Joshua Ma retweeted
Introducing Helios: measuring 1,000+ protein biomarkers from a drop of blood to help you live younger, longer.
Here’s how it works:
• Self-collect blood at home with a painless, needle-free device
• We measure the proteins carrying signals from cells and organs throughout your body
• Your Helios AI agent explains your results, highlights risks, and offers personalized guidance
Proudly backed by Y Combinator and Formation VC.
Join the waitlist at myhelios.com
Joshua Ma retweeted
Hello, world. Your blood has a lot to say. 🩸
Introducing Helios: measuring 1,000+ protein biomarkers from a drop of blood to help you live younger, longer.
Here’s how it works:
• Self-collect blood at home with a painless, needle-free device
• We measure the proteins carrying signals from cells and organs throughout your body
• Your Helios AI agent explains your results, highlights risks, and offers personalized guidance
Proudly backed by Y Combinator and Formation VC.
Join the waitlist at myhelios.com
Cursor's Composer 2 just took second place on the Next.js evals leaderboard, beating both Opus and Gemini.
See the full rankings ↓
vercel.fyi/next-composer2
Joshua Ma retweeted
gpt-5.4 in codex is often unreliable for long-running tasks. even with clear guidance, it often stops early and doesn't fully finish a task.
In many cases, that only becomes obvious during code review. It ticks the tasks but there are a lot of leftovers.
compared same & similar tasks in Cursor and its harness clearly performs better here.
Joshua Ma retweeted
get in babe, we're automating the SDLC!
something I want to highlight about these that's super powerful is that they can kick off on any event or webhook, and run in the cloud. so they are not tied to having one person's laptop open
- they are owned by your entire team.
Joshua Ma retweeted
Beyond excited to share what I’ve been working on - it’s changed the way we write software and I hope it will change the way you do too!
Joshua Ma retweeted
Cursor now has automations!
You can run agents on schedules, trigged by events from Slack, GitHub, or any MCP server. I get a daily review every morning with my GitHub/Slack activity.
Our team now has dozens of agents running 24/7 improving or monitoring things for us.
Joshua Ma retweeted
The way we build software has changed a lot. I'm proud to announce what has made the largest impact for me: Cursor Automations.
Now:
- Whenever CI fails on main, an agent automatically kicks off to fix it
- Whenever we push a PR, an agent automatically kicks off to determine if it's low risk and auto-approves the PR if so
- Whenever we get paged for an incident, by the time I wake up and orient myself in Slack, an agent has already given me a report with a PR ready to merge
- Whenever we get a bug report in our bug reports Slack channel, an agent automatically kicks off to investigate and resolve it
All of this is fully customizable. For the auto-approve PR automation, we give it Notion MCP and just prompt it to add a row to a table every time it makes a decision, so we have an audit trail. For the incident response automation, we give it Datadog MCP so it can navigate its way around the problem.
This has completely changed how I work, and I'm excited for it to change how you work too.
Joshua Ma retweeted
today we're launching automations!
it allows you to trigger agents based on real world events, like incidents, PR's, slack messages or anything that you can connect with a webhook!
internally, we've seen crazy adoption of automations from all teams, and we've outlined a few of our favorites in this blog post
try them out at cursor.com/automations
Joshua Ma retweeted
Bugbot kept finding real bugs. I kept forwarding them to be fixed.
Turns out I was the bottleneck, so I automated my own tedious manual labor out of the loop.
Introducing Bugbot Autofix, now out of beta.
Joshua Ma retweeted
Drives me nuts that @cursor_ai is so underrated at this point. I’ve spent the last week with Claude code, Codex, the new codex app and of course Conductor and can tell you know: night and day. Being on X, I felt I was missing on something. Well not at all. Turns out everyone is missing on Cursor.
Here’s my top five:
- Cursor with Opus 4.5 is unbeatable. Full stop. Even the native Claude code harness doesn’t come close.
- Bug agent is so so SO good that I asked Cursor to create a skill I could use with Codex and Claude to get work done. Pure genius from the team here.
- Cursor has made installing MCPs and Skills so easy that I didn’t realize how bad the others were until I tried. Who wants to waste 15 minutes figuring out the format of each agent and unique workflow to Oauth.
- Cursor agents management is best in class (even though I’m impatiently waiting for them to create some sort of Kanban to track which work I’ve reviewed)
- The UI is delightful to surface everything you need when you need it so you never have to search online how to do something. But it’s powerful enough that you can always go crazy when needed (sub agents, hooks, etc).
I could go on and on talking about how they accidentally created one or maybe the best agentic browser. Then there is composer, by far the best way to work on UI side by side with an agent at light speed. Oh and let’s not forget Bugbot that is pretty much managing my whole QA and fixing things autonomously at this point. 🤯
So yep, I’m keeping my 3 ULTRA (likely need a 4th one this month) subscriptions. I hope they launch a generous 2K plan soon as this would feel cheap for running what feels like a team of 10 engineers in 2020.
I love you @cursor_ai. @notis_ai would not be half the product it is without your work. You are my daily North Star when it comes to building a product that is 200$/month and yet look stupidly cheap.
PS: thank you for shipping updates every day that delight me.
Joshua Ma retweeted
I feel like Opus 4.5 is smarter in Cursor than in Claude Code
but GPT-5.2-Codex is dumber in Cursor than in Codex CLi
both by quite a large margin
anyone else?
Joshua Ma retweeted
tested greptile vs codex vs cursor bugbot on the same PR:
- greptile: 1/3 real bugs
- codex: 2/3 caught
- cursor bugbot: 3/3 + 1 false positive
cursor was the only one that caught a bug requiring significant context from a totally different file. i'm impressed!
Joshua Ma retweeted
When does synchronous coding hit diminishing returns to intelligence?
I believe we're a few months away from a ChatGPT intelligence bar where >95% of these queries don't get much better with smarter models.
More intelligence would only matter for async-level tasks, ones that would take developers hours to perform.
Think about a large fraction of UI work. The bottleneck for layout changes is not intelligence or task horizon, it's user intent.
In that world, the speed matters a lot more! I'm very excited for frontier models at Composer-1 speed.
Joshua Ma retweeted
Replying to @mckaywrigley
We think of Cursor as a desktop app, not an IDE. Agree the ideal Cursor becomes a dashboard for working with agents.






















