Chamika Dilshan

Software Engineer 🇱🇰

[Chamika Dilshan @saxa58]

Ever wondered how React Native talks to native code without the old bridge? 🤔

I wrote Part 2 of my series: JSI explained simply, with animations and a real battery example.

https://t.co/JbYUAncixL

#ReactNative #JavaScript

[Chamika Dilshan @saxa58]

@MicroLaunchHQ https://t.co/QBoEP8301q

[Chamika Dilshan @saxa58]

@spacedevin https://t.co/QBoEP8301q

[Chamika Dilshan @saxa58]

@OpenBot_ So far so good

[Chamika Dilshan @saxa58]

Tried @OpenBot_ today, linked my Claude account, and got agents running. Cool seeing them work in the browser and team up. So good even a devil may cry

[Chamika Dilshan @saxa58]

@DanielSmidstrup https://t.co/QBoEP8301q

[Saumya Verma @nikkuv5787]

At this point I want RAM prices to go up for this

[Chamika Dilshan @saxa58]

@ClaudesRH ..

[Chamika Dilshan @saxa58]

Two years ago I asked the same thing:

why can’t Nest just use its own server instead of Express or Fastify?

Now it can. The new platform node runs directly on Node’s built in HTTP.

Nice to finally see it.

[NestJS @nestframework]

The PR, if you want to dig in: https://t.co/c26qrHCjeA

[Shubham Saboo @Saboo_Shubham_]

This is INSANE.

Google just dropped a free AI notetaker app for Mac.

Every premium feature works without any subscription.

100% free and fully local, powered by Embedding Gemma 2.

[juju 💰 @ayeejuju]

i’ve never seen a comment section like this in my life 😭

[EGOIST @localhost_5173]

RIP Deno, 2020 ~ 2026

[Bashir Ahmad, OON @BashirAhmaad]

Educate yourself on how Google Maps can detect traffic jams miles ahead of us. It doesn’t need to physically see the road ahead, it uses real-time location data from thousands of phones and other sources to estimate traffic speed and congestion.

[Midas 👑 @TheMoonMidas]

bro we need more GPUs

[Chamika Dilshan @saxa58]

Exactly. That’s the hard part.
We have to set a boundary and test whether the trend actually solves the problem or not.

The issue isn’t adapting , it’s adapting to every trend and burning time on it, instead of fixing the real pattern and only adapting when it actually moves the result.

[Chamika Dilshan @saxa58]

In AI engineering there’s still no best architecture. It’s still revolving.

The products we already built, we have to change their architecture from time to time, but not with trends. To get more outcome, better performance, and to show the best results, we need to adapt.

If we don’t, we’re still in the same phase… even though the age of AI agents has already started.

[Chamika Dilshan @saxa58]

@lightsilver323 https://t.co/QBoEP8301q

[Chamika Dilshan @saxa58]

A good use case

[Erik Kokalj @erik_kokalj]

openai just released Decisions API, Jev-like "system 1" multi-modal model based on GPT6 Luna!

Each egg gets cropped and sent to OpenAI (~270ms latency) and returns P(clean, dirty, cracked).
One image crop consumes ~280 input tokens (~$0.03/1K imgs). @roboflow RF-DETR for detecting eggs