless to continue laying when I already have codex and Claude membership which I use way more

@CoderSteven @theo Composer 2.5

@nikolasbydesign @Kyle99338388 @husain_j53 @theo And “good luck”. Btw I also agree not everyone knows evening. As a junior I tried learning everything in a fortune 100 company and it was just MASSIVE. Well with AI I develop and work completely different now

@nikolasbydesign @Kyle99338388 @husain_j53 @theo I have to say you must have some solid documentation and a solid system because I was given a SHIT SHOW earlier in the year and if we didn’t have AI it would be impossible to work on. No documentation, no ex engineer present. Literally just tens of thousands of lines of code

@truiycr7 @ga15735 @dxbestani Not his fan but agreed on this one. If this is Bahrain, Lewis was cutting forgot which corner which gave an advantage. It was allowed for 20-30 laps before the FIA told Mercedes they had to stop it. Spain was such a brilliant strategy which RBR had to copy later

@ga15735 @dxbestani verstappen didnt really challenge lewis, i said it was a good drive but not a BRILLIANT drive. Lewis had better victories in Spain, Britain, Russia and the final 4 races of the season. Only ranks ahead Portugal

Hamilton had absolute no right to win this race. That Red Bull was like 3-4 tenths faster than the Mercedes. His tyre management, the solid defensive driving for once and that last stint was world class. God I miss this Hamilton so much.

@guerilla_fit @quickst8tus @elonmusk You should do some actual research how useful the moon resources will be for us

@quickst8tus @elonmusk Nothing that would be useable here on earth. Literally just stuff that would help them build…lunar bases. lol

🇺🇸🇺🇸 This will be awesome!! 🇺🇸🇺🇸

@elonmusk @elonmusk can you please just lead the way? I grew up loving NASA and space but as I got older I noticed @NASA is incompetent and we cannot depend on them for space expansion or exploration.

@JorgeCastilloPr Then how does codex exist in the ChatGPT app….

Technically this app will never exist for iOS. Apple is banning all vibe coding apps.

@Sourabhsinr @CodeWithAmann This and personally anything touched by Facebook (can hide under meta) I wouldn’t trust even if I’m paid 🤷‍♂️

@CodeWithAmann meta ai has no place in my dev workflow 🫠

Be honest As a developer, which AI model is the most useless here?

@droidbuilds What developer? I haven’t seen a windows developer ever by choice. Usually a company forces you to use a windows laptop due to partnership with windows products (ugh)

if macbook is so powerful. why do most developers still use windows?

@KaiXCreator Claude, ChatGPT and basically all AI agents are done/down for the foreseeable future 😃

I'm a vibe coder, scare me with one word.

@trikcode Yea when I did this I first asked for it to classify the files that are critically long like 600-800+ then worked on fixing each one individually while making sure it still works lol We can’t just go 100% blind with AI. Maybe 70% 😜

i asked claude opus 4.7 to refactor a large codebase. 68 minutes, millions of tokens burned - it finished nothing worked. app completely broken

@aryanlabde @cureteurazvan lol…I’m running 3+ projects on codex which doesn’t ever hit 5hr limits while Claude I need to take breaks or be massively token sensitive I wonder if they are running A/B test on people with different token usage 🧐

@cureteurazvan Fr man, codex limits are like CRAZY

As a vibe coder, which plan are you on right now? - $20 codex - $100 codex - $200 claude