bybass

build? development? or chill

[bybass @BybassBuild]

https://t.co/7fhp8h0MY9

[bybass @BybassBuild]

@realDonaldTrump just another video frow the president 🤩

[bybass @BybassBuild]

You can build and publish a multiplayer game in 2026 without paying for an engine, servers, hosting, or a backend.
And AI can write most of the boring code.
The barrier to making games is getting stupidly low.
The barrier to making a game people actually want to play is still exactly where it was.
I went deep into the Roblox economy, how games get built, what they cost, and where the money actually comes from:

[bybass @BybassBuild]

https://t.co/7fhp8h0MY9

[bybass @BybassBuild]

@birdabo Modelers reading “2 hours” and quietly closing Blender for the day.

[bybass @BybassBuild]

@RoVoider Modelers reading “2 hours” and quietly closing Blender for the day.

[bybass @BybassBuild]

A 16-year-old built a Roblox game that eventually hit 22.3 MILLION concurrent players.

So I went down the rabbit hole.

How much does it actually cost to build a Roblox game in 2026?

How much can AI do?

Where does the money come from?

And can a solo beginner realistically make anything?

[bybass @BybassBuild]

https://t.co/7fhp8h0MY9

[bybass @BybassBuild]

What’s wild is how fast the bottleneck moved.

A few years ago, having the idea was the easy part. Then came weeks of coding, debugging, UI work, assets, deployment, and all the other little things that quietly kill side projects.

Now you can describe the game, test it almost immediately, and spend your time deciding what’s actually fun.

That’s a much bigger shift than “AI can code.”

[bybass @BybassBuild]

CHINA IS LITERALLY BEATING THE SHIT OUT OF ROBOTS TO MAKE THEM BETTER.

Watch the one in the middle.

Two humanoids keep hitting it from opposite sides while it tries to stay on its feet.

It looks ridiculous, but this is exactly the kind of problem humanoid robots need to solve.

The real world won’t be a clean laboratory floor.

People bump into you. Things hit you. The ground moves. You lose balance.

A useful robot has to recover without a human running over to save it every time something goes wrong.

So while everyone is obsessed with robots doing backflips, this might actually matter more.

China isn’t just teaching humanoids how to move.

They’re teaching them how not to fall.

[bybass @BybassBuild]

THE GIRL YOU’RE LOOKING AT ISN’T HUMAN.
SHE’S A ROBOT BUILT IN CHINA.
Watch her face before you look at the neck.
The eyes move. The lips react. Tiny expressions appear while she’s standing right next to an actual human.
Then you see the exposed machinery underneath.
This is AheadForm’s Origin F1, built to make face-to-face interaction with machines feel less like talking to a machine.
And that’s where this gets interesting.
Most robots are obvious from across the room.
Metal face. Plastic shell. Mechanical movement.
You never have to wonder what you’re looking at.
But a receptionist doesn’t need to lift 100 pounds.
A hotel worker doesn’t need to do a backflip.
A customer service robot needs to look at you, listen and respond without making the interaction feel weird.
That makes the face incredibly important.
China already has companies racing to make humanoids walk, run and work.
Now some are working on the part we notice first.
The face.
Watch the clip one more time.
If you couldn’t see her neck, how quickly would you know?

[bybass @BybassBuild]

Watch the worker leave.

The robot doesn’t.

That’s the part of humanoid robotics that gets lost in all the backflip demos.

A supermarket doesn’t need Optimus to be brilliant. It needs it to open boxes, stock shelves, move carts and do the same boring jobs at 3AM that somebody has to do every single night.

Humans are actually very good at those jobs.

We’re also expensive to keep there.

We need shifts. Breaks. Training. Someone has to call another employee when we don’t show up.

A robot has its own problems: it costs a lot upfront, breaks, needs charging and still can’t handle nearly as many weird situations as a person can.

But watch where the competition is happening.

Every year the human costs roughly the same.

Every year the robot is supposed to get cheaper and better.

At some point those two lines meet.

And the first jobs affected probably won’t be the complicated ones.

They’ll be the jobs happening while the rest of us are asleep.

[bybass @BybassBuild]

THE SCARIEST ROBOT DEMO ISN’T ONE WHERE IT WALKS, RUNS OR DOES A BACKFLIP.

IT’S THIS ONE.

Watch its hands.

It holds the weapon, adjusts its grip, aims and fires without losing control of the gun.

That sounds like a small detail until you remember how much of humanoid robotics has been about teaching machines to handle objects made for humans.

Doors. Tools. Boxes. Cars.

Weapons are made for human hands too.

And that creates a problem we haven’t really had before.

A drone is built as a weapon.

A humanoid can theoretically walk through the same doors we do, climb the same stairs, use the same equipment and operate in places designed entirely around the human body.

You don’t need to redesign the environment for the machine.

You redesign the machine to fit ours.

That’s what makes footage like this uncomfortable to watch.

We spent years trying to make robots move more like humans.

Eventually someone was going to ask what else they could hold.

[bybass @BybassBuild]

HUMANOID ROBOTS ARE TAKING OUR JOBS.

Eventually.

This one still has some work to do.

And honestly, clips like this are why robotics is so fun to watch right now. One day they’re doing backflips, working in factories and handling tools.

The next day, gravity wins.

We’re watching these machines learn in public.

[bybass @BybassBuild]

THIS ROBOT JUST DID SOMETHING MOST HUMANOIDS STILL CAN’T.

Watch the recovery.

Falling is easy. Getting back up without a human stepping in is the part that actually matters.

Because a useful robot won’t spend its life walking across a perfectly flat lab floor.

It’ll slip. Miss a step. Drop things. Get pushed. Land in positions nobody planned for.

And every time that happens, calling a human over to rescue a $50,000 machine kills the whole point of automation.

That’s why videos like this matter more than another robot walking in a straight line.

A robot doesn’t need to be perfect.

It needs to fail without becoming someone else’s problem.

That’s a much harder benchmark.

[bybass @BybassBuild]

ONE DAY, THIS WON’T BE AN AI-GENERATED VIDEO.

A robot walks into an interview, signs a contract and gets the job.

Right now that looks ridiculous.

Give it a few years.

Companies are already spending billions trying to make humanoid robots useful outside of carefully staged demos. Tesla is building Optimus. Figure is putting robots into factories. Chinese companies are pushing humanoids into production at increasingly lower prices.

And none of them need to build a robot that can do everything you can.

That would be overkill.

A warehouse doesn’t need a robot that can write poetry. A hotel doesn’t need one that understands philosophy. A factory doesn’t care if its robot can hold a conversation.

It needs something that can do a boring physical job for hours, reliably, at a price that makes sense.

That’s where things get interesting.

Hiring a person means salary, training, breaks, sick days, turnover and finding someone willing to show up again tomorrow.

Buying a robot means a large upfront bill, electricity, maintenance and repairs.

For now, humans still win that calculation in a huge number of jobs.

But robot prices are falling, production is growing and the machines are getting more capable.

At some point, for some job, those two numbers cross.

And whoever employs that worker will have a very easy decision to make.

The robot in this video is fake.

The job interview eventually won’t be.

Which job gets there first?

[bybass @BybassBuild]

@0xSyrex Imagine spending years mastering culinary arts just to get replaced by a metal arm that never asks for a smoke break.

[bybass @BybassBuild]

@CytMutant The real advantage isn’t replacing one worker. It’s deploying the same improvements across thousands of robots without retraining a single person.

[bybass @BybassBuild]

@Dapper_0ne Bro passed the Turing test and immediately started dating the hardware.

[bybass @BybassBuild]

@0xTenzo Bro is fighting for his life against a guy named Dave with a shovel 😭

[bybass @BybassBuild]

@drakefomo The biggest Optimus customer might not be a person or a company. It might be an entire country.

[bybass @BybassBuild]

@arc_courier Optimus might lose the demo today and still win the entire race.