This is genuinely one of the most powerful Claude stories I have come across.
A 62 year old man in India had severe migraines for 25 years. Only when lying down. Neurologists. Nephrologists. Brain MRIs. Blood thinners.
Nobody could explain it.
His nephew brought everything to Claude.
Claude noticed the one thing 25 years of specialists missed.
The headaches were positional. It connected dots across nephrology, neurology, and pulmonology simultaneously. Something no single doctor was positioned to do.
Then it asked one question nobody had thought to ask in 25 years.
Does he snore?
Loud snoring for 25 years. Every doctor dismissed it as dialysis fatigue.
Sleep study done. Breathing stopped 119 times per night. Oxygen dropped to 78%.
CPAP machine. Headaches gone.
Claude did not replace his doctors. It just looked at everything at once.
That was enough.
One Claude conversation cracked it.
Good decision India's current shift isn't a sign of lowering its guard; it’s a realization that total economic decoupling from China is nearly impossible if India wants to become a manufacturing superpower. By allowing investments like the $370 million hybrid engine project, India ensures that the money is tied up in physical factories on Indian soil, employing Indian workers, under strict majority control of local partners. It's a strategy of leveraging Chinese capability to build Indian self-reliance.
Grok के अनुसार इराक के एरबिल शहर के Franso Hariri स्टेडियम में इराकी फुटबॉल मैच देखते हुए इस महिला का Live वीडियो एक इराकी स्पोर्ट्स चैनल ALFORAT पर दिखाया गया। इस महिला की नेचुरल सुंदरता की वजह से वीडियो सोशल मीडिया पर तेजी से वायरल हो रहा है।
apple vs micron vs memory
Proxy investing looks easy until the OEM starts playing the real game.
OEM gets the customer.
OEM gets the brand.
OEM gets the pricing power.
Proxy supplier gets the order and then gets squeezed.
Volumes rise.
Working capital rises.
Capex rises.
Receivables rise.
But margins may not rise.
That is the trap.
Never buy proxy just because TAM is big.
Ask only one question:
Can this company keep margin when the OEM negotiates hard?
If no, it is not a proxy.
India is a weird market. With obesity rates shooting up, I'd have bet on the sales of generic GLP-1s exploding once their patents expired. They now cost about Rs 1,000–2,500 a month, and there's growing evidence pointing to benefits well beyond weight loss, including cardiovascular, metabolic, and liver health.
Yet, generic drugmakers are quietly cutting their sales targets by 25–30%. At Rs 1,000–2,500 a month, it's cheaper than a gym membership. The real problem seems to be retention. GLP-1s are injectables, and you have to keep taking them. If you stop, you gain back the lost weight. It seems like asking someone to stay on a weekly injection indefinitely is a much harder sell.
A few other friction points:
For a variety of reasons, Indians haven't taken to GLP-1s with the same enthusiasm as Western populations. Could it be because Indian physicians are conservative when it comes to prescribing newer drugs?
Self-injecting is a pain for most people, and that friction and inertia might be stopping them from starting in the first place.
Given that there are now GLP-1 pills, I'm wondering if the adoption curve will change.
A single AI data center swallows 27,000 tonnes of COPPER.
The AI boom is a raw-materials supercycle.
One 1-gigawatt AI data center contains 25 minerals and 179,940 tonnes of material.
Steel dominates by mass at 132,660 tonnes, a staggering 73.8% of the total.
But copper is the one to watch.
27,212 tonnes per facility, 15.1% of the build.
Aluminium follows at 14,320 tonnes, with graphite and nickel filling out the heavy metals.
Then come the transition metals.
Cobalt, silicon, lithium, tin, and lead.
The top 10 alone account for 99.7% of all the mass.
And that's before the Rare earths, gallium, germanium, tantalum, and indium are all in there too.
Now multiply this by the hundreds of data centers being built worldwide.
Every gigawatt of AI compute means tens of thousands of tonnes of copper pulled from the ground.
This is why copper miners are the quiet kingmakers of the AI era.
The model needs metal before it needs intelligence.
I wrote an analysis on the 2 best positioned US stocks for this , link in replies
#BuzzingStock | Sterlite Tech Hits 5% Upper Circuit After Launching ₹1,500 Cr QIP
- Base issue size at ₹1,200 Cr; Green Shoe Option Of `300 cr
- Base issue will dilute equity by around 4.2%, rising to 5.3% if the green shoe is fully exercised.
@reematendulkar@STL_Tech@CNBCTV18News
#STLTech #QIP #Stocks #CNBCTV18Market
Vishal Sikka Returns
Dr. Sikka has launched Hang Ten, an enterprise AI services company – the exact business of Infosys & TCS. It has bagged global enterprise clients. Indian IT’s revenue displacement is coming.
Poetic Revenge Is Here
a. Backed by Mayfield (Silicon Valley VC giant for AI), Saudi Aramco, and top angel investors, Dr. Vishal Sikka has launched Hang Ten Systems.
b. Hang Ten’s stated purpose: “To help large enterprises adopt AI, using an AI-native model – to build, change, and run enterprise software at lower cost and on shorter timelines.”
This is exactly what India IT companies do at the upper end (premium segment) of their services model (using an army of interchangeable labour).
c. CEO of Siemens Gamesa (one of the global clients Hang Ten has already bagged) said: “Every enterprise I know is looking for trusted guidance on AI and help with dramatically improving major operational programs. I am very excited to see Vishal and his team tackle this challenge with Hang Ten.”
Where It Competes with IT
a. Vishal Sikka is going after complex, high-value enterprise software transformations that traditionally generate the highest margins and the most prestigious client relationships for Indian IT.
b. AI-native enterprise modernization programs are the new flagship projects that will anchor relationships and produce upsell. If Hang Ten (and other Silicon Valley startups) win these projects, Indian IT’s most valuable client relationships weaken first.
c. This is the classic Silicon Valley disruption pattern: attack the premium end first, commoditize it, then go downstream. Indian IT’s $100 million+ AI transformation engagements are likely to be their first target.
d. Commercial Off-the-Shelf Software (COTS) lies at the core of Indian IT. IT firms earn tens of billions of dollars annually helping clients configure, customize, integrate, and test SAP, Oracle, and Salesforce implementations.
Hang Ten wants to skip the COTS layer entirely where possible or to AI-automate the integration where it cannot be skipped, and build enterprise-native capability faster and cheaper than any IT service provider could.
How Hang Ten Will Do It
a. Expert FDE Bench: Hang Ten announced it will use “FDE Bench” (Forward Deployed Engineers), which is the reverse of Indian IT. Infosys/TCS operate on a Bench model: a pool of engineers waiting for a client project. The Bench is a cost.
FDEs are the opposite: elite AI-native engineers deployed directly into client environments to identify and solve hard problems (rather than being told what to do.) Palantir pioneered this model for the Pentagon. Sikka is scaling it for global enterprise AI services.
b. Reusable Skills Library (Knowledge-Compounding Flywheel): Every enterprise AI transformation Hang Ten completes will generate reusable AI skills, agents, workflows. These can be deployed faster & cheaper on the next engagement.
As knowledge compounds, what takes 200 Infosys engineers six months today will eventually take Hang Ten 20 engineers and three weeks. (Sikka calls it “Creative Destruction at AI Speed.”)
c. Agent Orchestrators: Hang Ten will use engineers who master agent orchestration (i.e., acquire the ability to coordinate swarms of agents toward complex outcomes.) The outcome rises with each advance in model & orchestration infra.
d. Continuous Transformation: Unlike IT’s fixed-scope engagements, Hang Ten proposes a continuous AI-native development & operations loop. The enterprise software keeps continuously improving, adapting, and responding to business change. This is an AI partnership moat.
Water is in the Living Room
a. The top five Indian IT firms employ 1.5 million engineers. Indian IT contributes 8% of GDP, and brings in $264 billion export revenue. If even 5% revenue displacement happens, it is $13 billion evaporated in one year. Stakes are too high.
b. Last week, Vishal Sikka told CNBC India: “The wave is here.” Nobody understood what exactly he meant. (“Hang Ten” is a classic surfing maneuver to ride the wave.)
c. Floods do not occur overnight. Water levels rise slowly until the moment all exit routes are blocked and people are trapped. Indian IT has had its warnings. “Now the water has reached the living room,” said Dr. Sikka.
Whether the Indian IT industry rides the wave or gets buried in the water grave of its own making will start becoming clear in the next few quarters.
@arabicatrader
Self doubt comes when position size is bigger than clarity. read it again.
Before buying big, decide 3 things
What proves thesis right?
What proves thesis wrong?
What price/business action makes me reduce?
If only price is correcting, but earnings, guidance, sector tailwind and management commentary are intact — hold.
If business evidence is also weakening — don’t call it conviction. Exit.
And sometimes thesis is right, but size is too big for your mind. Then reduce 10–20%. That is not weakness. That is buying mental peace.