
building @cscadeIndex · prev @GluexProtocol
Earth
Joined September 2021
- Tweets395
- Following966
- Followers403
- Likes6.4K
I'm going to ETHGlobal Mumbai @ETHGlobal!
Use my referral link below to earn a discount on your stake!
Link: ethglob.al/7ZMWDS6W58
3
116
Mani R. Bajpai retweeted
IBW 2026 "2 General Tickets" Giveaway 🎟️
How to enter:
→ Like & retweet this post
→ Comment whatever you like
We'll pick winners in 72 hours. Wish you all the best! 🤞
(Must be following @Build3DAO and @IBWofficial)
We are excited to announce that we are an official Community Partner for @IBWofficial 2026!
📆 Nov 1-2, 2026
📍 Mumbai
Use our code: IBWXBUILD3 and get an instant discount on your tickets (and Devcon tickets as well 😉).
Check below for more details 👇
been looking at why things feel slow.
not the model. the box.
@brendangregg USE method. three questions for every resource.
utilization: is it busy.
saturation: is work waiting.
errors: is it lying to you.
90% CPU with no queue is often fine.
60% disk with a fat queue is the fire.
busy ≠ drowning ≠ dying.
those are three different clocks.
most dashboards show the first number and stop.
that is how you debug vibes.
i would rather check the queue.
brendangregg.com/usemethod.h…
i agree to disagree
My friend’s Mac was stolen today, but thanks to @MumbaiPolice for their rapid action, he was able to recover it.
1
4
427
People keep treating AGI/ASI like the finish line.
It is not.
Once we get AGI/ASI, we enter the era of smart mechatronics. After robots can take ownership of physical work, we enter full-fledged intelligent biotech.
AGI/ASI → advanced mechatronics → biotech boom
Robots already look smart in demos. Walk, pick, wave, follow a prompt. Most are still early. Narrow skills. Brittle hands. A human jumps in the second the scene changes. Software intelligence is racing ahead of body intelligence.
That is the gap.
@OpenAI @GoogleDeepMind @xai and the other frontier labs are compressing the first layer: planning. Models can write protocols, design parts, debug control stacks, simulate a factory line, and propose the next experiment. Cognition is getting cheap.
Cognition is not ownership.
A robot takes ownership when it can see a messy real environment, plan several steps, use its body without a scripted demo, recover when something slips, and keep working without a human babysitter.
That middle layer is mechatronics plus embodied intelligence. Motors, sensors, force control, dexterous hands, on-device vision-language-action models. Not a chatbot with legs.
This is why the order matters. AGI without reliable hands just produces smarter PDFs. The moment those models can drive whole-body robots, the bottleneck moves from “can we think” to “can we execute.”
@GoogleDeepMind is already saying this out loud with Gemini Robotics 2: vision + language → motor control, whole-body humanoids from feet to fingertips, new robot bodies in hours, multi-robot teamwork with platforms like @Apptronik. That is the bridge between software AGI and physical work.
@BostonDynamics @agilityrobotics @Tesla @figure_ai are the body companies. Hardware that has to survive a factory floor, not just a keynote. Until those machines can run a shift without a teleop engineer, we do not have smart mechatronics. We have impressive prototypes.
Skip this layer and AGI stays trapped in text. Nail this layer and the scarce loop becomes: token → motor.
Then biotech.
A scientist today: read paper → design experiment → pipette → wait → write notes → repeat.
An intelligent lab: model proposes the molecule or protocol → robot runs it 24/7 → sensors write the result back → model updates → next run.
That is already starting. Self-driving labs. Robotic chemistry. Wet-lab automation. Organoids. Drug screens. @Ginkgo and the closed-loop biology stack are early versions of machines owning the experiment cycle. Clinical and diagnostics automation is pushing the same idea into hospitals.
Humans set the goal. Machines take ownership of the cycle.
So the stack is simple:
AGI/ASI — cheap intelligence and planning
@OpenAI @GoogleDeepMind @xai
Smart mechatronics — robots that can actually do the work
@GoogleDeepMind @Apptronik @BostonDynamics @agilityrobotics @Tesla @figure_ai
Intelligent biotech — biology compiled by models and executed by machines
@Ginkgo + robotic / self-driving labs
Chat was phase 1.
Robots taking ownership is phase 2.
Biotech at industrial scale is phase 3.
If you are building, stop only chasing leaderboards. The scarce thing is the physical loop.
"i agree to disagree"
Mani R. Bajpai retweeted
in life, i have learnt one valuable lesson:
if you have an undying belief -- a delusion so big that everyone thinks you are crazy -- that you willl be filthy rich, then almost 95% of the times, you will succeed.
ofc this means you put your thoughts and actions to right use.
Mani R. Bajpai retweeted
JUST IN: $222,000,000 shorts liquidated from the cryptocurrency market in the past 60 minutes.













