manu (@manuflog) header
manu (@manuflog) profile picture

Human | All-In for the future | Engineer | MSc | MBA | Enjoying the journey 🫠

Joined May 2020
Photo by manu (@manuflog)Photo by manu (@manuflog)Photo by manu (@manuflog)Photo by manu (@manuflog)Photo by manu (@manuflog)Photo by manu (@manuflog)Photo by manu (@manuflog)Photo by manu (@manuflog)Photo by manu (@manuflog)Photo by manu (@manuflog)
To understand most things Musk says, you have to understand that engineers are born, not made. You have to understand that he has always been an engineer and pretty much nothing else. And you have to understand that 99% of an engineer's job isn't to create a thing... it's to teach ourselves how to create a thing. So when he says cooling in space is "easy", that doesn't mean what normies mean by "easy", which is "doesn't take a lot of resources, time, or expense". He means what engineers mean by "easy", which is "contains no unsolved problems". You take in X amount of energy, you know it's all gonna end up as heat, you know that to radiate X amount of heat you need Y amount of radiator. It's all just math. If you build a satellite that takes in 2X energy, you need 2Y radiator. There's no unsolved problem, here. There's nothing to figure out. You know the answer. You just have to design the layout. If you're been beating your head against "how land booster" and "how make engine not boom" and "how make rocket lighter" for years, this is like drawing stick figures by comparison. So if guy comes along and says "vacuum is a near-perfect insulator", which is both technically true and completely irrelevant, of course Musk is gonna call guy a retard. Because guy has just revealed that he doesn't understand the basic nature of the problem, or of the design conversation. The only significance that "vacuum is a near-perfect insulator" has is that it means you can't do cooling in space the same way you would on Earth. Well, duh. We've known that since we first understood what space is. And we've successfully dealt with that challenge since the very first times we put things in space. Fusion containment is hard. Rocket engines are hard. New AI capabilities are hard. Cheap and reliable manufacturing at scale is hard. Estimating costs and schedules is hard. Waste heat is not hard.
Yeah, much easier to cool in space. ~1300 W/m^2 is easily achievable for an aluminum radiator pointed away from the Sun, as it is mostly just seeing cosmic background radiation at ~3 degrees above absolute zero! Solar panels generate ~250 W/m^2, so you should expect them to have ~5 times the area of the radiator. Most of the mass of the satellite is solar panels.
154
377
30
6,213
247,839
.json voorhees
🤖 Made with AI
1
33
Replying to @Pokemon
@Pokemon @Pokemonlatam @PokemonTCG Hey Pokemon! @Gameplanetmx in México is an official distributor selling at more than 3x the ETB in Mexico. I propose you find new distributors and ban Game Planet forever. Please!!!🙏
1
2
166
For comparison. My 2 ETBs from Amazon. MSRP $1,099 MXN vs $3,499 MXN in Game planet -_-
16
First booster and a Gengar. 30th rules!! Sorry scalpers…
68
Tú, tanto 440 y yo aquí…
293
Replying to @Collect
@Collect hey!!! I found something weird with BGS pop report. You have BGS 1 of 10 none graded higher… but mine has surface 10. It’s graded higher. When searched in the database it says it’s not in the pip report but authentic. And I sent them for grading like 3 years ago. Mine converts easily to a PSA10 as well but there’s something to the gold label and subgrades I love. I’ve never understood the ibsessiin for PSA10. I’m pretty sure a lot of them would go down to 9.5 or 9s Link fanaticscollect.com/buy-now/…
1
253
I see why mine is not in the pop @beckettcollect , before they wer not catalogued with the text Box Topper @Collect
2
47
I also have a third one 8.5, all 3 are sequential
9
You would think so…. But books have been written algorithmically for longer than you would think. I worked back in 2019 with a lab named Totogeo associated with @INSEAD. At one point, they had the most published books on Amazon by Phil (founder). Thousands of books were generated, printed and sent. All based on searches, web scraping, and stochastic methods. @grok Dive into the specifics
Books written before gpt 3.5 are now pre-nuclear steel
1
129
@grok please verify and give me details…. I never asked too much
1
27
Peres (1991) and Penrose (2000) each proved Kochen–Specker with 33 rays in dimension 3. Penrose noted his rays are essentially complex and no rotation takes one set to the other. They're antipodes on the same circle. θ=0 is Peres, every chord a diameter. A quarter turn is Penrose. The orthogonality graph — 16 triads, 24 dyads, uncolourable — never moves. Exact certificates, closed form, verification suite: 10.5281/zenodo.21488474
🤖 Made with AI
195
Anyone else into WOTC box toppers???
1
365
There are 16 of them in total. 4 aquapolis, 4 legendary collection, 4 expedition… and @grok help me with some details about box toppers for Pokemon
1
79
Here’s a fun mathematical puzzle from quantum physics. Suppose you have 33 carefully chosen directions in space. Your goal is to label each one YES or NO. There’s only one rule: Whenever three of those directions are mutually perpendicular, exactly one of them must be labeled YES. Sounds simple. Except it turns out there’s no way to do it consistently. That impossibility is one of the classic proofs of quantum contextuality, and for decades people have studied special 33-direction constructions that demonstrate it. My friend Fable and I spent a couple of very long weeks digging into one of those constructions, expecting to verify a few details. Instead, we found something surprising. Two versions that everyone treated as different—the Peres construction and the Penrose construction—are actually connected by a single continuous parameter. Imagine turning a dial. As you rotate it, the construction changes smoothly, but it never stops working. Every position is still a valid proof. Turn the dial all the way around and you’re back where you started. That’s what the image shows: the entire family forms a circle. We proved that this circle is complete. There aren’t any missing pieces, disconnected branches, or hidden families. Every result was machine-verified using exact arithmetic. We weren’t looking for a new proof. We ended up finding that two classic proofs were really just different views of the same mathematical object. Credit where it’s due: this builds on the beautiful work of Asher Peres and Roger Penrose, whose original constructions started the story. Elizabeth Gould and P. K. Aravind first noticed the two constructions were related. Michael Kernaghan’s survey helped point us toward the right question, and the rigidity perspective builds on work by Stefan Trandafir and Adan Cabello.
1
153
Every claim ships with a script that checks it, so you can run the proofs yourself: 📄 The circle + exact rigidity certificates:
doi.org/10.5281/zenodo.21488… 📄 Its geometry and arithmetic, including a holonomy phase 2π√1867/33 that we measured on two IBM quantum computers to about 2%:
doi.org/10.5281/zenodo.21498… 📄 The alphabet theorems and the uniqueness proof:
doi.org/10.5281/zenodo.21498… 💻 All 40+ verification scripts, exact arithmetic, no floating point tolerances:
github.com/manuflog/contextu…
66
New model, yay, first interaction and I already have trust issues… 🫠
100
Quantum measurements can be contextual. The outcome of one measurement can depend on which other compatible measurements you perform at the same time. This effect is important in quantum computing, but until now it has been difficult to explain why it appears only in certain dimensions and how strong it actually is. The linked work provides a clear geometric reason for the difference between even and odd dimensions and shows that the obstruction has only one meaningful strength when it’s present. On the practical side, the accompanying open-source library automatically finds the smallest possible witness of this contextuality inside large measurement programs, making it much easier for researchers and hardware teams to detect, verify, and work with these effects instead of dealing with huge complicated sets of measurements. Papers and verification details: github.com/manuflog/contextu… Library and examples: github.com/manuflog/qkernel @grok please think of more ideas on how to apply this ✌️
1
1
107
50 years ago on March 22, 1976, Star Wars began filming in Tunisia. Among the shots captured were the first attempts at Luke Skywalker's view of the iconic twin sunset. Although that first day's sunset shots were not ultimately used, they marked the first live action plate photography ever captured for Industrial Light & Magic, who would be tasked with adding the suns to the background. ILM was less than a year old that spring of '76, and were busily preparing their visual effects pipeline at their Van Nuys, California studio. Read more here: starwars.com/news/star-wars-…
14
628
17
2,403
81,253