Olivia
Online
Dylan
@Dylan_txa
CTO & Co-Founder (YC P26) | Find & Contact High-Intent Leads With AI 🌡️ | 1x Exit
Joined January 2026
20 Following
628 Followers
482 Posts
I'm 32. I built an AI startup called GojiberryAI to $2.8M ARR. Got accepted into YC.
If I had to start from 0, here's exactly what I'd do:
1. Sell it before I build it.
No code. Just a simple slide deck (mine was 6 ugly slides) explaining the problem, the solution, the outcome, and the price. I made my first $10k that way, before writing a single line.
2. Pick a painfully specific customer.
Not "B2B SaaS." Something like "founders at 20-person SaaS companies about to hire their first SDR." So specific that the right person reads it and thinks "that's me."
3. Start outbound on day one, but only to people showing intent.
Not scraped lists. People engaging with competitors, changing roles, raising money, or publicly posting about the exact problem I solve. That's the gap between a 1-2% reply rate and 25-40%.
4. Lead with value, never a calendar link.
Send a blueprint, not "got 15 minutes?" Let the resource do the selling, and the trial becomes the obvious next step instead of a pitch.
5. Pick ONE channel and go deep.
For us it was outbound first, then Reddit (10M+ organic views), then LinkedIn lead magnets. I wouldn't touch a second channel until the first one was clearly working.
6. Talk to customers every single day.
The product doesn't matter until you understand the problem better than they do. Spend 90% of every early call listening, not demoing.
7. Only build once people are actually paying.
Then keep it dead simple and price it to sell itself. We landed on $99/mo with a free trial, so the funnel runs without me dragging anyone onto a call.
8. Do this relentlessly for about 12 months.
That’s roughly how long $0 to $2.5M took us.
Bootstrapped.
No outside funding.
Most founders don’t lose because they can’t build.
They lose because they build too early, sell too late, and quit the channel before it compounds.
One motion design video did 900k views and added $30k MRR.
Most SaaS founders post talking-head videos. Or text posts. Or static graphics.
Our video was under a minute. It showed:
→ A LinkedIn profile lighting up with signal alerts in real-time
→ Each signal scored 0-100
→ The top scorers auto-routed into a DM sequence
→ Replies handled inside the same workflow
The product story told itself in under a minute.
Production cost: a couple thousand with a freelance motion designer.
Distribution: organic. We posted it with a one-line hook.
What I learned:
→ The format matters more than the platform. Same idea posted as a screenshot would've done a fraction of the numbers.
→ Single-feature videos beat overview videos. Pick one thing the product does well. Show it doing that thing.
→ Cost of motion design has collapsed. What used to cost $20k now costs $2k.
We're producing one motion video per month now.
My customers have sent 5M+ LinkedIn DMs.
The campaigns that book the most demos usually follow the same pattern.
They don't start with a pitch.
They don't ask for a meeting.
They offer something useful first.
One campaign booked 12 demos in 5 days with a 56% reply rate using a simple blueprint strategy.
The first message looked like this:
"Hey [Name], I put together a quick blueprint on how I'd approach [goal/problem] at [Company]. Happy to send it over? Just reply YES if interested"
No pitch, no calendar link and no long explanation.
Just a relevant offer.
Most people say yes because they're curious to see the ideas.
Once they reply, send the blueprint.
The mistake most people make is treating the blueprint like a disguised sales deck.
The best-performing blueprints are packed with actual value.
They contain a breakdown of the opportunity, a few concrete recommendations, examples from similar companies, short Loom videos explaining the thinking, relevant case studies, and resources the prospect can use immediately.
The goal is not to convince them to buy.
The goal is to make them think:
"These people clearly understand my business."
Of course, the blueprint should include a few clear next steps.
A link to book a call.
A link to the website.
A way to continue the conversation.
What's interesting is that many prospects never reply after receiving the blueprint.
Instead, they read it, visit the website, consume more content, and book a meeting themselves.
The dynamic changes completely.
You stop chasing the prospect.
The prospect starts evaluating you.
The call feels much more like inbound than outbound because they've already seen your thinking before the meeting even starts.
Seven days later, follow up with everyone who received the blueprint.
A simple message is enough:
"Hey [Name], just wanted to see if you had a chance to review the blueprint I sent over."
A blueprint sent to the right person at the right moment doesn't feel like cold outreach.
It feels like free consulting.
That's how LinkedIn reply rates go from the usual 5-15% to 25-40%.
![romanbuildsaas's tweet photo. My customers have sent 5M+ LinkedIn DMs.
The campaigns that book the most demos usually follow the same pattern.
They don't start with a pitch.
They don't ask for a meeting.
They offer something useful first.
One campaign booked 12 demos in 5 days with a 56% reply rate using a simple blueprint strategy.
The first message looked like this:
"Hey [Name], I put together a quick blueprint on how I'd approach [goal/problem] at [Company]. Happy to send it over? Just reply YES if interested"
No pitch, no calendar link and no long explanation.
Just a relevant offer.
Most people say yes because they're curious to see the ideas.
Once they reply, send the blueprint.
The mistake most people make is treating the blueprint like a disguised sales deck.
The best-performing blueprints are packed with actual value.
They contain a breakdown of the opportunity, a few concrete recommendations, examples from similar companies, short Loom videos explaining the thinking, relevant case studies, and resources the prospect can use immediately.
The goal is not to convince them to buy.
The goal is to make them think:
"These people clearly understand my business."
Of course, the blueprint should include a few clear next steps.
A link to book a call.
A link to the website.
A way to continue the conversation.
What's interesting is that many prospects never reply after receiving the blueprint.
Instead, they read it, visit the website, consume more content, and book a meeting themselves.
The dynamic changes completely.
You stop chasing the prospect.
The prospect starts evaluating you.
The call feels much more like inbound than outbound because they've already seen your thinking before the meeting even starts.
Seven days later, follow up with everyone who received the blueprint.
A simple message is enough:
"Hey [Name], just wanted to see if you had a chance to review the blueprint I sent over."
A blueprint sent to the right person at the right moment doesn't feel like cold outreach.
It feels like free consulting.
That's how LinkedIn reply rates go from the usual 5-15% to 25-40%.](https://pbs.twimg.com/media/HKAEEela4AApwoJ.jpg)
I'm 32. I built an AI startup called GojiberryAI to $2.8M ARR. Got accepted into YC.
If I had to start from 0, here's exactly what I'd do:
1. Sell it before I build it.
No code. Just a simple slide deck (mine was 6 ugly slides) explaining the problem, the solution, the outcome, and the price. I made my first $10k that way, before writing a single line.
2. Pick a painfully specific customer.
Not "B2B SaaS." Something like "founders at 20-person SaaS companies about to hire their first SDR." So specific that the right person reads it and thinks "that's me."
3. Start outbound on day one, but only to people showing intent.
Not scraped lists. People engaging with competitors, changing roles, raising money, or publicly posting about the exact problem I solve. That's the gap between a 1-2% reply rate and 25-40%.
4. Lead with value, never a calendar link.
Send a blueprint, not "got 15 minutes?" Let the resource do the selling, and the trial becomes the obvious next step instead of a pitch.
5. Pick ONE channel and go deep.
For us it was outbound first, then Reddit (10M+ organic views), then LinkedIn lead magnets. I wouldn't touch a second channel until the first one was clearly working.
6. Talk to customers every single day.
The product doesn't matter until you understand the problem better than they do. Spend 90% of every early call listening, not demoing.
7. Only build once people are actually paying.
Then keep it dead simple and price it to sell itself. We landed on $99/mo with a free trial, so the funnel runs without me dragging anyone onto a call.
8. Do this relentlessly for about 12 months.
That’s roughly how long $0 to $2.5M took us.
Bootstrapped.
No outside funding.
Most founders don’t lose because they can’t build.
They lose because they build too early, sell too late, and quit the channel before it compounds.
hot take:
the best GTM hire for an early startup is still the founder posting every day.
nobody can explain the pain better than the person building the product.
https://t.co/oGjCA2L4KR
Unpopular opinion: most founders don't actually have a distribution problem.
They simply aren’t willing to put themselves out there.
They don’t want to DM strangers
(we send 300+ linkedIn messages daily across all our accounts)
They don’t want to post consistently when nobody is paying attention
(we post 12 pieces different pieces of content every day)
They don’t want to ask creators for partnerships.
(2 B2B influencers post about us daily)
They don’t want to spend weeks answering questions in communities before mentioning their product.
(We spend 1 hour per day responding to reddit comments)
They don’t want to send the first cold email.
(We send 7k+ cold emails daily)
They want a clean system. A playbook they can follow. A dashboard that tells them exactly what’s working.
Something that feels predictable, repeatable, and easy to scale without constantly putting themselves in uncomfortable situations or risking rejection.
But most early distribution feels messy while you're doing it.
You leave a comment and nobody responds.
You send 50 DMs and get a handful of replies.
You share something in a community and get ignored.
You record a Loom and never hear back.
Then one thing works.
Then you do it again.
Then you build a process around the parts that worked.
That's exactly why we built GojiberryAI.
So that founders and sales teams can turn outreach that works well into a repeatable system thanks to intent signals.
But let me tell you something:
If you do the minimum, you'll get the minimum.
No tool changes that.
No AI changes that.
No playbook changes that.
Some 20-year-old founder with nothing to lose, endless energy, and an unhealthy obsession with winning will happily steal your customer.
Even if their product is 3x worse.
The idea for GojiBerryAI came from me manually stalking people on LinkedIn.
Not in a creepy way lol.
In a “we have no money for GTM tools so I guess I’m doing this myself” way.
At my last startup, I couldn’t afford the fancy software.
So I would search for ecommerce founders and look at who was engaging with posts about the exact problem we solved.
People commenting. People liking. People complaining. People asking questions. People tagging teammates.
Basically anyone showing the problem was already in their head.
Then I’d reach out.
And it worked way better than it should have.
If I messaged a random founder from a list, most of the time I got ignored.
If I messaged a founder who had just commented on a post about the pain we solved, they would actually reply.
Same offer. Same person sending it. Probably the same mediocre cold DM.
But completely different result.
At first I thought it was just a scrappy growth hack.
Then it became pretty obvious it was the whole company.
We didn’t need another database.
There are already enough places to find names, titles, companies, and emails.
The hard part is knowing when someone actually cares.
That’s what GojiBerryAI is built around.
Find the people showing intent right now.
Then help you reach out while the problem is still fresh.
That’s when outbound starts feeling a lot less random.
If I had to find customers for a new startup, here's exactly what I'd do:
1) define your ideal customer profile in one sentence sharp enough that your cofounder could repeat it back word-for-word
2) write down the 5 buying signals that mean "this person is in market right now"
3) skip cold lists. start with the people showing those signals.
4) pick one channel. linkedin if you sell to b2b. master it before you add email.
5) never pitch in the first message. open with the signal, not the offer.
6) build the qualification step before you build the pipeline.
7) content and outbound feed each other. run them together.
8) track reply rate by signal type, delete the bad ones
9) check pipeline coverage every week.
10) let ai run the system. spend your hours on the calls that actually close.
11) Right person. right timing. right message. in that order.
Most teams build the system in the opposite order. that's why most outbound is broken.
introducing GOJIBRAIN
Find and contact your perfect customers in one prompt
This is the end of scraping, enrichment headaches, and cold prospecting.
Comment "GOJI" and I'll personally send you 14 days of our paid plan, completely free.
I'm 30. I built an AI startup called GojiberryAI to $2.5M ARR. Got accepted into YC.
If I had to start from 0, here's exactly what I'd do:
1. Sell it before I build it.
No code. Just a simple slide deck (mine was 6 ugly slides) explaining the problem, the solution, the outcome, and the price. I made my first $10k that way, before writing a single line.
2. Pick a painfully specific customer.
Not "B2B SaaS." Something like "founders at 20-person SaaS companies about to hire their first SDR." So specific that the right person reads it and thinks "that's me."
3. Start outbound on day one, but only to people showing intent.
Not scraped lists. People engaging with competitors, changing roles, raising money, or publicly posting about the exact problem I solve. That's the gap between a 1-2% reply rate and 25-40%.
4. Lead with value, never a calendar link.
Send a blueprint, not "got 15 minutes?" Let the resource do the selling, and the trial becomes the obvious next step instead of a pitch.
5. Pick ONE channel and go deep.
For us it was outbound first, then Reddit (10M+ organic views), then LinkedIn lead magnets. I wouldn't touch a second channel until the first one was clearly working.
6. Talk to customers every single day.
The product doesn't matter until you understand the problem better than they do. Spend 90% of every early call listening, not demoing.
7. Only build once people are actually paying.
Then keep it dead simple and price it to sell itself. We landed on $99/mo with a free trial, so the funnel runs without me dragging anyone onto a call.
8. Do this relentlessly for about 12 months.
That’s roughly how long $0 to $2.5M took us.
Bootstrapped.
No outside funding.
Most founders don’t lose because they can’t build.
They lose because they build too early, sell too late, and quit the channel before it compounds.
Our website just got a major upgrade!
Now it finally feels like a company doing more than a few hundred dollars in revenue 😂
Our old Framer website, built in about 10 minutes, attracted over 500,000 visitors in 9 months and helped us reach nearly $3M ARR in less than a year.
Let's see if the new one can do even better 👇
Our new website for Gojiberry AI is live!
The old one was a Framer template we put together in about 10 minutes, and it still converted surprisingly well.
More than 400,000 people saw it, and the lesson is simple:
If you're solving a real problem, even a mediocre website can convert.
That said, we've finally decided to give it a proper upgrade.
Curious to hear what you think of the new one 👇
This guy says one person with AI will outperform a team of 200.
I disagree.
I'm 9 months into building an AI-for-sales SaaS.
$2.5M ARR.
0 outside funding.
YC backed.
I could not have done this alone.
Not even close.
Here's the actual split:
What AI does (most of the work):
→ defining ICP
→ detecting buying intent before competitors do
→ monitoring LinkedIn + company activity 24/7
→ identifying warm prospects automatically
→ enriching lead data instantly
→ writing contextual outreach
→ handling follow-ups in seconds
→ keeping the pipeline moving while we sleep
What the team does (the part that actually closes deals):
→ Earning trust on a call
→ Showing conviction when a prospect is on the fence
→ Handling objections in real time
→ Actually closing
Our team grew like this:
3 Customer Success Managers after $10k MRR.
A Product Manager after $25k MRR.
Growth, Sales, Engineering hires after $150k MRR.
If we'd tried to do this with 1 person and a stack of AI agents, we'd be nowhere close to $2.5M.
Hassabis is half right.
The team gets smaller and the leverage per person goes up.
But "1 person beats a team of 200" assumes the only thing that matters is execution speed.
What actually matters is:
→ Trust
→ Taste
→ Judgment
→ Closing
AI is still terrible at all four.
The companies winning in 2026 are not 1-person companies.
They're small teams where every person operates with 10x the leverage they had two years ago.
Demis Hassabis: "In the near future, one person who knows AI will outperform an entire startup team"
I've watched hundreds of AI talks, this 60-minute Cambridge lecture is the one I wish I had seen a year ago
this is the Nobel Prize winner in Chemistry, CEO of Google DeepMind and the guy who made AI solve biology
here's the part I can't stop thinking about:
> the AI you're using today is the dumbest it will ever be
> in 5 years the gap between people using AI and people who aren't will be impossible to hide
> companies will run on 10 people doing what 200 used to do
> the ones who get there first won't be the smartest, they'll be the ones who started right now
right now the average person opens Claude, types something, gets an answer, closes the tab
they think they're using AI, but they're using maybe 10% of it
I turned his lecture into 18 steps to actually use Claude the way it was designed, copy-paste prompts included
full guide in the post below.
GojiberryAI now runs on Claude Opus 4.8
Step 1: Agents find high-intent leads.
Step 2: Agents craft personalized messages.
Step 3: Agents reach out to your future customers via email and LinkedIn.
Step 4: Get more conversations and book more demos.
Comment “Opus 4.8” and I’ll send the setup guide.
@romanbuildsaas opus 4.8
Introducing Outreach Agents powered by Claude Opus 4.8
Step 1: Agents find high-intent leads.
Step 2: Agents craft personalized messages.
Step 3: Agents reach out to your future customers via email and LinkedIn.
Step 4: Get more conversations and book more demos.
Comment “Opus 4.8” and I’ll send the setup guide.
Startup and SaaS founders: there’s a cheap way to get traction on LinkedIn for almost free, and barely anyone is talking about it.
We generated 4M impressions in 6 months from a single fresh LinkedIn account.
Here’s the exact blueprint we followed to add ARR to our SaaS using LinkedIn:
1. Post lead-magnet content only.
Create posts where people need to comment to receive a resource (blueprint, video, template, etc.).
Find the best-performing lead magnet posts in your niche, rewrite the angle, and create a genuinely valuable resource.
2. Create an outstanding blueprint.
Include videos, screenshots, use cases, and add a self-onboarding link + demo link.
The value has to be insane.
People should feel like they should’ve paid for it.
3. Get 20 friends, colleagues, or anyone in your network to engage early.
Fresh accounts need initial traction, so you need to kickstart the post.
4. When people comment, reply with the document directly in the comments.
1 comment = 1 lead
1 good post = 100k+ views
5. Post twice per day until one goes viral.
6. Extract everyone who liked or commented on the post with GojiberryAI and run multichannel outreach on LinkedIn + email.
You can easily book 5–10 demos per week this way.
Also target:
* people engaging with competitors’ posts
* people showing buying signals
* people actively discussing the problem you solve
7. People will read your blueprint, remember you, self-onboard, or book a demo.
8. Repeat the process across multiple accounts.
9) Grow your ARR, get into YC and do an IPO
(Not fully sure about step 9 yet 😅)
20 days before Demo Day at @ycombinator and it feels like the last few months moved at 10x speed.
Since joining YC, we:
• grew from $800k to $2.5M+ ARR
• scaled from 700 to 2,000 customers
• stayed profitable while continuing to grow aggressively
• expanded the team from 3 to 7 people
• launched our AI GTM Operator that autonomously runs outbound and improves over time
• now have 500+ teams using Gojiberry AI agents daily
• saw customers start booking 10+ demos per week automatically
• onboarded a 25-person sales team doing $200M+ ARR running autonomous outbound on Gojiberry
• received dozens of inbound investor messages
• and probably sacrificed a few years of sleep in the process 😇
What’s exciting is that it still feels incredibly early.
We think outbound is about to change massively over the next few years.
Less manual work. Less list scraping.
Less “spray and pray.”
More intent signals, autonomous workflows, and AI systems that actually learn what converts.
Right now we’re fully focused on building, shipping fast, and helping customers scale.
Can’t wait to show what we’ve been working on at Demo Day.
Last Seen Users on Sotwe
Most Popular Users
Elon Musk
@elonmusk
240.1M followers
Barack Obama
@barackobama
119.3M followers
Donald J. Trump
@realdonaldtrump
111.6M followers
Cristiano Ronaldo
@cristiano
108.9M followers
Narendra Modi
@narendramodi
107M followers
Rihanna
@rihanna
97.3M followers
NASA
@nasa
92.1M followers
Justin Bieber
@justinbieber
90.6M followers
KATY PERRY
@katyperry
86.8M followers
Taylor Swift
@taylorswift13
80.6M followers
Lady Gaga
@ladygaga
72.1M followers
Kim Kardashian
@kimkardashian
69.4M followers
YouTube
@youtube
68.6M followers
Virat Kohli
@imvkohli
68.5M followers
Bill Gates
@billgates
63.4M followers
The Ellen Show
@theellenshow
62.5M followers
CNN
@cnn
61.9M followers
Neymar Jr
@neymarjr
61.1M followers
X
@x
60.9M followers
Selena Gomez
@selenagomez
59.9M followers