CoFounder of CYT. Robots. AI. The future is here.
Joined May 2022
- Tweets415
- Following1.1K
- Followers1.4K
- Likes1.6K
Mutant CYT retweeted
$94,000 A YEAR to babysit ROBOTS.
No, this isn't a joke. Figure AI is actually hiring someone for this.
And the job requirements are just beautiful:
Work from 10:30 PM to 7 AM
Stand for 8+ hours
Lift equipment weighing up to 50 lbs
Troubleshoot robot failures
Operate the robots YOURSELF if needed
That last one is my favorite.
If a robot can't do its job, YOU have to do it for them.
And once it learns how, it can take YOUR job.
But don't worry.
Maybe it'll hire you back someday. For the night shift.
Last chance to buy $NVDA under $185.
Bookmark this tweet.
NVIDIA’s CEO didn’t cheat on his wife.
China thought they were recruiting him. In reality, Trump sent him to burn their entire spy network.
“Agent H reporting: target infiltrated. I repeat, target infiltrated.” 🕶️
5
340
NVIDIA’s CEO didn’t cheat on his wife.
China thought they were recruiting him. In reality, Trump sent him to burn their entire spy network.
“Agent H reporting: target infiltrated. I repeat, target infiltrated.” 🕶️
Mutant CYT retweeted
NVIDIA may have found a vulnerability no driver update can fix.
Jensen Huang admitted to an affair. Now NVIDIA is investigating reports that the woman may have ties to Chinese intelligence.
All those billions spent protecting chip technology, and nobody thought to check the CEO's bedroom.
Manufacturing changed more in 250 years than most people realize.
> Hand tools and craftsmen
> Steam-powered machines
> Factories
> Assembly lines
> Industrial robots
> Fully automated production
> AI + humanoid workers
We went from making one thing by hand to factories that barely need humans at all.
And we’re still not at the end.
6
193
Mutant CYT retweeted
OpenAI: here's your GIF. Now count your tokens.
Yesterday they rolled out Images 2.6. Now you can animate pictures, make GIFs, and experiment with animation. Have fun, dear subscribers! Just not too much.
Because starting in December, Plus, Pro, and Business users will have a SHARED limit for ChatGPT and Codex. OpenAI has kindly included a meter with our new toys. God forbid we get too carried away with being creative.
Burned through your limit on a big project in Work? Great job! Now advanced models are unavailable to you in regular chat too. Wait for your limit to reset or get your wallet out.
They promise not to touch Instant. How generous! At least we don't have to pay extra to ask "how long do I boil eggs" yet.
But what annoys me most is the psychological part. Now, before every stupid idea, you'll start thinking: is it really worth my precious tokens?
Perfect atmosphere for creativity. You're paying for an AI subscription, sitting in front of the most powerful model, and saving questions like you're saving mobile data while roaming.
And they haven't even told us exactly how much we'll be allowed to use. A surprise, apparently.
And why did OpenAI decide to copy from Claude the exact thing I dislike about it?
Ethos isn’t a magic solution either.
A lot of accounts there have inflated reputations built through reputation triangles and coordinated reviews. You can game trust scores just like you can game Reddit.
The one thing Ethos does well is CANCEL projects that screw up badly. That part is actually useful.
Rich people are hiring GEN Z to make ChatGPT think better of them.
Yep, we've reached that point.
Business Insider uncovered a whole industry built around fixing people's reputations in AI. One agency has 24 Gen Z employees who literally get paid to hang out on Reddit and talk about wealthy clients and their businesses.
One rich guy got into some trouble involving an expensive car, and people on Reddit started calling him a douchebag.
So what do you do about that reputation? Obviously, you hire people to go online and tell everyone what a wonderful person you actually are.
And with a little luck, ChatGPT picks up that version of the story.
People are paying tens of thousands of dollars for this.
Come on, do we seriously need to make chatbots like us now? Some people are even hiring entire teams just for that.
And we're the ones asking ChatGPT whether someone is trustworthy, whether we should do business with them, or what kind of person they are.
Just imagine asking ChatGPT about some businessman while 24 Gen Z kids have been spending weeks on Reddit carefully polishing his reputation.
And you might never even know where ChatGPT got such a glowing opinion of him.
Meanwhile, we already have platforms like Ethos where you can check reviews of people and projects, see who left them, and figure out who you can actually trust. My friend @Biti8888 writes about these kinds of platforms.
But ChatGPT just hands you a nice, polished bio, and good luck figuring out how much of it is real and how much is paid PR.
I knew you could buy yourself a good reputation online, but now we're teaching machines to believe that crap too.
Makes me wonder how many of these wonderful people ChatGPT has already recommended to us.
Mutant CYT retweeted
I got caught at the US border at 19 and sent back home. Ended up so deep in debt I had to skip meals. Somehow I died at 96 as a Canadian citizen.
Still playing 1 in 8 Billion, and this life was a wild ride.
> Spawned into a poor family in rural Philippines 🇵🇭
> Tried moving to the US at 19. Got sent back and lost thousands
> Tried again at 28. Made it to Canada, eventually became a citizen 🇨🇦
> Wanted to become a boxer. Ended up working as a master electrician
> Got divorced at 63. Lost $101K but kept my house and eventually paid off the mortgage
> Beat cancer at 80. Died at 96 with $160K, almost entirely in home equity
The craziest part? Only 0.004% of boys born in the Philippines that year would live longer than me.
I never became a boxer, lost a fortune in a divorce, and spent years trying to get my life together. But somehow, I got 96 years out of this run.
Mutant CYT retweeted
Microduck decided Jensen Huang isn't the boss of him
Yesterday at GTC Berlin, NVIDIA showed how robots can be trained in simulation and then use those skills in the real world.
They brought out four Microducks for the demo. Jensen tried to get them to follow his commands, but not all of them managed to do a backflip, and one duck just wandered off. Jensen had to personally convince it to come back 😂
Microduck costs $399, runs on open-source software, and owners will be able to teach it new moves themselves.
I can already imagine what's gonna happen when those preorders finally start arriving.
NVIDIA put all that effort into teaching ducks to do backflips. The internet's first move will be teaching them to run away from their owners.
Although judging by Berlin, they've already figured that part out.
NO WAY. Looks like the conspiracy guys were RIGHT.
REPTILIANS TOOK OVER THE WHITE HOUSE.
After watching this, that old conspiracy suddenly doesn’t feel so easy to dismiss. Spielberg’s Disclosure Day came out this year, where a TV meteorologist is suddenly taken over by an extraterrestrial force and starts speaking a non-human language live on air. Now watch this video again.
Maybe fiction is getting a little too close to reality.
Mutant CYT retweeted
Thought X cut your reach?
You weren’t imagining it.
On October 7, the open xai-org/x-algorithm repo enabled a new source for For You: popular posts.
It pulls posts less than 24 hours old, up to 5 per author, with up to 500 candidates total.
But there’s a catch.
The system starts with the biggest active accounts by follower count.
Yeah, yeah. You got it. Size comes with a very real advantage.
So if a big account is already taking off, its post now has another route into other people’s For You.
At the same time, PostUnexploredWeight was cut from 0.02 to 0.015. That boost only applies when the viewer already follows you, so this specifically weakens the extra push to your existing audience.
Small accounts, well... you get it. I’m with you.
2050. MARS.
Tesla Optimus.
Cybertruck.
Living modules.
No humans outside without a suit.
Elon Musk actually did it.
The first cities on Mars won’t be built by people. They’ll be built by machines that don’t need oxygen, sleep, food or protection from the cold.
Optimus will work outside. Cybertrucks will move cargo between bases. Humans will stay inside while robots build the next piece of civilization.
This looks like science fiction today.
But 2050 is only 24 years away.