@aantoinelee at this point tmux or some ade like superset or conductor would be better 😭
English
Harsha Vardhan
90 posts
Harsha Vardhan
@harshahq
clanker whisperer | ethical hacker | design engineer | SWE | AI engineer https://t.co/b45OjMeee7
Sydney Katılım Şubat 2026
71 Takip Edilen12 Takipçiler
50x productivity with this cc setup
well engineered loops run longer and more autonomously which enables parallel workflows
cycling between 4 claude 20x accounts as limits are hit but the ROI for a $200 plan is insane
also exploring cursor cloud agents for more parallelisation. the built-in computer use can make for effective loops where agents test their own features
what's your setup look like rn?
English
@orcdev i wish it had more than one account support cause CodexBar shows almost triple across all the accounts :(
English
love this update of OpenUsage so much
now I can flex my usage
Robin Ebers · AI for Small Business@robinebers
OpenUsage 0.7.4 is is here and a massive upgrade! - OpenCode is back - Cost/Tokens now shows at the top - New models now show up in today/yesterday/etc - Lots of UX improvements + small bug fixes if you run into any issues, hit me up! github.com/robinebers/ope…
English
@aantoinelee you guys still using claude max? tokenmaxxing isnt the same until you cancel and get codex pro instead 🫣
English
Finally built my first Mac app called Keepme 💙
Where I save all my design inspiration
Built with Claude Code.
It took me 1.5 weeks to go from 0 → 1.
I focused on getting the functionality working first, then refined the design.
Building your own app feels like a superpower. You have complete creative freedom to shape every detail exactly the way you want.
English
@iMuffined Happened to me and checked platform.openai and turns out it fell back to api credit and used everything there until that ran out too then stopped
English
@rudrank @mopperson github.com/rudrankriyam/F…
Try it yourself! Tested it today on a Mac Studio and iPhone and it’s pretty amazing what is possible with it
English
So in beta 3, Foundation Models can go upto 8192 context window! I think for iPhone 17 Pro and Max
Anyone try it out?
English
Harsha Vardhan retweetledi
there are a lot of benchmarks that suggest 5.6 sol is the best model in the world right now, but the most reliable way to tell is that elon is obsessed with me again
English
When your ex anxiously comes running back cause bout to meet someone new 👀
Claude@claudeai
We're extending access to Claude Fable 5 on all paid plans through July 12.
English
@kr0der @brandon_galang Kimi k2.7-code is basically as close to composer 3 rn
English
for whatever reason my brain has a really hard time using opus 4.8 and fable 5 for development work
I always find myself spinning in circles, missing stuff, or getting out of flow with how much slower they feel compared to 5.5 or Composer 2.5 Fast
so excited for 5.6
English
@claudeai Sounds like they had no choice since Sol is about to come out. Not having Fable available for general use in the subs to compare with would’ve been way worse
English
As before, you can use up to 50% of your weekly usage limit on Claude Fable 5. After that, you can keep using Fable 5 with usage credits, or switch to another model to keep working within your remaining limits.
Read more: support.claude.com/en/articles/15…
English
@traviscurnutte @cursor_ai Personally Kimi k2.7 code is the closest to composer 3 I’ve felt in cursor
English
If @cursor_ai releases Composer 3.0 after Fable is taken off sub plans AND it benchmarks the same or higher at a substantially lower price?
You’ll see a massive influx to the positive in their P&L
Facts.
English
@whatdotcd I use Hermes daily and would love to try what you're working on!
English
Friends who use personal agents (Hermes / Openclaw / codex / cowork etc) daily reply to this so I can send you something cool to try
English
Harsha Vardhan retweetledi
@pondorasti you should add cursor support, would love to see the results combined with @harsha" target="_blank" rel="nofollow noopener">cursor.com/@harsha
English
Introducing tokenmaxxing.sh - the fastest way to get mogged about your token usage.
Subsidies won’t last forever, so might as well make the most of it. Start your loops, spin up your ultracode workflows, and start climbing the leaderboard.
The only purpose of this site is to bring some fun and competition to the sport of burning tokens.
English
Harsha Vardhan retweetledi
introducing matrix.
the ultimate multi-agent runtime that turns your product into actual revenue and automates 99% of what it takes to run a company.
probably the most different, most future-feeling agent product you've ever seen. the harness, the interface, everything is rebuilt from scratch.
think of it as the complement to codex and claude code. they write the code. matrix does literally everything else you actually need done. plug in your codex subscription and run it for free, or byok. fully free either way.
this video was edited by matrix and voiced by an agent. more real cases coming soon.
matrix is in beta preview right now. full launch next week.
retweet + comment and we'll send invite codes to the first 200.
English
Harsha Vardhan retweetledi
Harsha Vardhan retweetledi
