Patel Dhruv

Ex-@amazon | Seeking tech/AI roles 💼 |
Learning AI & building projects 🤖 |
Sharing tech & lifestyle insights 🌐 | Canada 🇨🇦 #JobSearch #AI

[Patel Dhruv @Dhruv_webDev]

Now , Intelligence = Self-believe

[Patel Dhruv @Dhruv_webDev]

@GetThatCash1218 @ArchiveExplorer @grok Same background , same life update videos ! Same face 🤣

[Dwarkesh Patel @dwarkesh_sp]

It was very fun to hear Ron tell the stories of early Jane Street.

Their first cluster was a pile of 6 Dell boxes in their office.

And it was important to them that it be physically in their office so that if something went haywire, they could just physically unplug the machine.

Goes without saying that they can no longer fit their 100s of k of GPUs in their office.

[Patel Dhruv @Dhruv_webDev]

@VaibhavSisinty @milindlabs Checkout its original version called " @heyclicky "

[Patel Dhruv @Dhruv_webDev]

3.)
The exchange rate is not calculated by current google rate , they have follow this rate for all trade with getting profit from both side of buying and selling stock .

[Patel Dhruv @Dhruv_webDev]

Hey @Wealthsimple, can someone explain this math? I made a $200 USD profit on $INTC, but your "exchange rates" turned it into a $393 CAD loss. 📈

#Wealthsimple #Investing #StockMarketwin
Read full thread 👇

[Patel Dhruv @Dhruv_webDev]

2.)
My $200 stock profit was instantly swallowed by a
5-cent artificial currency spread. It feels incredibly deceptive to advertise commission-free wealth building while wiping out retail traders with hidden FX spreads.

[Patel Dhruv @Dhruv_webDev]

Claude is getting cheaper and cheaper !

[Patel Dhruv @Dhruv_webDev]

Claude is getting cheaper and cheaper !

[Patel Dhruv @Dhruv_webDev]

@heyclicky @heyclicky can we get more ai power with Claude or chatgpt api !

It a whisper-flow with
CTO built in !

[Viraj Sheth @viraj_sheth]

time to sell the company I guess…

[Patel Dhruv @Dhruv_webDev]

Reiner Pope reverse-engineered the entire frontier AI infrastructure stack from publicly available API prices.
Full 2-hour blackboard lecture with @dwarkesh_sp is one of the most info-dense interviews I've ever seen.
Watch it if you want to actually understand AI works.

[Patel Dhruv @Dhruv_webDev]

🧵 7 things from Reiner Pope (MatX CEO) on Dwarkesh's pod:

→ Models 100x over-trained
→ Batch size = ~300 × sparsity
→ Inference "train" every 20ms
→ Gemini pricing leaks KV cache
→ Output 5x costlier than input
→ Context stuck at 200K
→ Google on SPINNING DISKS?

👇

[Patel Dhruv @Dhruv_webDev]

7/ Anthropic's cache pricing tiers reveal the memory hierarchy.
Cache for 5 min vs 1 hour → different prices.
Working from drain time (capacity ÷ bandwidth): • 5 min ≈ flash storage • 1 hour ≈ SPINNING DISK
Reiner: "I'm shocked to see spinning disk being used at all."