Fardeem
@FardeemM
Followers
2K
Following
7K
Media
215
Statuses
4K
eng @tryramp - big finder of beauty in the world
New York City
Joined February 2020
Building on top of what Sarim said, a few other things that I think makes a big difference in shipping speed:. - Being able to test complex flows easily. This means decoupling time if you're state depends on time so you can time travel to debug and test states. Honestly, if you.
Shipping velocity is directly proportional to data structure quality, especially in the age where UI can be increasingly treated as an ephemeral asset. More practically, this includes:.- the type of store you use e.g. relational, graph etc. - the way you structure your data.
2
0
11
The beauty of building on top of xstate is that you can make a +300 line PR and know it will work on the first run because everything is modeled using state machines.
1
0
5
All this planetscale talk reminded me how in 9th grade or smth, I was trying to put my wordpress blogs mysql database behind vitess . Other than solving the hyper scale issue for the 2 people visiting my blog every day, I actually learnt so much about databases. Kinda miss the.
1
0
5
this is true.
My mind is blown. If you follow me you’re going to be begging to give @hassantsyed your money in 1-2 months.
0
0
5
"Once you try it, you can't go back". Effect is definitely one of the coolest things I learned this year.
1
0
15
I think Im going to cry in joy.
Here's how to guarantee you've handled every known error in your system. from the very top of a potentially enormous call stack. Can your code do this?
0
0
1
Who is building the platonic coding agent (ie. the most perfect, pre-eternal coding agent)?.
0
0
3
Everyday before bed i look through linear for some simple tickets. I then give them to codex and wake up to a couple of PRs I can review. Almost all of them need more polish before I can merge but it’s a great headstart.
0
0
5
Tempted to host a coding and qawwali event in nyc. Get fareed ayaz to go crazy in the front while everyone is insanely locked in.
1
0
10
Using notifications to solve app retention no longer works. It used to work before. But people have evolved to ignore notifications right now. Build a product that solves of a painful problem or bust.
0
0
4
Iterating on prompts for an LLM app is actually quite hard (unless im fumbling this). Half my evals pass with one prompt. The other half passes with a different prompt. Clearly the model can do the whole range of tasks. But how do I make one prompt that makes all the evals.
0
0
3
Building an app of any meaningful level of complexity without effect and xstate should be illegal.
0
0
3
Been thinking about why this is the case. I don’t think codex is unequivocally the best. One reason though is codex has the best mobile app experience out of anyone. Also it comes free with chatgpt.
in the last 35 days, @OpenAI codex has merged 345,000 PRs on github. 345,000. AI is eating software engineering
1
0
3