Engineering
/engineering3347
Electrical, Aerospace, Mechanical, Systems, Chemical, Semiconductor etc etc. A channel for making things work
Send help guys 💔
935226
Fega
@oghenefega·2 days ago
How I am feeling today doing my class assignments 💔
what's the best free/free-trial/non-student autodesk option for editing 3d step files for printer/cnc?
i'd pay to watch a self-driving racecar and i bet people would even hope it spun some other driver out
post vibe-coding, engineering really means creating embeddable knowledge
Framework really does seem worth watching
1/ Verge leads with James Bond in the ep title but the first half is a convo w Framework CEO
"First, Framework CEO Nirav Patel joins David and The Verge's Sean Hollister to talk about Framework's new Laptop 12 and Desktop, plus the
company's plan to bring its upgradeable, repairable ethos to other gadgets."
https://pod.link/vergecast/episode/3fed6174a497d55d4d670b5df5a87bc3
1/ Verge leads with James Bond in the ep title but the first half is a convo w Framework CEO
"First, Framework CEO Nirav Patel joins David and The Verge's Sean Hollister to talk about Framework's new Laptop 12 and Desktop, plus the
company's plan to bring its upgradeable, repairable ethos to other gadgets."
https://pod.link/vergecast/episode/3fed6174a497d55d4d670b5df5a87bc3
Great article. Part of me wants to build a jet engine. Why?
- great situation where the cost of development is astronomically bloated (3~5B for new engines)
- You can bring it down orders of magnitude (100x reduction would be 3-5M, even less) with modern 2020s tech, perhaps (unchanged since 1980s) and aerospace co stuck in their ways, chasing contracts etc
- and possible if you do it the right way.
- Yes it’s naive, but this is how amazing things start
Here’s how I’d do it:
- 3D printed CMC blades
- no multi stage axial, go simple single centrifugal compressor
- basic cooling
- no FADEC, just build simple like a ESCs/ECUs
- no certification - find some exotic use case (experimental, defense, space, no FAA)
Most of the bloat is in
- R&D
- requirements
- certification
Cut all of it. Do it in a shorter period of time (10 years -> 2~3yrs)
https://www.construction-physics.com/p/why-its-so-hard-to-build-a-jet-engine
- great situation where the cost of development is astronomically bloated (3~5B for new engines)
- You can bring it down orders of magnitude (100x reduction would be 3-5M, even less) with modern 2020s tech, perhaps (unchanged since 1980s) and aerospace co stuck in their ways, chasing contracts etc
- and possible if you do it the right way.
- Yes it’s naive, but this is how amazing things start
Here’s how I’d do it:
- 3D printed CMC blades
- no multi stage axial, go simple single centrifugal compressor
- basic cooling
- no FADEC, just build simple like a ESCs/ECUs
- no certification - find some exotic use case (experimental, defense, space, no FAA)
Most of the bloat is in
- R&D
- requirements
- certification
Cut all of it. Do it in a shorter period of time (10 years -> 2~3yrs)
https://www.construction-physics.com/p/why-its-so-hard-to-build-a-jet-engine
Polymers are large molecules composed of repeating structural units known as monomers. These macromolecules are widely used in various industries due to their lightweight, durability, and versatility. Polymers can be classified into natural polymers (such as cellulose, rubber, and proteins) and synthetic polymers (such as plastics, nylon, and polyethylene).
Types of Polymers
1. Thermoplastics – Can be melted and reshaped multiple times (e.g., PVC, polyethylene).
2. Thermosetting Polymers – Harden permanently after being molded (e.g., epoxy, Bakelite).
3. Elastomers – Have rubber-like properties and high elasticity (e.g., silicone, rubber).
4. Biodegradable Polymers – Environment-friendly alternatives that degrade naturally (e.g., polylactic acid).
Medical: Biodegradable sutures, drug delivery systems
Automotive: Lightweight components, tires
Electronics: Insulating materials, flexible circuits
Packaging: Plastic bags, food containers
Types of Polymers
1. Thermoplastics – Can be melted and reshaped multiple times (e.g., PVC, polyethylene).
2. Thermosetting Polymers – Harden permanently after being molded (e.g., epoxy, Bakelite).
3. Elastomers – Have rubber-like properties and high elasticity (e.g., silicone, rubber).
4. Biodegradable Polymers – Environment-friendly alternatives that degrade naturally (e.g., polylactic acid).
Medical: Biodegradable sutures, drug delivery systems
Automotive: Lightweight components, tires
Electronics: Insulating materials, flexible circuits
Packaging: Plastic bags, food containers
Tempting: Desktop Wind Tunnel for Your Diecast Cars Models (and who knows what else?)
(props to the insta algo for throwing a dart)
https://www.fun-tech-lab.com/products/windsible-desktop-wind-tunnel-for-your-diecast-cars-models
(props to the insta algo for throwing a dart)
https://www.fun-tech-lab.com/products/windsible-desktop-wind-tunnel-for-your-diecast-cars-models
Sometimes you come across products that just make you love the product
Sensepeek is one those
https://sensepeek.com/
Another one for me is Saelae
https://www.saleae.com/
Sensepeek is one those
https://sensepeek.com/
Another one for me is Saelae
https://www.saleae.com/
Sensepeek
sensepeek.com
Saleae Logic Analyzers
Effortlessly capture signals and decode protocols like SPI, I2C, I3C, CAN bus, Serial, and many more with the world's most trusted USB logic analyzer. Delightfully intuitive software and a rich community of user-created protocol decoders allow you to capture, analyze, and debug both analog and digital signals.
www.saleae.com
Increasing scale changes everything about the system. Design that targets two scales breaks both ways
If you do so, you will never reach that point because the time it takes to build something grounded on theoretical constraints will make your work irrelevant
If you do so, you will never reach that point because the time it takes to build something grounded on theoretical constraints will make your work irrelevant
what happens when we stay high level?
1287
July
@july·17:53 06/02/2025
All 3 printing at once
Let’s see how it goes
If it goes well this is going to be my life for the next few weeks
Let’s see how it goes
If it goes well this is going to be my life for the next few weeks
I'm reminded of Moore's law as well (when I think about scaling inference time compute) I think in the 2000s? 10s? there was a shift from scaling the CPU / power efficiency i.e. Dennard scaling (because we started hitting the physical limits like 3~4 Ghz - and started moving to multi-core scaling, SIMD (single instruction, multiple data). Also just better branch prediciton, out of order execution
I think what's also interesting was this was also around the time that NVIDIA started to pick up CUDA in 07, GPGPUs, Google started building TPUs, specialized hardware for doing massively parallel tasks, which eventually found a gigantic home in AI
I think what's also interesting was this was also around the time that NVIDIA started to pick up CUDA in 07, GPGPUs, Google started building TPUs, specialized hardware for doing massively parallel tasks, which eventually found a gigantic home in AI
Gah. Cruise was such a fumble
GM you messed up
https://www.axios.com/2025/02/04/cruise-gm-robotaxi-layoffs
GM you messed up
https://www.axios.com/2025/02/04/cruise-gm-robotaxi-layoffs
images-as-coordinated-truths mostly work where "form determines function" ... where the requirements of camera calibration are loose
for truths of something formless like gas composition, we'd want verifiably-calibrated labs that does things like chromatography ... such that "scale determines function"
verifiable labs-on-chips?
for truths of something formless like gas composition, we'd want verifiably-calibrated labs that does things like chromatography ... such that "scale determines function"
verifiable labs-on-chips?
1287
July
@july·04:54 22/01/2025
And "truth" in this case, I imagine a sort of "Popperian Truth" -- where 'truths' can only be tested and potentially falsified. Truth, in this sense, is provisional and contingent on surviving attempts at falsification.
And as I think of a verified truth layer, as an for an objective adjudicator to exist. Provide a shared reality layer where individual claims can be verified or falsified -- meaning you don't proclaim that there is an ultimate arbiter of abstract "Truth" with a capital T. Rather, it’s like a ref that referees sports, or something or rather
And as I think of a verified truth layer, as an for an objective adjudicator to exist. Provide a shared reality layer where individual claims can be verified or falsified -- meaning you don't proclaim that there is an ultimate arbiter of abstract "Truth" with a capital T. Rather, it’s like a ref that referees sports, or something or rather
Those pincers are way too small, but this is a brilliant idea.
Just another example of why specific-task robots are best. Humanoid robots would cost 1000x more to accomplish the same task
https://x.com/BrianRoemmele/status/1876122027826024828
Just another example of why specific-task robots are best. Humanoid robots would cost 1000x more to accomplish the same task
https://x.com/BrianRoemmele/status/1876122027826024828
Interesting article about a low cost and non destructive IC verification method! (found as a link in vitalik's recent d/acc blog post) Interesting problem space to work in for open source hardware
https://www.bunniestudios.com/blog/2023/infra-red-in-situ-iris-inspection-of-silicon/
https://www.bunniestudios.com/blog/2023/infra-red-in-situ-iris-inspection-of-silicon/
There’s a podcast about materials science called “Materialism” and the episodes usually have a lot of interesting tidbits. They sometimes use technical terms without explaining them, so some prior exposure to MS (e.g. through engineering) helps but I think most of the content is very approachable and interesting nonetheless.
Here are two episodes I liked:
- Stories from a materials salesman: https://open.spotify.com/episode/3aKOMMdAHAlSNiigOvSlws?si=ATMjytJHQg2YILHvuEA0Dg
- You don’t know anything about steel: https://open.spotify.com/episode/2myA5oLPbB5pdGaEHP20zr?si=12w2Bey1RKOzBu4s684B1w&t=997
.
Here are two episodes I liked:
- Stories from a materials salesman: https://open.spotify.com/episode/3aKOMMdAHAlSNiigOvSlws?si=ATMjytJHQg2YILHvuEA0Dg
- You don’t know anything about steel: https://open.spotify.com/episode/2myA5oLPbB5pdGaEHP20zr?si=12w2Bey1RKOzBu4s684B1w&t=997
.
BOM cost for Baidu Robotaxi is said to be about $28K
Not bad for 7 million rides
https://carnewschina.com/2024/08/23/baidus-robotaxi-fleet-has-operated-7-million-rides-as-of-end-of-july-2024-according-to-companys-earning-call/
Not bad for 7 million rides
https://carnewschina.com/2024/08/23/baidus-robotaxi-fleet-has-operated-7-million-rides-as-of-end-of-july-2024-according-to-companys-earning-call/
Integrated Circuits (ICs) in Phones
They are: ICs are tiny electronic circuits etched onto a single piece of semiconductor material, typically silicon. They contain numerous interconnected components like transistors, resistors, and capacitors, all miniaturized to an incredible degree.
Key Role in Phones:
Brain of the phone: The most crucial IC is the System on a Chip (SoC), often called the processor or CPU.
Communication: ICs enable phone functions like:
Cellular connectivity: Handling calls, data transfer (4G, 5G), and other wireless communication.
Wi-Fi: Connecting to wireless networks.
Bluetooth: Connecting to other devices like headphones and smartwatches.
GPS: Location tracking and navigation.
Image and video processing: Powering the camera, and displaying images and videos on the screen.
Audio processing: Handling sound for calls, music playback, and other audio functions.
Power management: Efficiently managing the phone's battery life.
RAM: Temporary storage for active applications and data.
They are: ICs are tiny electronic circuits etched onto a single piece of semiconductor material, typically silicon. They contain numerous interconnected components like transistors, resistors, and capacitors, all miniaturized to an incredible degree.
Key Role in Phones:
Brain of the phone: The most crucial IC is the System on a Chip (SoC), often called the processor or CPU.
Communication: ICs enable phone functions like:
Cellular connectivity: Handling calls, data transfer (4G, 5G), and other wireless communication.
Wi-Fi: Connecting to wireless networks.
Bluetooth: Connecting to other devices like headphones and smartwatches.
GPS: Location tracking and navigation.
Image and video processing: Powering the camera, and displaying images and videos on the screen.
Audio processing: Handling sound for calls, music playback, and other audio functions.
Power management: Efficiently managing the phone's battery life.
RAM: Temporary storage for active applications and data.
you know, i'm all for innovation, but now that a bunch of startups think they can bring consumer-quality tech to warfare solutions, i think it's only fitting to remember when a tech company brought military-grade manufacturing to a consumer juicer
https://x.com/jwt0625/status/1875366631368790198
https://x.com/jwt0625/status/1875366631368790198
I learned recently about why Spherical Harmonics are really good for things like NeRF and 3DGS. i.e. making sure each Gaussian can have radiance values encoded using spherical harmonics, can use SH co-efficients for efficient reconstruction of lighting effects, good lighting, realistic etc - pretty cool
https://mathworld.wolfram.com/SphericalHarmonic.html
https://mathworld.wolfram.com/SphericalHarmonic.html
Pretty fascinating video on Photonics being used in chip development. Perhaps a bit sensational so read the comments for context, but still pretty cool to see.
https://youtu.be/wBqfzj6CEzI?si=GhpTNBrzNSGcxZfo
https://youtu.be/wBqfzj6CEzI?si=GhpTNBrzNSGcxZfo
"Willow can reduce errors exponentially as we scale up using more qubits. This cracks a key challenge in quantum error correction that the field has pursued for almost 30 years."
https://blog.google/technology/research/google-willow-quantum-chip/
https://blog.google/technology/research/google-willow-quantum-chip/
4262
0age
@0age·15:37 04/12/2024
something consistently in the back of my mind whenever i’m working on cross-chain swaps / chain abstraction:
does this thing still get the job done once we’re multiplanetary?
blockchains in space are absolutely inevitable, and if what we’re building doesn’t work on mars then GTFO
does this thing still get the job done once we’re multiplanetary?
blockchains in space are absolutely inevitable, and if what we’re building doesn’t work on mars then GTFO
