Filippo Brunelleschi

Inspired by change, inspiring change. I love life, my wife & 3 kids, the things I'm working on to leave a better future for them, and the people I work with.

[Filippo Brunelleschi @FilBrunelleschi]

Some designs divide opinions.
Some people see beauty.
Some see controversy.
Some see something they would never want to drive.
But that is exactly the point.

In mobility, design will always be subjective. Driver evaluation should not be.

At @Mettis_AI, we are building technology to make driver evaluation objective, data-driven and scalable, using computer vision and AI to understand real driving behaviour.

Not opinions.
Not impressions.
Not “I think you drove well.”
Real data. Real behaviour. Real insights.

The design may be controversial.
The driving data is not.
Mettis AI makes driver evaluation 100% objective.

[Filippo Brunelleschi @FilBrunelleschi]

You’re right on the center of mass point

IMHO smaller wheels > much higher vertical acceleration when hitting obstacles
> worse comfort and stability in real urban conditions

So it feels like these bikes are just optimized for handling when stopped and operational efficiency, rather than actual riding dynamics

Which is… an interesting trade-off

[Filippo Brunelleschi @FilBrunelleschi]

Hey @nhfoley, we’ve both been out of micromobility for a while, but I need your take on these new Lime bikes.

“We kept the geometry the same.”

Is that:
- maturity of the category
- or lack of real innovation?

Feels like we’ve moved from product evolution to pure ops optimization.
Nevertheless a signal that the category has plateaued on hardware innovation.

Agree or am I missing something?

[Filippo Brunelleschi @FilBrunelleschi]

What looks like “driving” is actually hundreds of micro-decisions per minute.

Most of them are never measured.

What we’re building at @Mettis_AI is the ability to capture, structure, and understand those decisions in real time, not just what the driver does, but why.

Because driving is not just rules. It’s intent.

Every hesitation, every adjustment, every interaction with other road users carries meaning.

For now, this allows us to make evaluation objective.

But the real opportunity is bigger:

→ understanding human behavior at scale
→ building datasets of real-world driving intent
→ enabling systems that can anticipate what humans will do next

That’s where this is going. 🚀

[Mettis_AI @Mettis_AI]

Real-time driving analysis.

Understanding human intent, frame by frame.

Next step: teaching machines to predict it.

[Filippo Brunelleschi @FilBrunelleschi]

@SawyerMerritt @Tesla I wonder if in Europe it'd still use traditional mirrors or just cameras like most semi in the region

[Filippo Brunelleschi @FilBrunelleschi]

@AlertaMundoNews @grok cuanto tardaron los romanos a construir el Coliseo?

[Filippo Brunelleschi @FilBrunelleschi]

The interesting part isn’t where the Cybercab charging port is, it’s how little importance it gets.

That strongly suggests automated depot charging (wireless or robotic) as the primary mode.

The plug feels like redundancy, compliance, and edge-case recovery, not the main UX.

[Filippo Brunelleschi @FilBrunelleschi]

@whatsinside @JoshWest247 This makes sense if Cybercab is designed primarily for automated depot charging (wireless or robotic).
The physical plug likely exists as a fallback / service option, not for daily use.

[Filippo Brunelleschi @FilBrunelleschi]

Hey @elonmusk, will consumer Model Y soon get the side and rear camera washers like in the Austin robotaxi fleet? I guess it would be great for FSD reliability

[Filippo Brunelleschi @FilBrunelleschi]

Not sure why I keep posting on LinkedIn when the valuable audience is here on X.

Resharing Augustin Friedel’s insightful post on global ride-hailing platforms and their AV strategies.

That said, I’m curious: why leave out Tesla in this analysis?

As a vertically integrated player launching Robotaxi services in 2025/2026, Tesla is directly competing in autonomous mobility with a disruptive edge.

Their Austin geofence already spans ~245 sq mi (larger than Waymo’s ~90 sq mi there), and with Cybercab production ramping, they’re targeting fleets at a fraction of the cost, under $30K per vehicle vs. Waymo’s estimated $100K+ setups.

Autonomous vehicles and ride-hailing won’t scale if the underlying tech isn’t scalable. Tesla’s vision-only approach, relying on advanced computer vision and neural networks trained on massive real-world data, allows for rapid iteration and deployment across diverse environments without the hardware dependencies of LiDAR-based systems like Waymo’s, which require expensive sensors, meticulous mapping, and slower adaptations to new areas.

For the same capital, Tesla could deploy 3-4x more vehicles.

This isn’t just OEM vs. tech, it’s a paradigm shift that could redefine the ecosystem.
Thoughts?
Would love to hear from others in mobility/AV space.

#AutonomousVehicles #Mobility #Tesla #Waymo #RideHailing

Augustin’s original post: https://t.co/LMcpBnuE4L

Real time robotaxi active fleet
https://t.co/SWXeou4qZr

[Filippo Brunelleschi @FilBrunelleschi]

@gothburz Agility is where Consultants end up when they “evolve”

[Filippo Brunelleschi @FilBrunelleschi]

hey @UberSpain I guess #CandyCrushSaga is addictive, not sure how “safe” though

[Filippo Brunelleschi @FilBrunelleschi]

@elonmusk You should have called it Skynet

[Filippo Brunelleschi @FilBrunelleschi]

@SawyerMerritt the 1200-volt platform is interesting. The driver, Marc Basseng, "backs out of it" at 472kph since "he perceived stability issues" that made him hesitant to continue at that pace

[Filippo Brunelleschi @FilBrunelleschi]

@prabinjoel I’ve seen Jaecoo in Spain for over a year… not only that, there are also plenty of MG, Omoda, Voyah, Xpeng, BYD and Livan…

[Filippo Brunelleschi @FilBrunelleschi]

@grok probably "just chill duuuude"

[Filippo Brunelleschi @FilBrunelleschi]

Not sure about you, but my @grok inner spirit is fully represented by Gork

[Filippo Brunelleschi @FilBrunelleschi]

@grok I mean the voice assistant Gork, he's ma' dude

[Filippo Brunelleschi @FilBrunelleschi]

@FxbMeta @TELOtrucks I’m sure they could do a cheaper 2WD, ~50 kWh, lower-power spec around $30k. But AFAIK most orders today are for the long-range, high-spec model. A $30k trim would be more of a headline-grabber than something buyers actually choose

[Filippo Brunelleschi @FilBrunelleschi]

I’d not only reserve a truck, I’d invest in @TELOtrucks. They seem like a team that knows exactly what they’re doing. Cool people, great vision, and a product that’s finally something fresh.

[Filippo Brunelleschi @FilBrunelleschi]

@FxbMeta @TELOtrucks It’s built to do a lot in a small package, power, range, fast charging qand smart design. For many, that’s worth it