Joe

@martinuke0 @adahstwt Totally. And with AI, no need for Postman anymore at all.

@adahstwt curl? duh

hey devs, do you know any alternative of Postman... really fed up with this

@ravikiran_dev7 Avoid the pain and wasted time and switch to a Mac.

I am a window user and I just installed Linux for the first time. What should I do next?

@ekidhaja @dev_maims This is exactly why I use one..

It’s not hype, it’s the balance. MacBooks give you a Unix-like environment, strong performance, solid battery life, high processing power, and hardware that just works. You spend less time fighting your machine and more time building. Plus most dev tools are built with Mac/Linux in mind, so everything feels smoother out of the box. And it’s also very aesthetically pleasing and sleek

What is the reason behind MacBook popularity among engineers?

@adamheimann @cursor_ai My take is exactly the same. Really impressed.

.@cursor_ai absolutely cooked with Composer 2! I spent most of the day using it on the most complicated porting job I’ve ever done, and it has been nothing short of adherent, accurate, thorough, and fast. The best way I can describe it is this: it feels like a model above Sonnet 4.6 and maybe slightly below, if not dead even with, Opus 4.6, at a fraction of the price and with better prompt adherence. WELL DONE! @ericzakariasson @jediahkatz @JuanRezzio @benln @leerob @mntruell @bdc @daniel_adornes @hugocharre @_nicole_zhang_ @mshwery @wilsonzlin @KSHartnett @SamSokolin @edwinarbus

@theo Who’s freaking cares! It’s faster, very good and the cost is dramatically cheaper. Seriously, who cares about the drama.

Cursor's new model might not be all that new. Did my best to break down the Cursor x Kimi drama

@jacobferus @iannuttall It’s ridiculously cheaper..

@iannuttall Is composer 2 pricing even close to claude code or codex?

new workflow for the weekend: - gpt 5.4 xhigh to plan - cursor composer 2 to implement - back to 5.4 xhigh to audit + fix - ship pull request - repeat

@VibeCoderOfek @iannuttall Not sure about 2x but definitely faster at a huge reduction in cost. I’m so happy I tried it.

@iannuttall Hybrid workflow gang rise up! This cut my iteration time in half yesterday. Anyone else hitting 2x or more with this exact flow?

@iannuttall @mhmazur Composer is faster and A LOT cheaper.

@mhmazur Yes still using conductor! I generally just use 5.4 for everything but composer 2 is really fast and cheap so I’m trying it out

@iannuttall I switched to this last Friday and it’s great and my token costs have plummeted. Honestly I’m quite amazed how good Composer is now.

@sparkycircuit @RealCandaceO @LangmanVince Weak reply. Come on, you can do better than that!

@JoeConlin @RealCandaceO @LangmanVince your ugly mother is a dem plant so stfu!

Must Watch! An Ex Candace Owens employee reveals what a narcissistic evil "b*tch" Candace truly is who "knows exactly what she's doing and it's f*cked up!"

@sparkycircuit @RealCandaceO @LangmanVince And Candace is a Dem plant so it’s even!

@RealCandaceO @LangmanVince Candace, “Vince Langman” is not a real person. It’s a russian troll farm bot account. Do not engage with it.

@robinebers @stepangrit You don’t need frontier models for everything and stop using big portions of the context windows. I can code all day and Max covers the month.

$200 actually isn't actually that bad the problem is that $200 isn't enough I spent over $2k Cursor in the past 30 days (caveat: because I do what I do, a lot of AI companies get me that access for free, including Cursor) but looking at this usage, I just have to think about my clients and ask: "would be willing to spend $500 or $1,000 a month on this at this point?" I'm honestly not so sure anymore

most people don't know this but... Cursor beats most harnesses because it has features that make your AI significantly (!) smarter than other AI tools the biggest three are probably: 1️⃣ CODE INDEXING Cursor indexes your entire app like a library indexes books, and lets agents ask plain language questions like "where do we talk to stripe" or "what is the database service" and Cursor answers these instantly this means your AI doesn't need to read 37 files, saves context window, acts faster and smarter - and Cursor does this for FREE btw (!) despite costing them a significant amount of money on the backend 2️⃣ PLAN MODE Cursor's plan mode is by far the best one in the industry - not only does it ask really good questions, but it also works across every model in the same way, uses plugins/mcps/skills/etc, but also easily lets you edit the plan before you execute it, and switch models between planning and implementing you can even create a plan with e.g. Opus, then ask GPT-5 to review it, and then let Gemini implement it (I don't know why you would, but you could) 3️⃣ PLUGINS so there's skills, mcps, commands... 🫠 nobody knows what the difference even is anymore or what to use when - so Cursor thought: let's make this ONE thing, call it plugins, and call it a day now they do only make it super easy to use them, but they even suggest them (!!!) if it knows that there is a plugin available in the marketplace that you're not using yet, and it thinks it would help TL;DR: Cursor is THE BEST agent to code, bar none this is based on thousands of hours, comparing tools, building apps with thousands of daily users, no longer reading code, and seeing hundreds of students and clients succeed NO, Cursor isn't cheap NO, if you're a student or indie builder, this is probably not for you but YES, it is the best if you have the funding to do it

@GergelyOrosz Boy enterprise dev if the can’t even get a Max acct.

I am hearing tons of complaints from Cursor customers at enterprise companies: A silent change put almost all models Cursor uses behind Max mode. Devs who used to manage to “spread out” monthly credits over a month see all of it used up in 1-2 days. Are furious + switching.

@trikcode Funny seeing all these people bitch about Cursor using Kimi. I’m a pro dev coding all day and switched from Sonnet on the majority of my tasks during the day and this is faster, better and much cheaper. Not sure what the problem is.

Tool companies like Cursor are starting to look less like wrappers and more like real AI labs.

@outsource_ @trikcode Wrong…it def better than just Kimi.

@trikcode No they just used Kimi 😂 BIG wrapper

@jc_gilbert @trikcode And your point is?? I’ve been using it for a day now and it’s fast, accurate and much cheaper than Sonnet, which I use for much of my daily code tasks.

@trikcode except it’s based on kimi k2

@david_nix Tired of these posts. It’s NOT the tool, it’s absolutely shit prompting. People, LEARN HOW TO PROMPT CORRECTLY! This will save you time, money and most importantly, improve the quality of what you’re building.

My experience with Cursor. "Write this code for me." Lookin good bro! Look at all this time saved! Run it. Wait... doesn't work. Wrong in subtle ways. Spend more time debugging than if I wrote the code. Who likes debugging more than writing?

@kapilansh_twt Might get better?? Think about where we were 6 months ago, or 12 months.

@JoeConlin ai might get better but ownership still doesn’t get automated

vibe coding is just a fancy term for "I have no idea what my codebase does" → AI writes 400 lines → you don't read it → it works → you ship it → 3am production fire → you have no idea where to start → ask AI to fix it → AI breaks 3 other things we're not building faster we're just breaking things at the speed of light and calling it innovation

@leerob @ClementDelangue Who cares! People bitch about anything nowadays. I’ve been using since yesterday and performance is great at a big drop in cost. Hopefully it helps drive down cost on the others. For now, I’m happy…

Since people really want me to say this: "KIMI K2.5" ‼️ Yes, that is the base we started from. And we are following the license through inference partner terms (e.g. Fireworks) I'm thankful for OSS models personally, good for the ecosystem.

Yep, Composer 2 started from an open-source base! We will do full pretraining in the future. Only ~1/4 of the compute spent on the final model came from the base, the rest is from our training. This is why evals are very different. And yes, we are following the license through our inference partner terms.