Head of Product Growth @BeaconsAi, 2x Start Up Founder, Formerly Creator Growth + Ecosystem Lead @Hoyoverse | PM @TwitchRivals | Dev @Microsoft
Joined October 2010
- Tweets139
- Following402
- Followers654
- Likes726
shi. retweeted
Today, we’re announcing Zeroset’s $5.2M pre-seed round to deploy world models into enterprises, co-led by @GradientVC and @2048vc, with participation from Leblon Capital and others.
Frontier intelligence is capable of enormously complex tasks. And yet with an abundance of general capability and mandate, automation still requires enormous handwork. The models and systems that exist today lack an understanding of the world they act in.
Robotics faced a similar problem. Building capable robots required more than compute-scaling on policies; they needed representations of their environments, encoding state and how it changed under actions. Over time, those representations shifted from hand-specified states toward learned models of the world.
What is missing in the enterprise is a general system that represents work and learns from it. At Zeroset, we believe the same paradigm that enabled progress in robotics can be applied to represent enterprise “worlds.”
We open with Nebula, a system that captures the state and traces created as people and agents work across systems, using them to model how an enterprise works and changes over time.
These models contain the decisions, dependencies, and unwritten processes bespoke to a company. They are derived from the enterprise, and they should belong to the enterprise.
Companies should be able to use the best models for their work and train them on their own operating knowledge freely. The underlying system does not need to change.
Nebula is currently in closed research preview. If your team is automating complex work or exploring how to learn from the work your organization already does, please contact us.
We are building the system through which that intelligence becomes increasingly specialized to the enterprise, allowing agents to act through it, and models to learn from it.
If you’d like to join us in building this future, we’re hiring researchers and systems engineers.
Read more here: businessinsider.com/zeroset-…
shi. retweeted
there's nothing more i dread than hitting the call button...
the gap between AI that talks and AI that does is where the next wave of real value lives and imo voice is one of the most underrated frontiers of that shift.
just think about how much of the world still runs on phone calls that no one actually wants to make. booking appointments. following up on a quote. confirming an order. chasing a reschedule. handling a routine customer request. each one feels trivial but together they're a massive tax on time and attention esp for businesses.
that's what makes @Heycalle interesting. it takes a task end to end: you hand it an objective, it makes the call, works the dialogue, and tries to get the outcome. that "last mile" of actually executing is the hard, valuable part.
and the surface area is enormous e.g. customer support, scheduling, reminders, sales, collections, verification.
CALL-E is also running its first virtual hackathon in July. Great way to build with voice AI and geek out with people thinking about agents virtually!
AI tools are getting to the point where they don't just talk anymore — they can actually do things for you.
@Heycalle is basically an AI that makes phone calls on your behalf. You give it a task like booking something, following up, handling a customer request, and it takes care of the call and tries to get it done.
What I find most interesting is how many real-world workflows are still stuck on phone calls — support, scheduling, reminders, sales — that feel small individually but add up fast.
If you're interested, CALL-E is hosting its first virtual hackathon in July! It's a great chance to learn, build with voice AI, and connect with others who are curious about AI agents.
Perfect for devs, operators, no-coders, and curious minds alike.
Watch as I direct CALL-E to casually call me and give me an emotional support
Learn more here 👉 heycall-e.com/
shi. retweeted
the @WSJ just revealed how the biggest brands, VCs, and creators are using clipping as their most powerful growth channel.
here's what they found:
• clipping agencies are delivering views at CPMs that blow traditional paid media out of the water
• campaigns create organic viral moments with real word-of-mouth conversation in the comments, not just passive impressions
• marketers are now rebuilding their entire campaign strategy around a distribution-first approach
and @CloutedHQ just raised $7M for exactly that.
shi. retweeted
In a world where short-form video and attention is the modern currency of marketing, Clouted takes out the guesswork for making videos go viral.
Check out our interview with @TechCrunch covering our $7M seed round for @CloutedHQ.
Clouted wants to take the guesswork out of making short videos go viral techcrunch.com/2026/05/20/cl…
Funny enough, there's enough content demand on the platforms on the topic of viewbotting right now that if you are middle-sized streamer (500-1000ish) viewers, you can actually try to bait large public culture pages into covering you organically through having a friend viewbot you and then "out yourself". Thus creating actual organic traffic
Paid ads work when your UA funnels are actually built, and whoever is running them understands intent
Roblox bought their first 1500 users from Google for $1 each. You can tie the 300 million users today back to those original 1500:
US’ 996 is actually more hours than China’s 996
The thing they don’t tell you about 996 is that in China there’s a built in mid day napping culture from Noon to 2PM which doesn’t exist here
Ok, getting consistent ride quality of a Jaguar is kinda nice when using Waymo. This predictability has increasingly become top of mind when deciding on which app to hail a ride from
AI hallucinations are oddly human.
I got married last weekend, and my best man cited that because of the confidence in my spoken words, it would sometimes give the perception that hypothesis’s sound like fact. That in turn has lead us on many chaotic type II fun moments, and sometimes drama in the friend group
This is similar to AI, where the the agent sends you off on a wild goose chase because it was so confident on what it states that you feel it absolutely has to be true. But as more context gets added, sometimes you arrive right back where you started, and conversely causing a sev0 with the dev team
shi. retweeted
Today we’re launching @markopolo_ai an AI that creates deeply personalized campaigns that run themselves. Convert 30-40% of warm leads vs. the industry standard of 10-15%.
People don’t abandon because they don’t want the product. They abandon because the follow-up is generic, late, or in the wrong channel/language. So me, @RubaiyatFarhan and our team has built an engine that automatically repairs the leak.
Our platform reads each visitor's behavioral signals and generates individual follow-ups across email, SMS, and voice, automatically.
No generic workflows. No A/B testing. Just LLM-powered campaigns that understand your customers intent and adapt in real-time.
Here’s how modern brands use Markopolo to recover millions in lost intent:
1. Reads real shopper behavior in real time
Hesitations, comparisons, timing, language preference, a full behavioral fingerprint for every visitor.
2. Creates unique 1:1 journeys instantly
Email → SMS → WhatsApp → AI Voice — whatever that specific shopper responds to. No flows. No templates.
3. Follows up in the right tone, language & moment
If a shopper buys in Hindi and researches in English, Markopolo adjusts. If they only respond at 7 PM on WhatsApp, it adapts.
4. Fully autonomous campaigns
No workflow builders. No segmentation maintenance.
Just automated, deeply personal conversations that bring people back.
The result?
30–40% recovery consistently. If @markopolo_ai doesn’t beat your current recovery rate within 60 days, we’ll refund you.