Rohit kashyap

Hey you, yes YOU. Stop running lighthouse tests. Understand what really matters, lighthouse scores have a place in your workflows and not your decision making. rohitpotato.xyz/thoughts/we-do…

i get that he’s young but there’s no way he didn’t have malicious intent - oss runs the machine you’re working on man - gotta have respect

@josefbender_ nowhere

What are you guys using Server Components for?

Read this great article about implementing raft consensus algorithm: notes.eatonphil.com/2023-05-25-raf…

@lokeshtrajan @mikejava85 🤣

@mikejava85 Under war time scenario, it’s not about achievements. I never said it’s an achievement. It’s the ability to migrate from software maintained by other countries (if we lose access) to a clone maintained by our own people. It’s self reliance

Why we ae blaming Galgotias University ? Last year IIT Madras just cloned postgres, renamed it to shaktidb, and released it as "indian database to reduce dependence on foreign interests".

Wrote something about service meshes a while ago: rohitpotato.xyz/thoughts/to-me…

Most teams running EKS eventually hit the same requirement. Pods need to talk to AWS services like S3, DynamoDB, or SQS. For years, the answer was IRSA ( IAM Roles for Service Accounts). >> The old way: IRSA << > You had to inject an OIDC token into every pod > Your pod had to read that token and call AWS STS itself > You had to configure IAM to trust your cluster’s OIDC provider > Your trust policy was tightly tied to the cluster identity > If you deleted or recreated the cluster, the trust setup broke > Your pods were responsible for proving their own identity to AWS It solved the problem, but it added a lot of operational friction. >> The new way: EKS Pod Identity << This is AWS’s new, native way to manage IAM permissions for Kubernetes workloads. - No OIDC setup. - No cluster-specific trust logic. How EKS Pod Identity works: - Your pod no longer carries any OIDC token - Your pod never talks to AWS STS directly - Your pod just asks for credentials locally on the node - The Pod Identity Agent handles authentication for you - The agent verifies the ServiceAccount and Namespace - The agent assumes the IAM role on your pod’s behalf - Credentials are injected automatically where SDKs expect them Now the platform handles identity, not your application. >> Why this matters << > Same IAM role works across multiple clusters > Better scaling for large workloads > No application code changes > Cleaner and simpler operational model The platform proves identity, not the pod. Trust is managed by EKS, not by cluster-specific OIDC setup. If you are still managing OIDC providers manually, Pod Identity is the better path forward.

@brankopetric00 i am starting to think a lot of your posts are AI generated, no shade, just asking

Breaking your monolith into 47 microservices doesn't fix your bad code. It just distributes the pain across more servers.

@brankopetric00 its easy to set up, though, especially with k3s up and running

Stop. Putting. Everything. In. Kubernetes. Your 3-user internal app doesn't need the same infrastructure as Netflix.

i think its mass psychosis. You all need a break from the constant fear mongering and see things as they are

everyone on twitter just shit talks all day

is it just me or have people on this app become irrationally angry, increasingly bitter & overall just vile people for no reason? why is every reply just unnecessarily mean? whatever happened to normal conversation and scrolling past things you dont like instead of instant anger

Am I going insane, or is no one thinking about the energy implications of AI? Ya, claude opus is great, but it costs a great deal. We take energy from someone to make sure we churn out CRUD applications. Cheaper does not automatically mean less resource consumption. What about the economy as a whole? If there's no one to buy stuff, what's the point of selling anything?

@arpit_bhayani Heyarpit, dm'd you about something, not sure if you got a chance to see it yet.

Most of us assume that stored procedures are cached globally, but this is not the case with both MySQL and PostgreSQL 🤯 Here's how it works... Both MySQL and PostgreSQL cache stored procedures on a per-session basis. Thus, different connections also do not share cached procedure structures. When we execute a stored procedure for the first time in a session, the server parses and converts the procedure body into an internal structure and caches it. In MySQL, this structure is called `sp_head`; while PostgreSQL creates prepared statements for each SQL command in the function. Subsequent calls within the same session reuse this cached structure, avoiding re-parsing. This is where the benefits of using a stored procedure kick in. So, in MySQL and PostgreSQL, if 100 connections each call the same stored procedure, each one builds and maintains its own cached copy. By the way, SQL Server and Oracle maintain global procedure caches shared across sessions. When the session ends, the database discards all cached procedure structures associated with it. This means the first call to a stored procedure in every new session pays the parsing cost. So, if we use a large number of short-lived connections each invoking stored procedures, we are going to be actually slower and more expensive, because of all the reparsing and caching over and over again. Something to keep in mind the next time you use a stored procedure. Pro tip: always read the official documentation, never assume things :)

@shiri_shh rohitpotato.xyz

Show me your app, website or project and I’ll share my honest thoughts👇

@ccccjjjjeeee @thdxr hahhaah

@thdxr The real winners are making vague posts

people say programmers are splitting into ai skeptics or ai believers but like i always say with every new wave a dichotomy shows up and in hindsight neither side matters, the framing was wrong, and you should have been looking at things completely differently

@manthanguptaa for the record, i hope i'm wrong, i was on the productivity boat for a long time, but with recent advancements and usage, my opinion has changed. The progress in the last 2 months has been unlike i've seen before, and its only going to get better/

nah, just think about the fact that these agents will soon know best practices, opus is already magnificent, soon they will be able to pull off architectural design, review each other's work, so need to oversee either. Maybe you need some people to look after, but would you really need this many software engineers?

I want my coding agents to pick up tasks from Jira and work 24x7 even while I am asleep. That would truly make me a 10x engineer

I’m not joking and this isn’t funny. We have been trying to have a baby since last year. There are various options, not everyone is aligned… I gave Claude Code a description of the problem, it impregnated my wife in an hour.

I prefer to see things as they are now, the promise of will or wont happen is a phenomenon that a lot of people use to fund their reality. I don't discard what you're saying, i am just saying its a scenario we'll have to see play out. Its not that simple, and spans much further than software engineering.

@rohitpotato @alkhalidsardar @leerob i think this is a short-sited view, AI once implemented across the board will solve the energy problem and emissions issue you are worried about 1000x better that any human effort ever could.

It’s been interesting to watch two drastically different reactions to this. 1. Yeah, AI is changing software engineering. Let’s figure out what that means. Cautiously optimistic. 2. Denial.