Full Stack Robotics Engineer | Founder @AIRO3D | 12X hackathon 🏆
Kerala
Joined May 2017
- Tweets4.2K
- Following581
- Followers2.8K
- Likes38.4K
devalopr retweeted
Do you see what I saw? Because I see a seesaw.
Superimposing two images we can find the pivot point using the perpendicular bisector of a chord theorem. The fulcrum is cleary slightly forward of the B pillar, not the rear wheels.
This implies a fulcrum near the center-of-gravity. Given the junkyard setting and the visible damage, this is probably a gutted junker with the engine pulled, which would shift the center of gravity back to where the pivot point is shown.
The car rests on its front wheels with a hidden fulcrum, and the rear wheels are already off the ground. As the robot tips the front up, the rear goes down. Watch closely and you'll see it lift too far, causing the rear wheels to bounce off the ground, as you see the front drops for a moment before the robot recovers.
The robot did not lift a car. It tipped a balanced shell.
The video suggests strength the robot does not possess.
devalopr retweeted
The new @Tesla Roadster with the Thrust Augment puts out more thrust than an F-16 fighter jet at military thrust.
• Roadster: >80 kN (~18,000 lbf)
• F-16 military thrust: ~79 kN (17,800 lbf)
(Military thrust is the maximum thrust a jet engine produces without using the afterburner)
devalopr retweeted
No country can have a "sovereign" Starlink alternative to just serve its own people. Here is why:
LEO isn't geostationary. LEO satellites overfly the whole planet. You need customers all over the world to have a hope of turning a profit. You need 1000s of satellites to have continuous coverage. It'll be prohibitively expensive to only serve 1 country. It'll be far outcompeted by broadband and 5G.
"What if they serve internationally?" You can't compete with the company that owns the rocket with >10X cheaper price/kg over anyone. This moat is only getting stronger as Starship comes online.
Another moat is the 1st mover advantage. LEO real estate is limited. As Starship can launch a lot more satellites and as Starlink mobile starts rolling out, this moat is also strengthening. The subscription cost will drop as the scale increases.
Even Amazon LEO and China will find it extremely difficult to compete with Starlink.
devalopr retweeted
Designing a robot from scratch is just plain hard.
Mechanical engineering, electronics, power, controls, software, AI and you some how have to put it all together in a way that also looks cool.
devalopr retweeted
ive been doing cad for 10 years and have gotten pretty good at it
but if i never have to do cad again i would be so happy
i could put all the effort into the engineering
devalopr retweeted
It's jarring to see how bad even the frontier models are at CAD and 3D design relative to how good they are at writing code
devalopr retweeted
Image to mechanical drawing is almost solved
This has never been done before AFAIK.
Update on my project homework machine:
Handwriting replication for cursive and semi-cursive has been solved. Only print style handwriting was realistic in my previous software.
Existing free handwriting generation softwares like 3dwriter and the inkscape plugins generate unrealistic output that can be easily caught.
The only existing option that can generate realistic handwriting is from quantum enterprises, it costs over $250 per font as it involves expert human labour to replicate your handwriting. My software can automatically record and replicate your handwriting from a sample.
The existing softwares cannot do accurate drawings either. I have almost solved that too.
The machine to do the actual writing is also being developed. It will be fully automated and easy to use.
devalopr retweeted
One 150 g hamburger has a water use footprint of approximately 2,400 L,
equivalent to roughly 83,000 AI images.
One cotton T-shirt has a water use footprint of approximately 2,720 L, equivalent to roughly 94,000 AI images.
One hour of Netflix, including the TV, router and network, is roughly 1.9 L which is the equivalent of 65 AI images.
This includes training + inference. Comparing the same types of water btw so apples to apples.
US golf course irrigation costs 10-18x the water requirements of all data centers in America. Paper mill production water use is 2-3x of all AI data centers in the US.
