M Boosiri

Exploring the intersection of technology and creativity. #Tech #Creativity #Innovation

[M Boosiri @mboosiri]

I’m walking 100k in September for every much-loved baby sadly not here today. Anything you’re able to donate truly means so much – I’m only £124.67 off my target. Thank you https://t.co/rHYdP4ExEy

[M Boosiri @mboosiri]

Two hackathons. Two very different problems. Two finals.

Built RiskWise: agentic AI for modelling business risk.
https://t.co/ESIoSLenrg

@TXODDSOfficial × @solana: Built BigShout - live sports predictions recorded on-chain
https://t.co/WnafAvUbqk

@althafurahman & Paul Zhan.

[M Boosiri @mboosiri]

@__mharrison__ Jetbrains mono

[M Boosiri @mboosiri]

@tensor_fusion That's one of the coldest post!

[M Boosiri @mboosiri]

@swyx How times changed, when this people gets the attention and becomes a thread of discussions!

[M Boosiri @mboosiri]

@aakashgupta Setting up tools also falls in this first category😄

[M Boosiri @mboosiri]

@svpino Library support?

[M Boosiri @mboosiri]

@aakashgupta There is a catch to the free option here. When we start using the Audio, Video capabilities, it will quickly uses up the limit, as each interaction counts as a call from 40/3hrs.
Basically, everyone expecting the 'her' experience would eventually need the 'plus' sub. #chatgpt4o

[M Boosiri @mboosiri]

@svpino 'Dunning
-Kruger effect" - is simply people overestimate their competencies. It's counterpart 'Beginner's Mindset', which is valued for excellence.
So no matter the new or Senior, one might fall behind without the right mindset, that with anything - And AI is just a name now!

[M Boosiri @mboosiri]

"Don't misinform your Doctor, your Lawyer, and certainly not to your AI models"

Benjamin Franklin might have updated his advice today

[M Boosiri @mboosiri]

What is a Neural Network?
"It's a mathematical abstraction of the human brain."
~ @karpathy

A definition both deep and beautiful is rare to find.

[M Boosiri @mboosiri]

@InternetH0F BODMAS - Bracket Off Division Multiply Add Subtract

[M Boosiri @mboosiri]

Keep setting goals.
If it feels out of reach:
- Raise the bar and aim for harder goals.
- Might still fail

Interestingly, we'll often surpass the original targets without even realising.
#experience

[M Boosiri @mboosiri]

@__mharrison__ Absolutely.
And after that, instead of thinking about the next 29 courses, do a project based on that one course we did.
We might be capable of creating a course by ourselves.

[M Boosiri @mboosiri]

@stats_feed Yes, the IITs, UPSC and similar indeed attract the brightest minds due to their stringent tests.
Is this intense competition a reflection of the country's population and lifestyle? As the list goes down, this competition lessens, yet the rewards remain comparably significant.

[M Boosiri @mboosiri]

@MKBHD the other way around with #AI

[M Boosiri @mboosiri]

Care for ML Models:
ML models tend to rote like a fruit if not treated well.
Actively monitor - Iteratively augment with data for sustained performance
- Watch out for Model/Data drifts
- Health issues in the pipeline
#mlflow #tfxserving could help.

[M Boosiri @mboosiri]

Deep diving 🪂 into a small yet complex dataset. Finding the right techniques is crucial for a desired range of accuracy. Consider a Material science🧑‍🔬domain, transforming these kind of datasets into accurate models is a familiar challenge #DataScience #MachineLearning

[M Boosiri @mboosiri]

Knock the door for 'more data' or Synthesise:
- Emphasising domain knowledge once again, with right domain knowledge and clear communication, it is more likely to get 'more data'