I can’t tell if I’m being spoiled by OpenAI with constant usage resets or if I should just accept a future (very soon) where I use the best models for everything and usage is not something to worry about.
I don't get the same flow state and reward I used to with coding, but I still absolutely love engineering. Im still building complex systems that coordinate with each other, automate repetitive processes, self-improve overtime. This has always been the job!
This is similar to how I work in Claude code. I create a "Group" for a long running stream of work. I have one session that creates a plan of attack and fork that session for each sub-task. When the sub task is done, that chat is archived.
Excited to try this out, and remove all the manual coordination I currently do.
Introducing Projects, a new way of working in Cursor.
Rather than creating a chat for every task, you work with a coordinator agent in a single, persistent thread.
Like @bot, your agent is always on, proactively manages work with subagents, and improves over time.
Neil Raina retweeted
Introducing Assembly Studio.
The first AI app builder designed to increase your revenue.
Comment something you want us to build and we'll build it for FREE.
Assembly Studio is here, the AI app builder built for real businesses, not weekend projects
xitter.cf/marlonmisra/status/208…
do you know someone that’s tried Claude code for their business, was blown away with what it built, but just couldn’t get it over the line to share with their customer for everyday use? Stay tuned, i have an answer!
Me: "move this behavior to another place"
Claude: "Done! And also here is a 500 word essay in comments explaining why its moved with references to all the tickets that were ever involved in building this"
new words that I've used recently thanks to llm training my speech patterns:
- prose
- disposition
- hysteresis
- provenance
I've never been a fan of using big words but ngl it does feel pretty satisfying to drop an S tier word in convo
There is long list of things I just don’t think about in my work anymore. resolving merge conflicts and rebasing is definitely near the top of the list. So many hours wasted
An engineers ability to produce more has sky rocketed, getting it over the line seems to be the bottleneck.
How are teams overcoming some of the existing human in the loop slowness in the age of agents? I’m talking about things like code reviews, testing, fine tuning solutions.
new @readme rollout is garbage. Tried the upgrade and it completely messed up our docs. now firefighting to figure out how to get thing reverted, and of course their support responds at a snails pace. PSA for anyone thinking of accepting the refresh.
every year over the past decade, i’ve worked harder than the year before. 2025 will be no exception—let’s keep the streak alive.
Excited to share some major updates from our latest launch! It’s an amazing feeling to build something that’s making a real difference for so many businesses. The day-to-day can be intense, but moments like these show how all the dedication and effort align.
Big fan of @cursor_ai so far. Only feature that feels like it doesn't quite work yet is the codebase chat. Im optimistic that will get better too.
There is a large number of South Asians software engineers, but I rarely meet any one named Dev (pronounced D-aye-v) despite it being a relatively common name. Why? Are they actively avoiding their namesake? 🤔
If you use Typescript but your compile step succeeds even with errors or you are using ‘any’ then you aren’t really getting the major benefits of using it in the first place.
Invest in the steps to make this happen in your app and it will really pay off!
At copilot.com we moved to TS some time ago but didn’t fail compiling on errors from the start.
Instead, we setup a step in our CI that counts the number of errors and enforced this number can only decrease with the goal of getting this to 0!








