Satyam Kumar

AI Engineer | Thinking aloud on Tech, AI, GenAI, Politics & Finance

[NDTV @ndtv]

'Don't Have Courage To Take Exam Again': NEET Aspirant Dies By Suicide https://t.co/G7AHcDydTh

[Akshat Shrivastava @Akshat_World]

1) We spent billions doing photography/marketing of politicians.
2) We spent billions running freebies.
3) We spent billions building pointless statues.
4) We handed over billions worth of resources to crony capitalists.
5) We spent billions pumping non-sense companies that mix sugar-cane juice with petrol.

Then we expect market to perform, 20% salary growth & 15 Lakh in each account.

Yeah sounds about right.

[Satyam Kumar @HeyItsGenAI]

India has its own LLM now

Which is also open source & built from scratch

Saravam AI
#ai #AISummit

[Satyam Kumar @HeyItsGenAI]

AI will not Disrupt IT Only

It will disturb whole ecosystem in IT Hubs

Even Real Estate prices will impact badly

#AI #stockmarkets

[Rahul Gandhi @RahulGandhi]

The AI revolution is here - bringing both threats and opportunities.

Our IT and services sector, a shining star of our economy, is at risk, and thousands of software engineers and professionals will lose their livelihoods if we do not prepare for the storm that is coming.

But we have opportunities too.

Data is the petrol which fuels the AI engine. As I said in parliament, India’s greatest asset is our brilliant people - and the enormous data we create.

In a few days, the government is hosting a grand AI Summit. It should have been an opportunity for India to assert leadership - to demonstrate how a country of 1.4 billion people can use our data to shape the global AI future ON OUR OWN TERMS.

Instead, a helpless PM Modi, has surrendered to the US ‘chokehold’ in the trade deal. Under the pretext of “clearing barriers to digital trade”, every move to use our data for our own benefit will be opposed.

Already, large foreign companies enjoy a near monopoly on our data through Facebook, Instagram, WhatsApp, Youtube, Amazon, Android, etc. With this deal, India will struggle to:

•⁠ ⁠⁠Safely store the data of 1.5 billion Indians in India
•⁠ ⁠⁠Get transparency in their source codes and algorithms
•⁠ ⁠⁠Tax the profits they make using our data

It’s a shame that our Prime Minister has been pressured to hand over India’s prime resource to a foreign power.

[Satyam Kumar @HeyItsGenAI]

Codex 5.3 just genuinely solved software.

It's over.

[Satyam Kumar @HeyItsGenAI]

Farewell SaaS
Hello MaaS

[Satyam Kumar @HeyItsGenAI]

@MrMekzy_

[Kr$na @krishdotdev]

The IT boom carried India for 30 years.
AI just killed it in 3.

> Across Top IT companies
Net hiring: 17 people.
Not thousands. Not lakhs.
Just Seventeen.

A billion people.
A dying service economy.
No innovation engine.
This isn’t fear.
This is math.

[Satyam Kumar @HeyItsGenAI]

We upgraded from GPT-X to GPT-X+1.

Cost doubled.
Latency increased.
Quality looked… the same.

Users still asked:
“Is this answer reliable?”

We learned users don’t care about model versions.

They care about being lied to confidently.

#LLM #OpenAI

[Satyam Kumar @HeyItsGenAI]

Our CTO asked if we were “AI-ready.”

So I turned off the internal chatbot.

No one noticed for two weeks.

Turns out our AI strategy was a slide deck talking to itself.

We are now officially “PowerPoint-ready.

[Peter Girnus 🦅 @gothburz]

Last quarter I rolled out Microsoft Copilot to 4,000 employees.

$30 per seat per month.

$1.4 million annually.

I called it "digital transformation."

The board loved that phrase.

They approved it in eleven minutes.

No one asked what it would actually do.

Including me.

I told everyone it would "10x productivity."

That's not a real number.

But it sounds like one.

HR asked how we'd measure the 10x.

I said we'd "leverage analytics dashboards."

They stopped asking.

Three months later I checked the usage reports.

47 people had opened it.

12 had used it more than once.

One of them was me.

I used it to summarize an email I could have read in 30 seconds.

It took 45 seconds.

Plus the time it took to fix the hallucinations.

But I called it a "pilot success."

Success means the pilot didn't visibly fail.

The CFO asked about ROI.

I showed him a graph.

The graph went up and to the right.

It measured "AI enablement."

I made that metric up.

He nodded approvingly.

We're "AI-enabled" now.

I don't know what that means.

But it's in our investor deck.

A senior developer asked why we didn't use Claude or ChatGPT.

I said we needed "enterprise-grade security."

He asked what that meant.

I said "compliance."

He asked which compliance.

I said "all of them."

He looked skeptical.

I scheduled him for a "career development conversation."

He stopped asking questions.

Microsoft sent a case study team.

They wanted to feature us as a success story.

I told them we "saved 40,000 hours."

I calculated that number by multiplying employees by a number I made up.

They didn't verify it.

They never do.

Now we're on Microsoft's website.

"Global enterprise achieves 40,000 hours of productivity gains with Copilot."

The CEO shared it on LinkedIn.

He got 3,000 likes.

He's never used Copilot.

None of the executives have.

We have an exemption.

"Strategic focus requires minimal digital distraction."

I wrote that policy.

The licenses renew next month.

I'm requesting an expansion.

5,000 more seats.

We haven't used the first 4,000.

But this time we'll "drive adoption."

Adoption means mandatory training.

Training means a 45-minute webinar no one watches.

But completion will be tracked.

Completion is a metric.

Metrics go in dashboards.

Dashboards go in board presentations.

Board presentations get me promoted.

I'll be SVP by Q3.

I still don't know what Copilot does.

But I know what it's for.

It's for showing we're "investing in AI."

Investment means spending.

Spending means commitment.

Commitment means we're serious about the future.

The future is whatever I say it is.

As long as the graph goes up and to the right.

[Satyam Kumar @HeyItsGenAI]

List of all LLM fine tuning techniques

[Ashutosh Maheshwari @asmah2107]

LLM Fine-tuning techniques I’d learn if I wanted to customize them:

Bookmark this.

1.LoRA
2.QLoRA
3.Prefix Tuning
4.Adapter Tuning
5.BitFit
6.P-Tuning v2
7.Soft Prompts
8.Instruction Tuning
9.RLHF
10.DPO (Direct Preference Optimization)
11.RLAIF (AI Feedback)
12.Multi-Task Fine-Tuning
13.Full Fine-Tuning
14.Mixture-of-Experts Fine-Tuning
15.Federated Fine-Tuning
16.On-Device Adaptation

[Manthan Gupta @manthanguptaa]

I spent the last few days prompting ChatGPT to understand how its memory system actually works.

Spoiler alert: There is no RAG used

https://t.co/zxvRRP2GK8

[Satyam Kumar @HeyItsGenAI]

AI vs AI

[kapilansh @kapilansh_twt]

Someone ordered eggs on Instamart and only one came cracked.

Instead of just reporting it, they opened Gemini Nano and literally typed:
“apply more cracks.”
In a few seconds, AI turned that tray into 20+ cracked eggs — flawless, realistic, impossible to distinguish.

Support took one look at the “proof”…
processed a full refund…
and moved on.

Just pause and think about that.
Our refund systems were built for a world where photos were trustworthy.
But now they’re up against 2025-level AI —
and they’re getting absolutely destroyed.
If even 1% of people start doing this,
quick-commerce unit economics won’t just suffer —
they’ll implode.

AI isn’t the villain here.
The real problem is verification systems stuck in the past.

Welcome to the era of AI vs AI.

[Satyam Kumar @HeyItsGenAI]

Gangsters of silicon valley.

[Satyam Kumar @HeyItsGenAI]

Try this once

[Satyam Kumar @HeyItsGenAI]

😆

[Aiden Bai @aidenybai]

just so we're clear:

Antigravity is a Windsurf wrapper

Windsurf is a VSCode wrapper

VSCode is an Electron wrapper

Electron is a Chromium wrapper

Chromium is a C++ wrapper

C++ is a C wrapper

C is an Assembly wrapper

Assembly is a Machine Code wrapper

Machine Code is a Binary wrapper

Binary is a Physics wrapper

Physics is a Math wrapper

Math is a Logic wrapper

Logic is a Philosophy wrapper

Philosophy is a Humans wrapper

Humans are a Carbon wrapper

Carbon is a Star-forged-matter wrapper

Stars are a Gravity wrapper

Gravity is… definitely not an Antigravity wrapper

[Satyam Kumar @HeyItsGenAI]

Weibo's new open source AI model VibeThinker-1.5B outperforms DeepSeek-R1 on $7,800 post-training budget

https://t.co/EBC1YVkPTK

#LLMs #DeepSeek #GenAI

[Satyam Kumar @HeyItsGenAI]

Sad
#Delhi #RedFort