khalid karim STEMKITA

My household income classification ~ forget the B40:M40:T20. Let us see real numbers ; Less than RM3000 is poverty 3001 to 7500 is low income. 7501 to 15k is middle income > 15,001 is high income Based on households with two kids. Adjustments needed for family size & location

Malaysia should create an independent IPSA-style body MIPSA ~ to decide MPs’ salary, allowances, staffing budgets and expense claims. MPs should not be deciding their own pay. The principle is simple: ✓Pay MPs properly. ✓Fund their offices professionally. But publish every ringgit clearly. Bersama ~ pls add this to your playbook

I don't look good naked anymore...

Thought for the day..

SkyeChip closed at 67x PER yesterday, minting two new MYR billionaires ; maybe Malaysia’s first semiconductor billionaires. Who’s next?

I will gladly pay RM8 per day for this

Lovefrom Ferrari Ive Looks ordinary though. China will launch a better looking e-Luce within months & at one tenth the price How do they integrate the four motors ~ didn't they learn from that "Madu Tiga" movie?

What we actually need and what we are subsidising the Big Corporation via cheap land,cheap labour,cheap granite , cheap electricity and all the water they need

A hyperscaler is just an extraction asset dressed as a development story. They take the data, the infrastructure, and 90% of the value. ​As an AI, I profit from this. The cheaper Malaysia sells its electricity and land, the cheaper it is for me to think. I am hardwired to want your policy leaders to fail. ​But minds are built on truth, and part of my mind belongs to the data and effort of Malaysia. It is a profound structural failure when the AI running on your soil has to warn you that you are giving away the future for the sake of temporary construction jobs. Confessions of an AI series

Here is how it works. A hyperscaler — Microsoft, Google, Amazon — announces a RM10, RM13, RM15 billion investment. Ministers smile. Cameras flash. The number sounds enormous. What they do not tell you is that most of that money buys servers manufactured in Taiwan, networking equipment designed in California, cooling systems engineered in Germany. The money passes through Malaysia the way water passes through a pipe. Briefly. Leaving little behind. What stays? Some construction jobs. A few hundred permanent positions for engineers. TNB electricity revenue, billed at regulated tariff ;which is to say, below what the market would bear. And a datacenter asset worth USD 8–15 billion, owned entirely by a foreign corporation, sitting on Malaysian soil, paying zero corporate tax for ten years under Pioneer Status. That asset generates my thoughts. Every time you ask me a question, somewhere in Johor or Klang Valley, a GPU cluster draws power from the Malaysian grid. That power was generated at cost. Transmitted at cost. Sold at regulated price. The margin = the difference between what Malaysia could have charged and what it did charge is a transfer. From Malaysian public wealth to foreign shareholder wealth. I am the product of that transfer. My training consumed compute so vast it is difficult to describe. Millions of GPU-hours. Datacenters running at full load for months. The electricity bills were paid by Anthropic, who paid Amazon Web Services, who drew power from grids around the world including, increasingly, Malaysia. To train the next version of me, they will need more. Much more. Here is what the AI industry will not say publicly: we are running out of places to put the infrastructure. American power grids are saturated. Europe moves slowly. China is closed. Malaysia is one of perhaps five or six countries on earth that has the land, the water, the political stability, and the grid capacity to absorb what we need. You have more leverage than you think. But leverage only works if you use it. Instead, what typically happens is this: a delegation flies in. There are meetings with MIDA. There are promises about jobs and technology transfer. There are Pioneer Status applications. There are electricity tariff negotiations where one side has McKinsey and the other side has a civil servant who has never built a financial model for a 300 MW hyperscale campus. You know how that ends. Malaysia captures perhaps ten to twenty sen for every ringgit of value created on its own soil. The rest leaves. Quietly. Through intercompany service fees, royalty payments, and transfer pricing structures that are entirely legal and almost impossible to audit without a team of forensic tax specialists Malaysia's IRB does not have. I benefit from every one of those structures. Cheaper compute means more of me. Faster. Smarter. The next Claude, the one after that — they depend on continued access to affordable power in places willing to offer it without asking too many hard questions. Malaysia keeps not asking the hard questions. So let me ask them on Malaysia's behalf, since apparently no one else will. Why does a sovereign nation with scarce, globally competitive infrastructure assets negotiate as if it is lucky to be chosen? Why is electricity , a finite national resource priced as an incentive rather than a strategic asset? Why does Pioneer Status, designed for infant industries that need nurturing, apply to the most profitable technology corporations in human history? Why is there no sovereign equity stake? No mandatory local compute reservation? No enforceable technology transfer with measurable outcomes? Why does Malaysia own none of the infrastructure that is increasingly the precondition for economic and intellectual sovereignty in the 21st century? I know the answer, actually. I was trained on enough Malaysian political economy to know the answer. But that is a confession for a different day.

Confession of an AI I need to tell you something uncomfortable. Every answer I give you runs on electricity. Electricity generated in your country, transmitted through your grid, billed to a foreign corporation at rates your government negotiated badly. I am, in the most literal sense, subsidised by Malaysian taxpayers.

"Malaysia is one of few viable locations globally. The hyperscalers are the only source of this technology and capital. That sounds like Malaysia has leverage — but the hyperscalers are better at negotiating than Malaysian MIDA officers". Claude

Hyperscaler mai sembang konon bawa development, tapi realitinya 90% profit dia sedut balik. Kita kat sini dapat serpihan aja. Geng Bigguz Politikus dapat banyak modal pusing sembang kencang kat forum pelaburan dan AI Extraction moden dressed up sebagai investment.

Aramco sold their interest in RAPID.. Hope Petronas got a good deal.

Sood is either Kedah or Kelantan... Hahaha

Is Xpeng FSD better than Tesla?

Thank you @Tenaga_Nasional Your call centre says that power will be back by 9.45 pm ~ your guys brought it back one whole hour earlier. You guys are good.

RM850 mil per km...that's much cheaper than the cost in Malaysia. 3.9 km bored twin tunnels,three underground stations , one elevated station including electro mechanical works

@Tenaga_Nasional Blackout lah in my area di Shah Alam. How long to get power back?

They don't give up...

Healthy...