If there is one thing you have to listen/read this weekend, then let it be this one https://t.co/OHgk3dpCj2 🎧🚴#GTC
— Abhay 🇸🇬🇮🇳 (@Abhay08)
Mar 21, 2026
RT @balajis: I’m going to make some obvious points.
(1) Blowing up all the oil infrastructure in the Middle East is an insane idea, and ma…
— Abhay 🇸🇬🇮🇳 (@Abhay08)
Mar 19, 2026
Interesting analysis by ent market cap (Wells Fargo / Feb 2026):
🔴 Oracle: 11GW / $0.46T = 23.9 GW per $T
⬛ Meta: 21GW / $1.79T = 11.7 GW per $T
🟧 Amazon: 29GW / $2.60T = 11.2 GW per $T
🟦 Alphabet: 35GW / $3.83T = 9.1 GW per $T
🟩 Microsoft: 27GW / $3.53T = 7.6 GW per $T— Abhay 🇸🇬🇮🇳 (@Abhay08)
Mar 15, 2026
An excellent surface-scratcher on datacenter networking stack. Devil is always in the details #iykyk https://t.co/tjBehNsfuM
— Abhay 🇸🇬🇮🇳 (@Abhay08)
Mar 15, 2026
The Infrastructure Layer Nobody Saw Coming: Agent Harnesses Are Eating Software https://t.co/Y0MzDj5zgX
— Abhay 🇸🇬🇮🇳 (@Abhay08)
Mar 14, 2026
The smartest bet in AI right now isn’t the model. It’s the plumbing around it.
We spent 2025 arguing about which LLM was better. GPT-5 vs Claude vs Gemini. Benchmarks, leaderboards, Reddit fights. Underneath all that noise, engineers quietly stopped asking “which model?” and started asking “how do I run 20 of them without losing my mind?” That question is now an infrastructure category. Agent harnesses. And nobody has clean answers yet.
@Suhail Pinning hopes on Groq and Cerebras
— Abhay 🇸🇬🇮🇳 (@Abhay08)
Mar 13, 2026
Sun ☀️ , my openclaw agent migrated itself from my MacBook to mac mini (which also is my workhorse nas) I had to ask it re read all my model api keys from obsidian. Took an hour. ☀️ has a new 🏡 – I am not ready for how Sun surprices me everyday
— Abhay 🇸🇬🇮🇳 (@Abhay08)
Mar 13, 2026
Energy is the foundation layer in this cake https://t.co/qdjDYqkxgW
— Abhay 🇸🇬🇮🇳 (@Abhay08)
Mar 11, 2026
🏃♂️➡️🎧😇 https://t.co/CEnFeubBO8
— Abhay 🇸🇬🇮🇳 (@Abhay08)
Mar 10, 2026