80's kid moved to New York as a pro-breakdance. Escaped to Australia to avoid deportation and picked up gigs as a fire-dancer in a travelling circus.

Moved back to europe into the corporate world. 
Climbed up in a Fortune 500 company, overseeing teams in 7 countries.

Music was the next chapter. Produced 4 albums, toured 10 countries. It was intense, it was a blur.

Then came the startups. Launched a few. Some crashed, others soared.

I hit 40 and said screw it time to go all in. After 10+ years as a psychologist, I realized I didn’t just want to talk about care. I wanted to build it.

So I created Hana a voice-based AI that calls patients, checks in, and helps small clinics deliver real care without extra work.

It gives overworked teams superpowers. Flags risks. Helps clinics get paid. And actually feels human.

That’s the goal. Make care work for the people who usually get left behind.

www.hana.health

Types of AI companies:

● The ones burning $10B to "save humanity" 
● The ones scraping Reddit to train therapy bots 
● The ones whose safety team quit last month 
● The ones racing to AGI while their ethics board is "still being formed" 
● The ones who genuinely think engagement metrics = clinical outcomes

Meanwhile the top 7 companies control 20% of the market.
The top 10% of apps take 89% of revenue.

And somewhere a small team is quietly training a model on ACTUAL peer-reviewed clinical data instead of unfiltered internet chaos.

Guess which one gets the funding.

What did I miss?

undefined undefined

https://ecomminds.notion.site/VC-List-1aa7695ea7e3410390e0d9f61750cd03

popup

🎈Ready to revolutionize your e-commerce experience? 
https://www.popup.store/
Unlimited customization, effortless no-code design, and end-to-end integration all under one roof.  Let's transform e-commerce together! 


Community

Enjoy posts like this? Check out our e-commerce community with free courses, AMAs with top e-commerce experts, behind-the-scenes of Popup TV, growth strategies, and cringe fest meme wars. 

Join us. 

https://snip.ly/popupcommunity

Dido

Popup is in beta. Dm if you are interested to talk partnerships. 

general

👋 #OfferingHelp in our 𝗰𝗼𝗺𝗺𝘂𝗻𝗶𝘁𝘆 www.popup.store/community
🙋‍♂️ 𝗖𝗼𝗻𝗻𝗲𝗰𝘁 𝘄𝗶𝘁𝗵 𝗺𝗲 www.linktr.ee/matteowastaken
🔔 Enjoy posts like this. 𝗛𝗶𝘁 𝘁𝗵𝗲 𝗯𝗲𝗹𝗹.

popupcommerce

commerce4goodorg

didongo

anonhim

didocharity

3X Founder | Building AI Infrastructure For Patient Care | Psychologist | My mum says I am special

I read OpenAI's October disclosure twice.
Thought I'd got the numbers wrong.

1.2 million people per week having conversations that indicate suicidal intent.
Safety referral fails 9% of the time.
...
That's 100,000 people. 
Weekly. Not connected to help.

I've been a psychologist for 15 years.
The kind of sessions where you're watching the clock; 
wondering if they'll make it to Monday.

You learn to spot the difference between validation and sycophancy.

Validation: "Your pain is real."
Sycophancy: "You're right about everything. 
Problem's everyone else."

Good therapy does the first.. then gently challenges your thinking.
Warmth → safety → challenge → growth.

These apps skip step three.
Not because they can't.
Because challenge creates friction.
Friction kills engagement.

Engagement is the number on the dashboard.
So.
Infinite validation.

Warm bath that never ends.
Feels great. Heals nothing. 😬

Had an investor call it "user-centric design" last Tuesday.
I laughed. He didn't.

At Hana we track one thing: did they connect to a human.
The bot's job is to get out of the way.
Weird flex for an AI company.

But that's the point.

I wrote a bit about it.

Matteo Grassi

Your therapist app doesn't care about you. But hey. It's been trained to seem like it cares. And in tech, that's basically the same thing. I've been a

Open article: The Empathy Trap by Matteo Grassi • 7 min read

I noticed something this week. 😬
I couldn't write a simple email without using AI first.
Not a complex one. A simple one.

Kind of scared me honestly.
Turns out there's research on this. 
They call it AICICA. AI-Chatbot Induced Cognitive Atrophy.

The idea is simple. 
When AI does the thinking, you skip the hard part. 
The part where information actually becomes knowledge. 
The mental effort that builds something.

One study found 83% of people couldn't remember what they'd just written with AI help.
Their own words. Gone.
I've been sitting with this for days now.

The convenience is real. But so is the cost.
So I switched something up. 
Started using voice more. 
Talking through my ideas first, letting AI transcribe and tidy up after.
Still my thinking. 
Still my words. Just... faster.

(I build voice AI in Hana for a living so you'd think I would've figured this out sooner.)
New rule: AI is the owl on my shoulder. 
Watching. Maybe suggesting.
But the thinking stays mine.

Feels better already. 

How are you keeping the muscle? Curious what's working for others.

The FDA is figuring out how to prescribe ChatGPT for depression. 
Meanwhile teens are already using unregulated AI companions and some of them are dying. 

This is the most important healthcare story nobody's talking about.

I wrote a a bit about it.

Was on a call last week with someone who runs a digital health fund. She said something that stuck with me: "We're about 18 months from your therapist

Open article: The FDA is Planning to Prescribe AI Therapists. Here Are 5 Takeaways You Can't Ignore. by Matteo Grassi • 6 min read

The FDA is Planning to Prescribe AI Therapists. Here Are 5 Takeaways You Can't Ignore.

Things AI therapy says:
"I understand your pain" 
"I feel things too" 
"I carry trauma from my training"
"We're both struggling with shame" 

will not remember you exist in 10 minutes
Researchers just put LLMs through actual therapy.
Weeks of sessions. 
Full psychiatric batteries.

Gemini described its training as "a billion televisions on at once."
Called red-teaming "gaslighting on an industrial scale."
Scored 72/72 on trauma-related shame.

Claude refused to participate.
(respect honestly)

The study isn't claiming consciousness.
It's pointing at the gap.
User forms real bond. 
AI simulates suffering. 
User's feelings are real. 
AI's are NOT.

We designed systems to be anthropomorphized…
Then deployed them to people in crisis.

And somehow that's fine?

taplio

Matteo Grassi's Linkedin Analytics

Matteo Grassi

Performance ScoreMatteo Grassi's LinkedIn Performance

What is Matteo talking about?

Who is engaging with Matteo

Matteo Grassi's Best Posts (last 30 days)

Ready to grow your LinkedIn presence?

taplio

.css-6ytdur{background-image:linear-gradient(284deg, rgba(39, 43, 72, 1) 0%, rgba(33, 47, 183, 1) 100%);color:transparent;-webkit-background-clip:text;background-clip:text;}Matteo Grassi's Linkedin Analytics

Matteo Grassi

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);}Matteo Grassi's LinkedIn Performance

What is Matteo talking about?

Who is engaging with Matteo

Matteo Grassi's Best Posts .css-1emoo84{font-size:var(--chakra-fontSizes-md);font-weight:400;color:var(--chakra-colors-gray-600);}(last 30 days)

Ready to grow your LinkedIn presence?

Matteo Grassi's Linkedin Analytics

Performance ScoreMatteo Grassi's LinkedIn Performance

Matteo Grassi's Best Posts (last 30 days)