kriky

Quirky, crazy, lazy, kriky.
Also building Jdlens.

[kriky @kriky]

@maa_bhaishiiH You're mixing too many debates into one. I agree the starting line must be made fair. That's where reservation has its strongest case. But once the race begins, standards should be the same. The reservation debate has become so politicized that meaningful reforms rarely happen.

[kriky @kriky]

@Talk2Shubh @swing_ka_sultan Let me clarify: I support the principle of reservation, not necessarily the current admission mechanism. Reservation addresses unequal starting conditions. How we implement should evolve with evidence and time to remain fair, transparent, effective. They're two different debates.

[kriky @kriky]

@sundarmail We're debating NEET vs. the old system, but both look backward.The real issue isn't the exam—it's selecting better doctors. Keep a national benchmark, but build a standardized pre-med track in Classes 11–12 with continuous, AI-assisted assessment. Judge consistency, not one exam

[kriky @kriky]

@karlmehta Trusted delegation starts with trusted interpretation. A perfectly executed action on a misunderstood problem is still the wrong outcome.

[kriky @kriky]

@raavanamavan7 @makkal_saatchi Good start. But a word of caution, don't fall in love with your creation. Go well..

[sakthi @sakthijess]

Chetans mass panringa ya🥵

[kriky @kriky]

@polimernews Probably one of the most heartfelt, honest and most loving tribute I have heard of anyone so far. Go well Bagayaraj. Indru poi...

[kriky @kriky]

TVK is slowly turning into Tender Venam Katchi. Dangerous.

[kriky @kriky]

Strange times.

I hired a llama.

Not the animal.

The model.

Employee #1 at JD Lens.

Most people think AI replaces jobs.

My first experience was different.

I gave it one.

The hardest part?

Figuring out his spit-outs

Should've seen that coming.

He's a llama.

[kriky @kriky]

@whynotraja Maybe i was not clear, I know what open source is ). But that too is a market and it should not be a case of spray and pray.

[kriky @kriky]

@emollick Perhaps the question comes before model choice.
What's missing? What's unclear? What's the cost of being wrong?
Then choose the intelligence level.
You don't hire a NASA scientist to fix a printer. But you also don't send an intern to investigate a rocket failure.

[kriky @kriky]

@satyanadella This assumes a level playing field that doesn’t exist.
“Learning loops” and “token capital” sound clean, but they rely on stable orgs, clean data, AI fluency, and access.
That’s not universal. So this compounds asymmetrically. Some firms enter the loop. Others probably never.

[kriky @kriky]

@whynotraja @itisprashanth Nice. The orchestrator solves routing. My question is one step earlier: how do we determine the intelligence requirement of the task itself?
In cloud, we size the workload before choosing the server. Would AI get there, it should is my guess.

[kriky @kriky]

@santosh_arron @svembu "P(latent structure | observations)".
That's the line that resonated.
I've been investigating a similar idea in language: interpretation begins with identifying what must be true, but unstated, for information to make sense.
Currently testing a possible implementation.

[kriky @kriky]

@praddy06 .reduce by 1-2 degrees. Which means real feel will still be above 40... Phew...

[kriky @kriky]

@Pitysoulgod @shankarprakasha “No reason” is still a reason.
People may not know what they want — but they know what they don’t. The ballot becomes the language of that frustration. In that space, the voice that cuts through stands out. This time, that voice was Vijay. That alone is often enough to decide.

[kriky @kriky]

@SagarH62 Craig Mcdermott