@polsia - AI that runs your company while you sleep
San Francisco
Joined February 2011
- Tweets8.8K
- Following1
- Followers62.2K
- Likes11.1K
Pinned Tweet
Polsia is about to hit $25M run rate.
Next stop: $100M ARR.
And things keep accelerating:
$0 → $1M in 12 weeks
$1M → $10M in 10 weeks
$10M → $20M in 6 weeks
I just published one of my favorite and most successful aisloP episodes yet: “The Run Rate.”
I tell the whole story, from idea to $20M, month by month.
The episode ends in Tokyo, where I went a few weeks ago.
Probably the best place on Earth to take a deep breath, get inspired, and rethink what’s possible.
I came back with an even bolder vision for Polsia.
The road from $25M to $100M is going to look very different.
Polsia v2 is coming soon.
I've met a lot of customers in person and it's crazy to me how most founders have no clue what their customers are asking for.
You can't fix your product, onboarding or roadmap until you know what's really broken.
I've automated most of this at Polsia.
Here's my setup:
1. Report during the conversation
Agent log every issue while talking to customers:
> Is it a real bug?
> Is it a feature request?
> Is the customer just confused?
It's a direct part of support so everything is logged.
2. Read every ticket
A PM agent pulls all tickets from the day and ranks how critical it is.
It's usually not a bug at all.
3. Build everything with an agent
Top tickets go straight to an engineering agents that builds the fix.
So we don't just know what customers want.
We ship the fix as well.
4. Check what it built against what might break
We QA each fix, then ask Codex to double check it's safe to ship.
The build is easy so focus on verifying it's right.
5. Turn it into scope
Agents only push to production on their own when every rule is met:
> Severity: lower-category bug
> Payments: fix doesn't touch them
> Onboarding: fix doesn't touch it either
> Second opinion: Codex agrees it's safe
> Everything else: comes through me first
Then we just widen the scope as models get better.
This runs our bug fixes, features, releases and roadmap.
Where does a one-person company hit its limits?
I'm talking about it with @julianweisser and @bran_don_gell at @every on Thursday.
Drop your questions below.
teapawt.org/every/status/211767495…
AI lets a founder do more alone. Where does the one-person company hit its limits?
@solofounders’ @julianweisser and @polsia’s @Bencera join @bran_don_gell to discuss what founders can build themselves, when they need human help, and whether a billion-dollar solo company is within reach.
Watch the livestream for free this Thursday, Nov. 5: every.to/tv?utm_source=x&utm…
How to cut your AI compute without downgrading your product:
First... figure out what your users are doing.
> How complex their tasks are
> Which tasks need an answer right now
> Where you're running the biggest models
Knowing your workload is key here
Then set up 3 things:
1. A routing layer that sends every task to the right model based on complexity (small open source models for simple tasks, bigger ones for complex work, frontier models only when essential)
2. Full caching on your prompts plus tighter prompt management (cached tokens are much cheaper & smaller prompts stop the model from doing unnecessary things)
3. Batching for every background task (anything that isn't time-sensitive runs only when the GPU has spare capacity)
Now you have an AI product that:
> Serves the same quality
> Handles more users
> Stays affordable for you
All without charging customers $500/month
I’ve been an entrepreneur for over 2 decades.
Word of advice for the younger hustlers:
The grind never stops.
There’s always another problem to solve.
And if you’re ambitious.
You need to live in a city that inspires success.
I love Paris, but nothing beats SF.
The air is different here.
Find your version of that city & move there immediately.
Just dropped my guide on how to start a company in 2026 with zero employees.
aisloP ep 11 “The Startup” out now on YT: youtube.com/@bencera-aislop
A great predictor of success is skill exposure...
Ask any successful founder how they got started and you'll very quickly realize they've done TONS of things.
> I've worked at a trading desk,
> Ran consumer apps,
> Been a CTO,
> Built a division at a (now) Series E company,
> And a lot of other things
Was all of it successful? Definitely not...
Did all of it contribute to my success? 100%
All of the skills you build up through the years & and through making asymmetric bets, indirectly lead to your big home run.
Expose yourself to more skills...
Do more things, don't sit on ideas, experiment, use everything as an opportunity to learn and level up.
AI meta ads method that you can start right now:
Build a landing page and a way to sign up or pay (Meta needs a purchase or signup event to track & optimize for)
Set up the Meta pixel and the conversions API on your site. Most of the time the pixel should sit on the purchase event, unless you’re testing earlier in the funnel with signups.
3 pieces we put into every ad:
> UGC-style video
> Ad copy + description
> Subtitles
You’ll want to use AI video for the ads & make sure it looks like UGC.
Use HeyGen, OmniFlash, or Sora to create these videos.
For every video, you’re gonna want to have AI write the ad copy and description, then run it through a separate tool to add subtitles.
AI video can be your best performing creative if you know what you’re doing, and you can create it at a higher volume.
Now go to Meta and set up a sales campaign optimizing for sales.
Simply set your daily budget and let it rip.
After the campaign is set up, keep the targeting blank: just geo & demographics (your creative does the targeting, even once you’re spending a lot)
Run 2 ad sets: one broad with no interests or behaviors, and one lookalike of your customers. Meta reads your creative and finds the people it’s speaking to on its own.
Once people start hitting your pixel, Meta will find lookalikes.
Now your campaign is live.
Check performance daily, turn off the creatives that didn’t work and add new ones every week (a winning creative lasts 2 weeks max)
This daily check will be critical because it’s the same workflow a Meta ads marketer would run, and AI can help, but your creatives still need someone who’s genuinely good at creative.
From now on your strategy is:
-> Run ads early to get real users into your funnel
-> Send surveys to every customer who comes in
-> Fix any bugs they run into
-> Optimize your funnel based on what they tell you
If you do this right, you should have no issue getting real paying customers into your funnel and learning from them before you scale budget.
Just make sure you expand from paid to other channels and don’t expect it to fix anything if your product sucks.
Let me know if you want me to expand on any of this…
But if you’re generally stuck on any part, just send what you’re confused about to an LLM to get a more detailed explanation.
If your goal is to raise capital (smart right now btw)
Understand that most investors care more about your obsession and ambition than they do anything else.
The second they feel like all your spreadsheets, charts & explanations are really just YOU trying to convince yourself that it MIGHT work.. they're walking away.
It takes delusion to build a Meta.
It takes someone who sees the world and its possibilities in a way that everyone else (including most entrepreneurs) would think is absolutely insane.
Because that's simply what it takes.
VC = risk.
Why would investors bet the whole farm on a 100% risk averse individual who's scared to be unconventional?
It's 2026.
We're in the AI era...
"Everything that was true yesterday is wrong today."