Yurii Filipchuk 🇺🇦 -- e/acc

Founder @ Stealth AI |TechStars AI Startup Weekend SF

I build Principle, a simulation engine that stress-tests business decisions before you commit capital.

I also run partnerships and sponsorships at Techstars Startup Weekend SF and Catalyst Bay. We're a community of 10,000 founders doing IRL events in San Francisco.

Chaotic Good is my operating system. Move fast, challenge conventional thinking, put powerful tools in more hands. I believe that's how you actually make the world better. Big believer in pay-it-forward culture. If you have an idea on how we can help you or you want to help us, reach out.

The LLMs you choose say something about you. I'm team Claude for work. NotebookLM and Gemini for research. ChatGPT and Groq for personal life questions.

Previously CEO of Party.Space. Spotify was my first client, then Google, Zapier, Grammarly, UC Berkeley. Raised $2M from top European VCs.

I advise early-stage founders on fundraising and GTM. Been on both sides of the table. Invested through CYFRD: 15 deals, 2 exits.

San Francisco. Born in Kyiv.

Community Lead: Vibes & Sponsorships | Pitching Mentor

At Techstars Startup Weekend SF AI, we're hosting a fast-paced, 3-day event where tech enthusiasts come together to bring AI startup ideas to life. My role is to create an energetic and welcoming atmosphere, build connections with important sponsors, and guide participants in making their pitches clear and impactful.

Let's play a game.

18,000 followers. No idea who 17,900 of you are.

LinkedIn told me to post "thank you for being part of my journey" but that felt weird.

So instead -- introduce yourself.

Drop a comment: → What you're building → What you're stuck on → Who you're looking for (investors, clients, co-founders, hires)

I believe in surface luck. The more random connections you make visible, the more "coincidences" happen.

Worst case: you meet someone interesting in the comments. 

Best case: someone here has exactly what you need.

I'll start:

Building OpenBabylon -- strategic simulation and foresight engine for enterprise decision-makers. Running pilots on M&A strategies and what happens when the AI bubble bursts. 

Looking for feedback from people in charge of strategy and strategy consultants.

Your turn.

Founder @ Principle | Strategic Foresight Engine | Chaotic Good

two types of people worked this weekend.

type 1: got trapped by a friday 5 pm “quick sync” they were too polite to dodge.

type 2: founders who don’t remember what weekends are.

if you’re type 1 — next time accept then reschedule. watch the meeting evaporate.

if you’re type 2 – no advice. just respect. the hustle doesn’t need fixing.

graphical user interface, text, application, email

Enterprises are treating AI like Ozempic – suppress the symptom and call it solved.

Headcount too high? AI agents.
Decisions too slow? AI summaries.
Strategy too fuzzy? AI dashboards.

None of this addresses why your org got bloated or slow in the first place.

One pilot we ran: exec team thought they had an “execution problem.” Simulation showed the real issue was misaligned incentives between two business units.

AI would’ve made them fail faster.

Are you solving symptoms or sytems? 🤠

graphical user interface, text, application, chat or text message

This month Catalyst Bay is hosting something special.

Immad Akhund (Mercury) and Rajat Suri (Tribe, Lima, Lyft) are recording their Founders in Arms podcast live -- with Max Mullen (Instacart co-founder) as a special guest.

Three founders who've collectively built companies worth billions. In one room. With 50 people max.

The format is what makes this different: 

1/ Networking to actually meet Raj and Immad before we start 
2/ 45 min conversation -- no fluff, real talk 
3/ 15 min live Q&A where YOU can ask the hard questions 
4/ More networking and food after

This is gated for a reason. Founders with revenue traction or at least pre-seed raised. Application required. We're keeping it tight so the conversation stays high-signal.

Huge thanks to the Founders in Arms team for trusting us with this. And to Juliana Chyzhova for leading this one, with Nazar Gulyk backing her up. Special shoutout to Bryan Wish for the intro that started it all.

Proud to be part of the journey and watch this community grow into something that attracts this caliber of guests.

And please -- don't @ me asking for a spot. Not this time. My DMs can't help you here. 😅

For those who applied but haven't heard back yet: if you have a killer question for the panel, drop it in the comments

PDF quotes won't survive 2026.

One founder built his entire company on killing them.

Salesbricks 🧱: $18.3M raised. Hypothesis Studio, Gaingels, Greenpoint. "Shopify for B2B SaaS." YC alumni network.

23 reports and predictions from VML, HBR, CB Insights, McKinsey, Gartner say the category explodes.

We ran them through our LLM world model at Principle. 

Here's the consensus:

1/ Agents Don't Read PDFs (VML, HBR) Walmart and Maersk already use AI agents to close thousands of negotiations → Agents query APIs, not attachments → PDF = invisible → Machine-readable checkout = answer-ready

2/ Reasoning Token Shock Kills Flat Pricing (CB Insights, InfoTech) AI compute costs up 20x → Salesforce pivots to "Flex Credits" → 64% of innovator IT departments mandate consumption-based → Per-seat pricing breaks
e

3/ The Sales Tax Gets Automated (McKinsey, Claude State of AI Agents) 30-50% time savings from workflow automation → 7-12% revenue increase → 48% of orgs cite internal process automation as highest-impact zone → Quote-to-cash is ground zero

4/ B2B Checkout Becomes Protocol (CB Insights, Stripe/OpenAI) Stripe + OpenAI launch "Agentic Commerce Protocol" → Infrastructure for machine-to-machine purchasing gets standardized → Structured data wins

5/ $720B SaaS → $4.6T Service-as-Software (InfoTech) Outcome-based billing replaces seat-based → Hybrid models (seat + consumption + drawdown) become default → Legacy billing tools can't handle complexity

6/ 90% of B2B Purchasing AI-Intermediated by 2028 (Gartner) $15T+ in business spend flows through agents → Procurement teams stop scheduling demos → The "Magic Hour" is 9 PM - 12 AM

Consensus says the "messy middle" of B2B sales -- quoting, contracting, billing -- gets eaten by whoever makes it machine-readable first.

Jonathan Festejo calls it "removing the Sales Tax." His data: 30% of revenue happens while sales reps sleep.

What's your take on this: Gartner says 90% of B2B spend goes through agents by 2028. Are we 18 months out from PDFs being a disqualifier?

No-code AI agents won't survive 2026.

Two MIT roboticists bet their dissertations on the opposite.

StackAI: $19.5M raised. 100+ enterprise customers. Top 5 U.S. Defense Agency. Nubank. MIT Sloan. YC W23.

25 reports and predictions from Nadella, Huang, Karpathy, Gartner, McKinsey say the category collapses.

We ran them through our LLM world model at OpenBabylon. Here's the consensus:

1/ Hyperscalers Bundle Everything (Nadella, Gartner) 

Microsoft Copilot Studio ships native → AWS Bedrock adds agents → Google Vertex AI follows → Standalone gets absorbed

2/ Developer Frameworks Win (Karpathy, LangChain) 

"Real" AI teams write code → No-code dismissed as toys → LangChain becomes standard → Visual builders get left behind

3/ 235 Competitors, One Category (Tracxn, CB Insights) 

Flowise, Relevance AI, Voiceflow, Zapier AI → Everyone pivots to "agents" → 12 YC startups entered in one batch → Margins collapse

4/ Enterprise AI Stuck at 1% (McKinsey, Forrester) 

"Only 1% reached AI maturity" → "80% report no material earnings" → Pilots die in purgatory → Budgets freeze

5/ Compliance Becomes Commodity (Gartner, Deloitte) 

SOC 2, HIPAA, GDPR now table stakes → Hyperscalers already certified → Startups spend years catching up → Or die trying

6/ Vertical Eats Horizontal (Huang, a16z) 

Generic "AI agent builder" saturates → Vertical specialists win deals → Horizontal becomes infrastructure → Race to zero

Consensus says no-code gets crushed between hyperscalers above and dev frameworks below.

StackAI built the opposite: Defense contracts requiring HIPAA. 30+ LLM providers for vendor lock-in paranoid enterprises. Federal partnership with Diné Development Corporation. One hedge fund cut deal prep 70%. One healthcare client saved 475K hours.

Bernard Aceituno -- you said "risk is the ultimate differentiator" and fired your SMB customers to go enterprise-only.

Consensus says horizontal no-code dies first. What's the thesis they're missing?

Unearned wisdom to say from the outside -- but Amazon modeled logistics, conversion, delivery. They never modeled how Hermès would react when Prime members could "add Birkin to cart."
 
With that many brilliant people in the room, I can see how it happens. Optimization muscle so strong it crowds out adversarial thinking.

undefined undefined

Founder

Mentor

Founding Member

English

Ukrainian

Artur Kiulian

Amazon Can Ship Anything Overnight. Except a Future.

Pavle Lucic

Senior UI/UX Designer @ Toptal | Prototyping, Branding, UI Development

I am a Senior UI/UX Designer at Toptal, a network of the world's top 3% of software engineering, design, and finance talent. With over 10 years of experience in the design industry, I have a proven track record of creating visually engaging and user-friendly interfaces for in-car navigation systems, SaaS apps, and marketing websites.

My background in business and economics gives me a unique perspective on how to design UX and UI that are not only aesthetically pleasing, but also functional, efficient, and user-friendly. I am passionate about solving complex problems, collaborating with diverse teams, and learning new skills. I am also Adwords Search Certified, which demonstrates my ability to optimize the online presence and performance of my clients. My goal is to help tech companies increase their revenue by improving the user experience of their products.

Senior UI/UX Desiger

This is how I work now:

1. Think
2. Write
3. Plan
4. Ask
5. Wait
6. Then design

It’s slower. But it works.

Because UX isn’t just what you make.
It’s what you understand first.

(Repost ♻️)

P.S. What is your process?

Senior UI/UX Product Designer | UX Strategy | Figma

No alternative text description for this image

This might be a controversial post but it's true.

I don't have a formal design degree.

I'm pretty sure a lot of self-taught designers experience this too.

Examples:

Every time I share my work online, someone asks "where did you study?"
I can feel the judgment coming.

Every time I'm in a room with "formally trained" designers.
I hear the whispers: "They probably just watch YouTube tutorials."

Every portfolio review, every job interview, someone implies "you're not a real designer without the credentials."

But then...They see just ONE of my projects and:

That skepticism closes immediately
I get genuine compliments in my DMs
I get referrals + new client inquiries

So...

Why do we judge credentials before we judge the work?
Especially with creative fields? 😊

Fear that talent and hustle might matter more than a piece of paper. 
Which, honestly, they often do.

To my fellow self-taught designers who've "made it work" I love you. 

Keep building. Keep proving that skills speak louder than diplomas.

Keep letting your work change people's minds. It's worth it.

Pavle ❤️

Most designers aren't designers.

(they're technicians who happen to use design tools)

Yesterday I lost 2 hours on a shelf.

Just me obsessing over form and function.
And here's the thing...

That "is" design.

Not the logo you send on Friday.
Not the mockup in your portfolio.
Not the deliverable in your inbox.

Design is how you "think."

- How you can't ignore a bad detail.
- How you see a room and want to fix it.
- How you solve problems no one asked you to solve.

Just think about this:

1. The Bauhaus school understood this.
↳ They didn't teach "tools." They taught living.

2. Japanese carpenters understood this.
↳ No nails. No shortcuts. Every cut is intentional.

Design is form.
Design is function.
Design is creation.

It doesn't "clock in and out."

If you only design when you open your laptop...
And stop when you close it...

You're not a designer.
You're a technician who uses design tools.

Nothing wrong with that.

Just don't confuse the two.

P.S. Agree or disagree? 👇

Loading states matter.

Blank screen = user thinks it's broken.

Progress bar = user knows how long.
Spinner = user knows something's happening.
Skeleton screen = user sees structure forming.
Instant feedback = user doesn't notice the wait.
Steps = user knows where they are ("Step 2 of 4").
Active waiting = user is entertained while waiting (tips, facts).

Never leave users guessing.

Show them the system is working.
Even if it takes time, show progress.

(Repost further ♻️)

17 genuine tips for using illustrations to improve UX

(I wish someone told me number 8 sooner)

1. Empty states feel less scary
2. Error messages feel friendlier
3. Users remember visuals 6x better
4. Onboarding completes 40% more
5. They reduce mental effort instantly
6. They help users know where they are
7. You will add illustrations way too soon
8. They reduce support tickets by 30-40%
9. Permission requests make more sense
10. Complex steps feel simpler with visuals
11. They answer questions before users ask
12. Visual metaphors beat long explanations
13. Users understand faster than reading text
14. They build trust during frustrating moments
15. Empty spaces feel intentional instead of broken
16. They guide without overwhelming with instructions
17. Strategic placement turns confusion into confidence

Repost this free value for others ♻️ Thank you!

P.S. Hey, which one (1-17) do you like most?

If there's one thing I'd eliminate in UX design...

It's birthday pickers like this. You know the ones:

1. Scroll wheels flicking through 40+ years to find 1986
2. Calendar grids tapping "previous month" 480 times
3. Three separate dropdowns, praying the order is right

There's a much better way.

It's called "TIF = Text Input First". 
Rooted in real behavior.
Backed by research. 

TIF is simple:

1. Let users type their birthday
2. Auto-format date as they go
3. Offer a picker as backup only

Why it works:

People know their birthday by heart.
Typing takes 2 seconds.

Scrolling through a picker takes 15+ seconds of frustration.

Hope this helps! (Repost if so ♻️)

P.S. Anything worse than a scroll-wheel date picker?

I was in a meeting last week.

One designer talked for 60 of them.

References to everything. 
"What if we..." on repeat.

Opinions on everything.

Loud. 
Confident. 
Influential.

Everyone nodded along.

Then someone asked:
"So who's going to build this?"

Silence.

The loud one? 
"I'm more of a strategy person."

The quiet one in the corner?

She opened Figma.
Had 3 options ready.
Built while they debated.

That's when I realized:

The best designer in the room?
Often the one who says the least.

Repost this ♻️ if you've seen this happen too.

Good UX is boring.

Users complete tasks without thinking.

They don't notice the design. They just finish.

Bad UX is memorable.
Users remember frustration. Confusion. Friction.

Stop trying to impress users.

The best interfaces are simple.

P.S. When was the last time you noticed good UX?

A client came to me with a "broken" product. 

Here's what happened:

They spent 8 months building an app.

It worked perfectly. But nobody used it.

Refund requests? Daily. 
App store reviews? 2.1 stars.
Support tickets? 200+ per week. 

They thought the product was broken.

It wasn't.
The design was.

I changed 5 things:

1. Moved the main CTA
2. Simplified the onboarding
3. Fixed the mobile navigation
4. Added visual feedback on buttons
5. Reduced the menu from 15 items to 4

4 weeks later?

Retention doubled.
Reviews jumped to 4.4 stars. 
Support tickets dropped 73%. 

Same code. Same product. Same features. 

I didn't "make it pretty."
I made it usable.

That's what designers actually do.

Repost ♻️ if this changed how you see design.

P.S. Have you ever blamed your product when design was the problem? 😊

My first client ever:

"Make the logo bigger."

Quoted Steve Jobs.
I argued for 20 minutes. 
Sent 5 articles about white space. 

They made it bigger themselves in PowerPoint.

My latest client:
"Make the logo bigger."

I said: "Let's test it. I'll send you 2 versions. You pick."
They picked the smaller one.

What changed?

I stopped convincing. 
I started collaborating.

People don't resist ideas. They resist being told what to do.

Give them ownership. 
Watch them choose wisely.

♻️ Repost if you wish you knew this earlier.

Entry-level means no experience.

It's for juniors to begin their careers, learn, and grow.

Yet job posts say:

"Entry-level position"
"Requirements: 3-5 years experience"

That's not entry-level. That's mid-level.

Stop lying in job descriptions.
You're not finding candidates because you're asking for the impossible.

Entry-level = zero experience required.

If you need experience, it's not entry-level.

(Repost for job seekers ♻️)

P.S. Have you seen this contradiction in job posts?

Mini-masterclass on receiving design feedback:

(save this for your next critique)

Well-delivered feedback = great insight
Well-received feedback = great growth

That being said... Let's dive into 5 levels of receiving feedback:

Level 1: Defensive 
↳ "But here's why I did it this way..." 
↳ You're protecting ego, not improving work

Level 2: Silent 
↳ You nod. You take notes. You change nothing. 
↳ Feedback theater. Everyone sees through it.

Level 3: Selective 
↳ You only hear what confirms your choices 
↳ Cherry-picking isn't growth

Level 4: Receptive 
↳ You listen. You ask questions. You implement. 
↳ This is where most good designers live.

Level 5: Hungry 
↳ You seek feedback before anyone offers it 
↳ "What's not working? Where did you struggle?" 
↳ This is where great designers live.

Getting feedback is a skill. Just like giving it.

And most of us never practiced receiving.
We just practiced defending.

Time to switch.

Repost this ♻️ for your design team.

P.S. What level are you at right now (1-5)? No judgment 👇

Old design workflow:

1. Sketch idea
2. Open Figma
3. Design for hours
4. Export and code
5. Test and iterate

New workflow with Claude:

1. Describe idea to Claude
2. Claude codes it fast
3. See design instantly
4. Iterate with prompts
5. Ship faster

I made glass icon effects this way.

No design tools. 
Just conversation.

The bottleneck shifted from execution to clarity of thought.

Can you describe what you want clearly?

That's the new design skill.

(Repost this shift ♻️)

P.S. How has your workflow changed with AI?

I just found out the best color generator for designers.

uicolors. app/generate

Save it for later or repost ♻️

Three reasons Inter dominates:

It's variable

One font file. 
It's optimized.
Multiple widths. 
Multiple weights. 
Smaller load times.

Letter spacing, x-height, character distinction all tuned for screens.

It's free
No legal.
No licensing. 
No procurement. 
Just download and use.

Hard to compete with that combination.

P.S. What font makes your product memorable?

Feedback feels personal.

It's not.

They're critiquing your design. Not you.

Hard to separate:

"This button is unclear" = design feedback
"You made this unclear" = feels like attack
Same words. Different interpretation.

Train yourself to hear:

"The layout is confusing" not "You're confused"
"This needs work" not "You're not good enough"
"Try another approach" not "You failed"

Your design is separate from your worth.

(Repost if this resonates ♻️)

P.S. Does feedback feel personal to you?

Here's the biggest lesson my son taught me:

Yesterday, I cut my finger.

He saw blood. He didn't freeze. He didn't cry.

He ran to get a napkin. 
Wrapped it around my finger. 
Taped it with whatever he found.

Then he grabbed a pen.

And drew a little stick figure on the bandage.
A tiny man.

"That's you, daddy. So you're not alone."

I looked at this ridiculous napkin-bandage. 
Tape peeling off. 
Little man with "hair".

This is what support actually looks like.

Show up. With what you have. 
And make sure they know they're not alone.

That little stick figure?

Best medicine I ever got.

Repost this ♻️ to spread a good message today.

"Wait, you actually use AI for prototyping?"

I get this reaction a lot from other designers.

Yes, I use AI tools daily. 

When did it become normal for designers to hide that they use AI? 
Like it's some dirty secret that makes your work "less valid"?

And...

↳ I prototype 10x faster than I did 2 years ago. 
Yes, full flows in minutes. 
Yes, clients can't tell the difference. 

↳ I refine everything myself. 
I bring the soul. 
AI gives me the skeleton. 
Never shipping raw AI output. 

So to answer this once and for all:

"Is it cheating?"
"Is AI doing your job for you?"
"Is your work even original anymore?"

No. It's leverage.

Try it. 
You'll see.

It's not that scary. 

(Repost if it hits home ♻️)

P.S. What's YOUR resistance to AI prototyping?

taplio

The Battle

Yurii Filipchuk 🇺🇦 -- e/acc

Pavle Lucic

Followers

Posts

Engagements

Likes

Performance ScoreYurii Filipchuk 🇺🇦 -- e/acc's LinkedIn Performance

Performance ScorePavle Lucic's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Yurii Filipchuk 🇺🇦 -- e/acc

Pavle Lucic

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Yurii Filipchuk 🇺🇦 -- e/acc's LinkedIn Performance

Performance ScorePavle Lucic's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreYurii Filipchuk 🇺🇦 -- e/acc's LinkedIn Performance