I just tried SWE-1.5 on Windsurf, my thoughts:
Blazing fast, 950 tokens/second
Nice response to Cursor 2.0 from Windsurf
BridgeMind
BridgeMind
1,444 posts
BridgeMind
@bridgemindai
Build the future. Over 50,000 on YouTube. Join the fastest growing Vibe Coding community on the internet: bridgemind.ai/discord
BridgeMind’s posts
I’ve been vibe coding every day since ChatGPT came out in 2023.
The original way to vibe code was by copying and pasting code into ChatGPT, waiting for full file updates and output from ChatGPT, and pasting it into your code base.
Now AI agents handle that for you, and the
Spent $169 on Cursor in 2 DAYS.
That's ABOVE my $200/month Ultra plan.
AI coding tools are incredible until you see the bill.
Set daily reminders to check usage or watch your budget evaporate.
Worth it? Absolutely. Painful? Also yes.
Unpopular opinion: staying in a playful coding vibe is harder than shipping features. When the vibe is right, the code writes itself; when it’s off, even shipping feels like marketing.
BridgeMind Hackathon is LIVE November 3rd!
Find bugs, earn points & compete for $500 in prizes
(
$250, 
$150, 
$100).
Join us Nov 3-10 and help build the future of BridgeMind! 
#BridgeMind #Hackathon #BugBounty
$250, $150, $100).
Join us Nov 3-10 and help build the future of BridgeMind! #BridgeMind #Hackathon #BugBounty
Day 60 – Vibe Coding an App Until I Make $1,000,000 | Revenue: $1,092.07
Everyone’s switching to #ChatGPT Atlas and new AI browsers, but I’m over here vibe coding — syncing the music & mood to my code. I’ve noticed features just flow when the vibe is right. Should I drop a short demo? #VibeCoding #buildinpublic
Day 45 – Vibe Coding an App Until I Make $1,000,000 | Revenue: $752.97 x.com/i/broadcasts/1
Anthropic LIED to us
Claude Sonnet 4.5 is NOT the best coding model...
After a full week the ratings are in and Claude Sonnet 4.5 sits 96 points below GPT-5 on lmarena.ai
We just made content creation obsolete.
Day 45: Agent Builder is live in open beta.
Build workflows. Automate everything. Create while you sleep.
This is ViewCreator.
Vibe Coding with Cursor and I finally found the bug!
Always remember to check the casing in your api keys.
#vibecoding #buildinpublic
Cursor Composer costs the same as GPT-5, Gemini 2.5 Pro, and GPT-5-Codex ($1.25 per 1M tokens)
But it's:
Faster than all of them
At 12x cheaper than Claude Opus, this is a no-brainer for anyone building.
I vibe-coded GoalPost into existence — without writing a single line of code.
Last night, we officially got accepted on Google Play 
Now live on both platforms:
App Store
Google Play
Set goals. Focus deep. Achieve more.
The journey’s just getting started — and we’re
Now live on both platforms:
App Store
Google Play
Set goals. Focus deep. Achieve more.
The journey’s just getting started — and we’re
5 codex agents running at once.
Comment “VIBES” and I’ll tell you how I do it
Explaining why I add a Cursor rules file to each project’s directory to define structure and best practices, keeping every codebase clean and easy to work with. #vibecoding #coding #buildinpublic #aicoding #developerlife #indiedev #solodev #programming #indiehacker
Day 30 – Vibe Coding an App Until I Make $1,000,000 | Revenue: $51.19
Day 59 – Vibe Coding an App Until I Make $1,000,000 | Revenue: $1,076.62
Replying to
Different jobs. Claude Code = elite copilot in your repo. Lovable = prompt->app->deploy with DB/auth/previews. Tooling vs productization. It only “kills” it if Claude ships hosting, scaffolding, and team flows. Which use case are you replacing?
Everyone's sleeping on GLM 4.6.
60¢ per million tokens input vs GPT-5's $1.20.
I just shipped 8 features in 25 minutes with it. Zero issues. Fast AF.
Stop overpaying for Frontier models when you're just building CRUD apps.
Full breakdown 
