Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
jayantbhawal
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
6 ms
·
31.
▲
by
jayantbhawal
2y ago
I mean... feels pretty obvious from my own experience.
32.
▲
Couple left with life-changing crash injuries can't sue Uber
(independent.co.uk)
11 points
by
jayantbhawal
2y ago
|
1 comments
33.
▲
South Korea set to criminalize having or watching sexually explicit deepfakes
(cbsnews.com)
6 points
by
jayantbhawal
2y ago
|
0 comments
34.
▲
by
jayantbhawal
2y ago
I'll have what he's having
35.
▲
by
jayantbhawal
2y ago
Totally fair. Speed is just one piece—it's key to balance it with quality and stability. Faster cycles are great, but only if they don’t compromise long-term project health.
36.
▲
by
jayantbhawal
2y ago
I'm looking forward to not needing to use useMemo/useCallback anymore!
37.
▲
by
jayantbhawal
2y ago
I get where you're coming from, and can see a lot of people seeing it that way. But to me, metrics like 'PR to merge time' help highlight bottlenecks, even in open source. It’s not just about churn, it’s about ensuring contri
38.
▲
What's New in React 19
(vercel.com)
8 points
by
jayantbhawal
2y ago
|
2 comments
39.
▲
by
jayantbhawal
2y ago
We track PR merges as potential deployments to measure flow and efficiency, even if not every merge triggers an immediate deploy. It’s more about tracking how fast the code moves through the pipeline.
40.
▲
by
jayantbhawal
2y ago
Good point, but I’d say lead time still matters. Even in open source, slow responses can kill motivation. Quick feedback keeps contributors engaged and makes the project feel more active.
41.
▲
CERN prepares to expel Russian scientists – but won't cut ties
(nature.com)
5 points
by
jayantbhawal
2y ago
|
0 comments
42.
▲
by
jayantbhawal
2y ago
This is AMAZING! I hope this is released for public use at some point. I'd love to run it through some of my older photos to see what it does with them.
43.
▲
by
jayantbhawal
2y ago
I feel poor looking at this.
44.
▲
by
jayantbhawal
2y ago
Did I catch a virus?
45.
▲
Almost all new car sales in Norway last month were EVs
(qz.com)
101 points
by
jayantbhawal
2y ago
|
115 comments
46.
▲
Is FreeCodeCamp Sacrificing Quality for Speed with Their Rapid Deployments?
(middlewarehq.com)
7 points
by
jayantbhawal
2y ago
|
0 comments
47.
▲
Swift Deployments: Are They Swift or Recklessly Rushed?
(middlewarehq.com)
8 points
by
jayantbhawal
2y ago
|
0 comments
48.
▲
Vue.js: A Speed Demon or Just a High-Risk Gambler?
(middlewarehq.com)
8 points
by
jayantbhawal
2y ago
|
0 comments
49.
▲
Did React's Repo Miss the Speed Limit? A Look into Fast Features and Bug Delays
(middlewarehq.com)
7 points
by
jayantbhawal
2y ago
|
0 comments
50.
▲
by
jayantbhawal
2y ago
Analysis paralysis is so real and frustrating with the leadership. The shame is the intellectualisation of that crap and inactivity. As a builder, I’d love to sip my coffee and build impactful stuff.
51.
▲
Klarna's AI chatbot: how revolutionary is it, really?
(blog.pragmaticengineer.com)
5 points
by
jayantbhawal
2y ago
|
1 comments
52.
▲
A Software Engineering Culture Test
(blog.pragmaticengineer.com)
5 points
by
jayantbhawal
2y ago
|
0 comments
53.
▲
by
jayantbhawal
2y ago
Oh yeah that was the first thing that came to my mind
54.
▲
by
jayantbhawal
2y ago
This is pretty neat! I'm glad there's another open source option for trying out DORA four keys.
55.
▲
by
jayantbhawal
3y ago
Hey pacifika, This includes data from your private repos to show a more complete picture for those that may have a lot of closed source contribs. I totally get your hesitation to try it out, hence the link to the repo. You'd notice tha
56.
▲
by
jayantbhawal
3y ago
Hey all! I'm one of the makers of Unwrapped, really appreciate the support so far, and of course, any feedback or direct contributions would be super appreciated. In terms of code, or design! Try it out here: https://unwrapp
57.
▲
by
jayantbhawal
5y ago
Was one of the first folks to try this thing out. Found this to be a solid way to learn some Redis internals. 10/10