Founder @RLVNTStudios ⚡️/ all things Branding, AI, Coding, Robotics, Data layer for Physical AI/ @ProductHunt
Connect here!
Joined February 2018
- Tweets1.1K
- Following691
- Followers305
- Likes2.5K
Pinned Tweet
Short story in a long format video // no apologies asked
I’m going to keep my content as fresh and unedited as possible
Met an awesome awesome founder who is launching next month on @ProductHunt and he offered me a JOB!!
Watch this
#PHCommunity #ProductHunt #AKUpvotes
I’ve been so inactive here but I just read this report and I KNOW WHAT THE FUTURE OF JOBS FOR CREATIVES WILL LOOK LIKE
I remember reading @contraben’s and @contralabs’s report on DISTRIBUTION being the MOAT in 2026 and beyond
The next quarter of the 21st century will take CREATIVES beyond AI’s capacity to allow people to create. Anybody and everybody will be able to CREATE but the future belongs to people with DISTRIBUTION
So, if you’re in the creative field, I’d highly recommend building your own jobs by building a distribution network that will enable you to sell your creative ideas to people who really need it
I have been building mine on @ProductHunt and for over 1136 days, consistently. I’ve worked with the most incredible founders, helped them build their own distribution systems and grow their products
So excited to see the future of creativity
AI is making creativity more valuable than ever before and..
we have 3 years of data to prove it.
Introducing the Creative Intelligence 2026 report from @contralabs.
Full report: c87.to/creative-intelligence…
hello from the other side people
I’ve been quiet
not gone
just busy building, breaking, questioning, rebuilding and generally making a mess of things
also AI changed almost everything about this selfie
but apparently not enough to make me unrecognisable
so HI 👋 again
I have things to show you
What do you have to show to me rn?
🤖 Made with AI
Akriti retweeted
The next big robotics layer emerging is the harness.
NVIDIA is building agent harnesses. Zetta is building a closed-loop embodied harness, while Ludo uses a harness to connect VLM reasoning with memory, speech, navigation & manipulation.
A new layer is forming between foundation models and robots: managing context, tools, execution, verification, recovery and feedback.
The model is becoming the brain. The harness is becoming the operating system for physical intelligence.
And the real moat could be the loop: every robot action creates data that makes the system better.
Akriti retweeted
$399 for an open-source biped with the full RL + sim2real stack is kind of insane.
The interesting part isn't collecting walking videos. It's measuring the adaptation cost: how many real-world failures/trajectories does a policy need to go from sim → real → robust?
Could be a cheap physical AI benchmark hiding in plain sight.
We built a small biped robot you can teach new tricks to.
Train it in simulation, run it on the real thing. Meet Microduck 🦆
$399, shipping before Christmas.
pollen-robotics.com/microduc…
github.com/pollen-robotics/m…
Startup Day 3, 4, 5, 6 update: (sorry not sorry for the looong post)
Okay so I disappeared for a few days
Not because I stopped working
I was on vacation and escaped to the mountains for a bit, which was probably the only productive thing I did for my brain
But work did not stop
One thing I’ve been spending a LOT of time on is actually testing astrology readings
And honestly… I have questions
I’ve been trying different charts, different readings, different platforms and seeing how accurate they actually are
And the biggest problem I keep noticing is how generic so many of them feel
Like yes, technically this could apply to me
But it could also apply to 50,000 other people
Nobody likes generic stuff
Especially when you’re literally giving someone a reading about THEIR life
People want to know what is happening to them
Not what could happen to “people with this placement”
I’ve been looking at my own readings for a very long time now and most of them barely talk about what is actually happening in my life
And that’s the part I keep coming back to
How do we make this actually personal
How do we make someone read something and go
“wait… how the hell did it know that”
That’s the bar
Not another generic horoscope that tells you you’re entering a “period of transformation”
We’ve all been transformed enough
We also kept going back to the product, the features, the experience and what we can actually build differently
The cofounder has been slightly less sleepy this week btw
Slightly
Progress is progress
And somewhere between testing readings, figuring out the product and disappearing into the mountains, I realised we’re probably asking a much bigger question than
“how do we build an astrology app”
It’s more like
“how do we make astrology actually feel like it was made for you”
Anyway
I’m back now
Posting again
Building again
And hopefully not disappearing for another 4 days
Day 6 has just started
Still very intentionally not building your average astrology app
Akriti retweeted
25X less demonstration time, with zero training updates, yet it gets 71% of the performance. The model isn't just learning faster. The cost of teaching a robot is starting to collapse.
59% average task success from just 3–12 seconds of demonstration.
v/s
83% with 5 minutes of data + 10 gradient steps.
Startup Day 2 update:
We locked in the colour system today
Small detail, but it’s already shaping the entire feel of the product
Spent a lot of time digging into Gen Z behaviour… what they like, what they reject instantly, and honestly… how do you even design for a generation that changes its mind every 3 seconds
From a product standpoint, this was actually important
We’re not just building “another astrology app” // we’re trying to understand what modern spiritual users actually want in terms of UX, tone, and daily engagement loops
We also mapped out core features, especially the ones that don’t really exist in current astrology products yet
This is where things started getting interesting
We did a full competitive teardown of existing apps in the space
A few clear takeaways:
We LOVE @AstroSageSays
We LOVE @drikpanchang
The depth of information is genuinely strong, but they’re clearly built for a more traditional Vedic astrology audience, the most accurate one tbh, and often skew toward an older demographic
That’s a very different user intent than what we’re targeting
Also, the paywalls… are aggressive
From a retention standpoint, it creates friction very early in the user journey
@AstrotalkApp, on the other hand, is a completely different story
They’ve clearly nailed distribution and marketing
Strong funnel execution, strong demand capture
We also reviewed a few other Indian apps and noticed something interesting: globally, there are surprisingly few dominant players in this category
The market feels fragmented, which usually signals early-stage opportunity
That stood out
Very interesting signal
On a lighter note:
The cafe owner today realised I’m building a startup and literally offered me a permanent workspace in his cafe
No strings attached
Small moment, but it genuinely reinforces something I’ve been feeling: people do support founders in very organic ways when you’re building in public
It matters more than it sounds :)
Meanwhile, my cofounder sacrificed 2 hours of sleep to push this forward with me
Tragic but appreciated
Tomorrow we go deeper:
More features
More UX decisions
More market validation
More clarity on what we’re actually building
Day 2 done
Still very intentionally not building a generic astrology app
Also… quick “investor ID” style update for anyone new here:
I’m documenting this entire build-in-public journey day by day: product decisions, research, chaos, everything
So I’m curious:
If you’re following along, what are you building right now? What stage are you at?
Or are you just here for the ride and the aesthetic cafe updates
Be honest
And also… should I keep romanticising startup life one overpriced coffee at a time
Also open to cafe recommendations in Delhi NCR where founders actually work out of, would love to explore more of those spaces
Akriti retweeted
Physical AI evals are starting to become a real category.
Until recently, most VLA evaluation was basically:
LIBERO → success rate → leaderboard.
Now we're seeing a much broader stack emerge:
• Allen AI — unified VLA eval across 18+ simulation benchmarks
• LeRobot — one eval interface across multiple sim benchmarks
• PhAIL — real robots + production metrics like throughput and failures
• Robocurve — independent, real-world robot evaluation
• RoboDojo — bringing sim + real-world evaluation together
The interesting part isn't just more benchmarks.
It's the move from:
“Can the robot complete this task?”
to:
“How reliable, fast and general is this system in the physical world?”
I think independent physical AI evals are going to become increasingly important as robot models start looking more and more similar on demos.
Startup Day 1 updates —
Using @grok to find me people on @X who are inti the #Astrology space and are investing and building in this space aggressively
I follow more people than needed fr on almost all my socials
Unlearning the fuck I thought I knew about myself and my life
Working on the brand name and design
Trying to wake up my sleeping cofounder to remind him of what we’re building and scaling together
That’s it ig
I always end up stretching my to-dos
Uff
Lmk if yall are interested in this space, I’ll build a waitlist soon too
Might as well reveal my BIG AND FAMOUS @telegram PERSONALITY (who is also the sleeping CF here)
Not building your average astrology app, I promise
Akriti retweeted
More and more high-quality robotics datasets are being open-sourced.
This is great for the field, but it also makes one thing clear: collecting raw data alone may not be enough.
The interesting part will be figuring out what data actually matters for learning useful, general-purpose robot behavior.
Another major Chinese company had open-sourced a humanoid robot dataset--
EgoLive from JD.
This is a large-scale first-person perspective dataset recording real-world human tasks. It contains 1680 hours of high-fidelity binocular video data at 60fps, involving 65,866 data segments and 346 different tasks. The data comes from real-world business scenarios in retail, logistics, healthcare, and industry...
Earlier, JD established China's first embodied data collection community in Suqian, Jiangsu; simultaneously, their RoboBase project started construction in Guangzhou in the Q2, with plans to establish over 80 robot bases nationwide over the next five years.
Humanoid robot data is truly a new era gold mine.
AI just killed one of the easiest ways to impress people.
“We use AI.”
Cool.
So does everyone else.
Couldn’t agree more with Nithin Kamath.
AI is becoming the new website.
Having it isn’t impressive anymore.
The interesting question is:
What can you do with it that 10,000 other founders haven’t already done?
The moat is moving from the tool → to the taste.
From access → to insight.
From “we use AI” → “here’s what we built because we think differently.”
AI lowered the barrier to creation.
Now it’s raising the bar for originality
I’d been thinking of ideas to #buildinpublic
I have one
I have a great partner as well
Let’s see how this journey unfolds