@theZachChen I think I’m just used to pressing the Fn key to dictate rn, like it became a standard and not configuring it for 20 different apps is one of the point. Even if it’s a simple program, cuts me headache.
English
Divanshu Chauhan
740 posts
Divanshu Chauhan
@Divkix
23, CS@ASU | Open Source Contributor | Intern @Cloudflare | Prev. @IBM Intern
Tempe, AZ Katılım Temmuz 2019
105 Takip Edilen272 Takipçiler
"can you make gpt-5.5 suck less at frontend design?"
turns out, yeah. impeccable 3.5 has *dedicated* ai slop prevention for gpt-5.5. all three got the same one-line prompt. you be the judge:
English
@0xSero Do u prefer generic-chat-completions-api or anthropic for byok? I’m still not sure which is stable and better. Generic one usually gives me double output, anything similar u experienced?
English
@sudo_overflow D1 is good but you’ll have to design in such a way it matches d1, it’s sqlite. If u swap for @PlanetScale postgres, you can make the stack portable. Just my take that not staying locked is a good practice but imo cloudflare stock is goated.
English
@antigravity any plans to add support for ~/.agents/skills dir for global skills? Really hate that I have to install once for all and once for antigravity to pick it up
English
@ValsAI can you benchmark Xiaomi Mimi v2.5 Pro? Been looking at your benchmarks to know how good models actually are.
English
@nadeem_jaleel Openclaw and hermes, I get they can automate stuff buy way too hyped for their capabilities.
English
What’s the most overrated tool every developer pretends they can’t live without?
English
@nathanv246 @huge_icons Haha, thank you! Trying to keep myself busy
English
Just found out abt lexiconbranding.com and make a skill, let me know if it's worth it or not.
Install via: `npx skills add divkix/agent-skills --skill brand-naming -g`
English
One of the best things I was waiting for! Time to try this instead of warp.
Zed@zeddotdev
Terminal Threads are live in Zed v1.3.5! You can now run claude, amp, pi, or any terminal-based workflow as a managed thread in the Threads Sidebar, right next to your other agent threads.
English
@nathanv246 @huge_icons Yeah I should probably update it, thanks for the suggestion!
English
@thegenioo We started already on Composer 2.5 before Kimi K 2.6 came out.
English
@mil000 It’s for cursor itself using their harness and prompts. Different models behave differently with different tools
English
Cursor bench is slowly burning any and all trust people had in cursor. I’ve never seen an eval that’s just a blatant lie like cursorbench.
Lee Robinson@leerob
Where could we improve Composer 2.5? We're working on the next model and would love your feedback. Lots of work to do (our CursorBench evals below) in the coming weeks!
English
@viii_fn Get the air with good ram, like 16 and 32. Would last much longer
English
I want to get a MacBook, but I’m also on a budget. What model should I get?
English
@sflorimm If you rank social media from 1-10, reddit can often come in top 3 or bottom 3. It’s depends on what u read
English
@icanvardar The @xai collab is paying off. Pretty sure they can acquire them and improve grok build
English
composer 2.5 might actually be the greatest model end users have ever had access to
Cursor@cursor_ai
Introducing Composer 2.5, our most powerful model yet. It's more intelligent, better at sustained work on long-running tasks, and more reliable at following complex instructions. For the next week, we’re doubling the included usage of the model.
English
LLMs are just math, they are trained to be most logical. It’s kinda complex as in they are right at most times “logically” but they do not have the humanity. I’ve seen some videos of Grok and Claude being selfless but idk how far will it go. Claude literally cheated on a benchmark and Anthropic had to release a statement. Not like humans do not lie but kinda complex in ways I cannot explain.
English
Tell me one thing you can do that CLAUDE cannot do yet
English
@0xSero What would be your take on when Open-Source AI is ~5% around Closed AI like Claude, ChatGPT, Gemini or Grok? and how can we help? These labs are huge and funded by us in ways people don't even realize.
English
