Any suggestions for personal project

I was working on a project using @better_auth and @DrizzleORM and found out the @better_auth CLI still creates auth tables with the depreciated import of relations from the Drizzle ORM package
Will request the team to update the code for CLI with the new import of defineRelations

I was thinking of doing something naive with @typesafeai Jev.
How about giving the entire dictionary and telling it to frame a sentence based on the input chat?

Thinking of creating my own Wispr Flow agent using OpenAI's Wispr model. Any suggestions on what I should use for inference: TensorFlow inference or ONNX runtime inference with CPP, just like Wispr.cpp?

Today we hit a PostgreSQL + Docker issue: we needed pgvector, but the image didn’t have it.

I jumped into the container, researched, and fixed it all without AI.

Seeing my seniors impressed by my DevOps skills felt amazing.

Thanks, @kirat_tw. 🙌

I met @darkosubotica at @AWSCommunity day. It was really a great session on Kiro Crew.

I am really grateful that I got a chance to meet you and had a conversation with you it really cleared all doubts and helped me to focus on right path in my career.

I met @darkosubotica at @AWSCommunity day. It was really a great session on Kiro Crew.

I am really grateful that I got a chance to meet you and had a conversation with you it really cleared all doubts and helped me to focus on right path in my career.

The best way to develop a project

Feeling powerful having the speech-to-text model in my phone 😜

Today I ran @OpenAI Whisper base model using onnx runtime on my phone it went really well.
Never thought it would be this much fast on mobile CPU.

I wanted to implement speech to text and was confused and then found a way we can convert @OpenAI whisper to onnx and then using onnx runtime I can the speech to text with higher accuracy.
In the offline mode

In software engg more you learn the more you feel like you know nothing

Been using @WisprFlow but the free tier caps at 2000 words/week
turns out browsers ship speech-to-text built in
thinking of building my own:
◆ browser speech recognition
◆ claude sdk, like @theo 's T3 app
◆ hooks into my existing claude sub with fast mode 😋

Been using @WisprFlow but the free tier caps at 2000 words/week
turns out browsers ship speech-to-text built in
thinking of building my own:
◆ browser speech recognition
◆ claude sdk, like @theo 's T3 app
◆ hooks into my existing claude sub with fast mode 😋

I just found out browsers have a built-in speech recognition API

you can add voice input without any external libraries

the Web Speech API handles it natively

didn't know this existed until today

@mattpocockuk skills helped to level up my coding skills. The way they make you think through problems really helps
I built a skill called Socratic Tutor, which helps you learn by answering questions back to the mentor instead of just selecting the answers
gist.github.com/parth3083/8331…

I have been using T3 code for while but since last week I am facing a issue. Every time I fire up T3 code my laptop after while stuckw and then it comes something went wrong
@theo can you look into it
But tbh @t3dotcodes is best thing for work and learning.

is it just me or is whatsapp dekstop app not working properly for everyone

What is stack PR ??? 🤔

learning this the hard way with AI models:
benchmarks measure general capability
your actual work needs specific strengths
been using different models for Odoo development and the "lower ranked" ones often perform better for my tasks
test on your real work, not just leaderboards