USA🇺🇸 is opposing FCRA
UK🇬🇧 is opposing FCRA
Pakistan🇵🇰 is opposing FCRA
Missionaries are opposing FCRA
Opposition is opposing FCRA
Churches are opposing FCRA
NGOs are opposing FCRA
India🇮🇳 must pass FCRA to restrict the flow of shady foreign funds🔥
How much of India's infrastructure was built after 2014?
• 100% of Dedicated Freight Corridors
• 98% of Solar Capacity
• 85% of the Expressway Network
• 79% of Tap Water Access
• 75% of Metro Rail
• 71% of Port Capacity
• 69% of Railway Electrification
• 60% of 4-Lane National Highways
Infrastructure is built over decades.
But some decades build more than others.
@stark0xbt With all due respect to JH, he is full of sh*t. All the reasons he listed were either unproven, liabilities or losing market share. Basically they realized SEC is toothless in trump-landia so they could get away with saying anything.
Local AI hardware = capacity * bandwidth * software stack
- Capacity tells you what fits
- Bandwidth tells you how hard the box can breathe
- The software stack tells you how much of the spec sheet you can actually cash out.
Hardware by Memory Bandwidth
- Mac Studio M3 Ultra: up to 512GB @ 819 GB/s
- RTX PRO 6000 Blackwell: 96GB @ 1792 GB/s
- RTX 5090: 32GB @ 1792 GB/s
- RTX 4090: 24GB @ 1008 GB/s
- RX 7900 XTX: 24GB @ 960 GB/s
- Radeon PRO W7900: 48GB @ 864 GB/s
- AMD Radeon AI PRO R9700: 32GB @ 640 GB/s
- Intel Arc Pro B65: 32GB @ ~608 GB/s
- Tenstorrent Wormhole n300: 24GB @ 576 GB/s
- Tenstorrent Blackhole p150: 32GB @ 512 GB/s + 800G
- MacBook Pro M5 Max: 460-614 GB/s
- MacBook Pro M5 Pro: 307 GB/s
- DGX Spark: 128GB @ 273 GB/s (coherent + CUDA)
- Mac mini M4 Pro: 273 GB/s
- Ryzen AI Max / Strix Halo: ~256 GB/s (~96GB usable GPU)
- MacBook Air M5: 153 GB/s
- Snapdragon X2 Elite: 152-228 GB/s
- Intel Lunar Lake: 136 GB/s
- Snapdragon X Elite: 135 GB/s
- Mac mini M4: 120 GB/s
- Arc Pro B60: 24GB @ ~456 GB/s
Verdict
- GPUs are still the bandwidth kings
- Apple wins: stupid amounts of memory, don't want to shard across GPUs
- Apple loses: when raw tokens/sec & concurrency matter more
- DGX Spark: coherent memory + NVIDIA stack
- Strix Halo / Ryzen AI Max: first real x86 unified-memory contender
- Tenstorrent: fully OSS stack, excited to see this mature
Fitting != serving
Even if it fits, you still pay for
- bandwidth during decode
- KV cache growth
- dequantization
- batching + concurrency
- scheduler quality
- framework overhead
The only mental model that matters:
1. What must fit?
2. What bandwidth tier do I need?
3. What software stack can actually deliver it?
In short:
- NVIDIA -> fastest raw speed
- Apple Studio M3 Ultra -> biggest one-box memory
- Strix Halo -> first real x86 unified
- DGX Spark -> coherent NVIDIA dev appliance
- AMD / Intel Arc -> rising alternatives
- Tenstorrent -> fully opensource stack
Ask: "which bottleneck am I buying?"
Do not ask: "which hardware is best?"
Abhijeet Dipke would think twice before coming on this show again 🚨
Abhijeet Dipke - Is being a Modi opposer crime in today's world? Is it a crime?
Reporter - No, it's not. But is supporting Modi a crime? Because you people behave like supporting Modi makes you no less than a criminal
Abhijeet Dipke - You are a reporter, you are media, you should be unbiased.
Reporter - What is Dhruv Rathee? He is a social media reporter. He has been against Modi for years. Why don't you call him out and tell him to be fair?
Abhijeet Dipke - you would not be saying all this if you had lived logically
Reporter - When a Hindu supports a leftist party, you all call them wise but when a Hindu supports a Hindutva party, suddenly it's dumb for you. Can you tell me even five Muslim creators who support BJP? I bet you can't even name one. I can name 50, 500 or 5000 Hindu creators who support a leftist party. It's Hindus who think logically and it's leftists who think with hatred
She made Abhijeet Dipke completely speechless. He couldn't even name one or two creators 🙏
In 1963, India told the man who'd go on to found Intel to make 94% fewer chips than he asked for.
Robert Noyce flew to Delhi to put Fairchild's first offshore plant in India. He wanted 10 million transistors a year.
The DGTD capped him at 600,000, asked him to keep equity under 50% and take a royalty ceiling, then sent him off. He set up in Hong Kong in 1965 at a dollar a day, then Korea. Five years later he co-founded Intel.
We didn't lose semiconductors in the 2020s.
We lost them in a Delhi meeting room in 1963, to a man who thought a production ceiling was industrial policy.
Sixty years and a PLI scheme later, we're still trying to buy back a seat we were once offered for free.
- Deepseek V4 flash GA: 284B that acts like a 1T
- MiniMax H3: Frontier video gen now open source.
- Qwen 3.8 27B: 27B that acts like a 400B.
This is probably the greatest week in open source LLM history.
Bro, this is actually a huge AMD moment.
Kimi K3 is so massive that it needs 16x B200s across two NVIDIA servers, but it fits inside one 8x MI355X AMD server because AMD gives you much more HBM memory per GPU.
That single AMD node hit:
- 952 tok/s total throughput
- 118 tok/s for a single user
- Nearly 4x the throughput per node of the 16× B200 setup
- Better performance per dollar than both B200 and B300
But the craziest part is that ROCm mostly worked out of the box.
@wafer_ai only had to make a few relatively small fixes. No months of kernel engineering. No custom kernels at all.
Even the slow time-to-first-token problem came down mostly to one attention kernel not loading because Kimi had 12 heads instead of a supported shape. They simply padded 12 to 16, used AMD’s fast existing kernel, and cut cold-prefill time by roughly 2–3x
AMD’s bet on packing more HBM into each server is going to become extremely important as frontier open models keep getting larger.
If AMD keeps improving ROCm and day-one model support, data-center operators will have to seriously consider these GPUs.
The CUDA moat isn’t dead yet, but this definitely puts a big dent in it.
🚨 REPORTER TO ELON MUSK: "You exaggerate the scale of Europe's problems!"
ELON: "Your people are asleep at the switch and do not realize the magnitude of the issue."
REPORTER: "Your interventions risk making it worse...it's authoritarian."
Days later: 60,000+ BARBARIANS INVADE SPAIN
As it turns out, Elon was right, and the Enemy of the People was wrong, despite thinking she knew everything!
Data doesn’t lie. Narratives do.
For a very long time, India was trapped in ‘learned helplessness,’ a defeatist mindset where ambition felt naïve. Our latest research quantifies the exact moment this flipped.
The decline in the Defeatist Narrative Quotient (DNQ) is the "leading indicator" for India's growth and rising expectations.
(1/6)
Udham Singh assassinated Michael O'Dwyer in London as a revenge for Jallianwala Bagh massacre. Today on this day, he attained the martyrdom through hanging execution.
M. K. Gandhi observed this as act of insanity & Swatantryaveer Savarkar praised to Udham Singh for his audacity.
i'm not fine after reading this
a guy with a laptop got his own silicon chip manufactured with software Google gave away.
the commercial version of that software rents for up to $1,000,000 a year.
he never signed a vendor contract and never paid for a seat. he wrote the logic, pushed it to GitHub, and nine months later a wafer came out of a fab. getting silicon with your name in it used to take a company. now it takes nine months.
his slot was 160 by 100 microns: about 1,000 logic gates, 8 inputs, 8 outputs, clocked past 50 MHz. more than 600 designs have already come back this way.
this is the open silicon stack. Google and SkyWater published a complete 130nm manufacturing process, DARPA funded the software that turns code into a fab-ready file, and all of it sits in public repositories.
turns out the whole flow fits in a git push:
- write the logic in Verilog, or drag gates around in a browser if you have never done this
- push to GitHub, an action runs synthesis, placement, routing and sign-off on the commit
- OpenROAD finishes place-and-route with no human in the loop, code to fab file inside 24 hours
- SKY130 carries the fab's real design rules, so passing the checks means it is manufacturable
- a few hundred designs share one wafer, which is the only reason a person can afford the masks
nobody puts this part in the thread: 130nm is roughly where the industry stood in 2001, and you wait six to nine months for silicon. this does not get you a GPU. it gets you a real object with your logic inside it.
bookmark this. the video is one of those designs opened layer by layer, and every rectangle in it exists in the file that was sent to the fab.
1-Gas cylinder on missed call.
2-Payment of online bills.
3-UPI payment.
4-DigiLocker.
5-Jan Aushadhi Kendra.
6-Ayushman Card.
7-New Parliament Building.
8-Free vaccination.
9-Pradhan Mantri Jan Dhan Yojana.
10-Surgical strike.
11-Operation Sindoor.
12-Triple Talaq Bill.
13-Aadhaar linking.
14-FASTag.
15-Chenab Bridge.
16-Jammu Tunnel.
17-Ganga Expressway.
18-Kashi Corridor.
19-Vishwanath Corridor.
20-Kedarnath reconstruction.
21-Shri Ram Temple construction
22-Sukanya Samriddhi Yojana.
23-Construction of toilets.
24-Doubling of railway tracks.
25-Expansion of railway stations.
26-Fewer days for ticket booking.
27-Ticket cancellation facility.
28-Golden Quadrilateral Scheme.
29-Bharatmala Project.
30-18 AIIMS construction
31-S I R
32-Mahatma Gandhi Setu reconstruction.
33-Bhupen Hazarika Setu.
34-Army modernization.
35-Make in India.
36-Drone revolution.
37-Direct money into farmers' accounts.
38-Ration digitization.
39-Free ration.
40-Kanya Sumangala Yojana
41-Implementation of GST.
42-Major reduction in GST.
43-Preservation of cultural heritage.
44-Construction of ponds.
45-Har Ghar Nal Se Jal.
46-22 hours of electricity.
47-Unprecedented rebate in income tax.
48-Abolition of Article 370.
49-553 new railway stations.
50-1500 new overbridges.
51-Removal of over 200 obsolete laws.
52-ED's record-breaking raids on corrupt individuals.
53-Year-long waiver on FASTag for 3000 rupees.
54-Implementation of VVPAT in EVMs.
55-Kartavya Path route.
56-Pictures of cultural heritage on currency notes.
57-Reconstruction of Nalanda University.
58-Indigenous production of Tejas.
59-Self-reliance in cell phone manufacturing.
60-Mudra Yojana.
61-Ujjwala Yojana.
62-Suraksha Bima.
63-390 new universities
64-7 new IITs
65-7 new IIMs
66-16 new IIITs
67-300 new medical colleges.
68-Construction of Pamban Bridge.
69-Rishikesh-Karnaprayag railway project.
70-Uttarakhand Char Dham road construction.
71-BrahMos production factory.
72-Assault rifle manufacturing in Amethi.
73-Vande Bharat modern trains.
74-Law on Waqf Board property.
75-Bio-toilets in railways.
76-Implementation of labor laws for workers' welfare.
77-Implementation of GRAM JI Yojana.
78-India has already surpassed China in road construction.
79-12 lakh youth provided government jobs.
80-Over 8 lakh commercial licenses issued.
81-80,000 new petrol pumps.
82-Construction of Shaurya Sthal.
83-Merger of railway budget into general budget.
84-Over 9000 new schools.
85-Bulldozers on illegal constructions.
86-Fencing on borders.
87-Open license to army against terrorism.
88-Vikas Kaushal Yojana.
89-Demonetization.
90-FASTag facility by abolishing toll gates.
91-Chandrayaan.
92-Mangal Mission.
93-Multi-satellite stabilization.
94-Reduction of flood-affected areas in Bihar.
95-End of queues.
96-Smart City expansion.
97-Urea coating for farmers.
98-PF withdrawal made convenient.
99-Creation of over 5 crore new jobs in private and domestic sectors.
100-If the farm laws had been passed, brokerage would have ended, and illegal encroachments by many Bangladeshis in mandis would also have ended.
Therefore, stand firmly with BJP and Modi ji, and understand the opposition's conspiracies.
This is absolutely true. Insta bans posts which are RW. Heck it even censors images with such words like Hindu. A few months ago I had made this graphic which went viral on X, but when I tried uploading it to my Insta account - it simply would not even allow me to post it even after trying dozens of times.
It’s absolutely 100% rigged against Hinduism or Indian nationalism.