quanto

Your answer shows why flat surfaces were acceptable for a bomber (with an asterisk) given the constraints.

Then, the question becomes why is the B-2 bomber (and latest stealth fighter jets) not angular like F 117.

Scientific computing has come a long way since the days of F-117. We do not need flat surfaces to scatter the beams away. Finding optimal curved surfaces that scatter the beams away is just a harder math problem.

The whole mirror analogy for angular facets breaks down when you realize airplane is a moving object and at some point in its trajectory, one of the facets is likely to be facing the radar (maximum cross section).

If I remember from my numerical analysis days, it was the ease of numerical computing that necessitated the facets. Think 3D graphics from the 90s. All polygons.

The numbers seem to agree. According to its SEC filing, “in part helped by progress in AI, revenue per full-time equivalent Spooner increased from $1.12 million in 2023 to $2.57 million in 2025, and was $0.97 million in Q1 2026.”

Is there a typo somewhere?

There are so many derisive comments here.

David Ha, CEO and co-founder, was one of the youngest managing director at Goldman Sachs before doing ML at Google. His ML publications were considered top-notch almost a decade ago. I had high hopes for him when he raised money and founded Sakana.

I do agree with some comments here that perhaps this particular product is not well thought out. I also agree with the criticism that David calls Sakana a frontier AI lab while making money just selling AI B2B applications to Japanese businesses. I also agree with the assessment that Sakana has abrasive and antagonistic, sometimes openly hostile, recruiting tactics. I also agree that his then-impressive publications may have lost their luster in the age of LLMs.

However, the man is clearly driven; and he and his team may have more to offer in future. I admire the man for not taking the conventional AI-research career path.

There could be geometrically tiny optimizations that lead to an outsized impact in noise and flow by turbulence reduction. While optimizing an impeller with computational FSI (fluid structure interaction) is not as hard as before, it still is not trivial. And it's these (perhaps small) optimizations that justify Noctua being 5x more expensive than generic black fan.

Interesting. This observation only reinforces the hypothesis that my colleague had. Well optimized components procured with exclusivity.

No, not Rentech. While Rentech is truly marvelous, it has been around for a while. There is a new kid in town, and this kid is very very good. To the point that it drove out other veteran shops out of the same market it traded.

I have an interesting discussion with a senior colleague: why ASML? why are they by far the best? Their competitors are a few generations behind.

The colleague claimed that there is no special magic. It's not that ASML is using some otherwise unknown laws of physics nor is any single step or component particularly special or novel. It's just that they meticulously optimized each step, and the sum of such steps is the winning solution.

In fact, this is probably why it's so hard to copy ASML. If there was a single magic component, a single or few engineers could be poached away to a competitor to copy it. However, copying a well-optimized company with many simultaneous optima is a much harder task.

Our discussion was in the context of why our quant hedgefund competitor was performing so well, far above the market norm. By nature and design, quant finance is an incredibly efficient field (and most techniques are more or less known by veterans), and we had thought unlikely that one fund could do so much better. Our conclusion was that this fund must be the well-optimized ASML of our field. My colleague happened to know the founder and indeed that was his personal impression as well.

Only Nvidia currently rivalled Google’s combination of AI hardware and integrated chip software, he added.

Does the phrasing imply that Google sees itself as #1 and Nvidia as a close #2 that may rival its greatness? not to mention other big contenders.

the output quality is so good that it triggered my LinkedIn PTSD. and at least for this particular output, the input-output pair is reversible.

input: I just ate lunch.

output: I’m thrilled to share that I’ve just successfully completed my midday refueling session!

Nutrition is the ultimate foundation for high-performance leadership and sustained productivity. Taking this intentional break has allowed me to sharpen my focus and realign my goals for the afternoon sprint.

How are you optimizing your energy levels today to drive impact? Let’s connect and discuss the power of strategic self-care in the workplace.

#Leadership #Productivity #HighPerformance #WellnessAtWork #GrowthMindset

What are the techniques (and the papers thereof) that you consider to be spectacularly good before 2021 for depth estimation, monocular or not?

I do some tangent work from this field for applications in robotics, and I would consider (metric) depth estimation (and 3D reconstruction) starting to be solved only by 2025 thanks to a few select labs.

Car vision has some domain specificity (high similarity images from adjacent timestamps, relatively simpler priors, etc) that helps, indeed.

A mech eng PhD said that he watches Gundam/Macross for drama and Romance (as in Romanticism).

The challenges you mentioned, and techniques to address them, are not unique to quantum physics. I am still not understanding how quantum physics require "new" kind of numerical analysis. And what are these new kinds of techniques you hint at? Could you give me some examples of unique techniques that arose from quantum physics and are not used elsewhere?

PDEs are PDEs, regardless of where they come from Newtonian or quantum. Would you care to elaborate why you think quantum requires a new kind of numerical analysis?

when I can get away with it I'd prefer to prove things with real numbers and assume magically they transfer to floating point.

True for some approaches, but numerical analysis does account for machine epsilon and truncation errors.

I am aware that Inria works with Coq as your link shows. However, the link itself does not answer my question. As a concrete example, how would you prove an implementation of a Kalman filter is correct?

During the days I was studying/working with Coq, one visiting professor gave a presentation on defense software design. An example presented was control logic for F-16, which the professor presumably worked on. A student asked how do you prove "correctness", i.e. operability, of a jet fighter and its control logic? I don't think the professor had a satisfying answer.

My question is the same, albeit more technically refined. How do you prove the correctness of a numerical algorithm (operating on a quantized continuum) using type-theoretic/category-theoretic tools like theorem provers like Coq? There are documented tragedies where numerical rounding error of the control logic of a missile costed lives. I have proved mathematical theorems before (Curry-Howard!) but they were mathematical object driven (e.g. sets, groups) not continuous numbers.

This paper reminds me of a class assignment in grad school where the prof asked the students to write a compiler in Coq for some toy Turning-complete language in a week. Having no background in compiler design or functional programming, I found it daunting at first, but eventually managed it. The Coq language's rigor really helps with something like this.

I wonder if AI's compute graph would benefit from a language-level rigor as of Coq.

Could you elaborate? Which two was he bad at? How did he ruin people's lives? Honestly asking.

How is Apple's nano-textured display different from ThinkPad's famed matte display?

Both China and Japan were agrarian and primarily rice-eating too, so if the presented narrative was true, what made Korea special?

Job losses could shave 30 cents off each item purchased by 2027.

This is incredible. It's far less than I would imagine. It represents how well optimized the warehouses are. If we roughly estimate a median product price to be $20, then the automation represents less than 2% cost saving. Of course, Amazon is at a scale that this is still net positive despite all the R&D cost. But if automation was to reduce the cost of living, there are probably better areas to focus on.

I am amazed how a narrative could be formed by select samples. The Korean peninsula has very little arable land, and much of Joseon Dynasty's history was marked by famine and mass starvation.

What a great project.

What's interesting is that action tokens are learned from video. In other words, the training dataset does not include actions like "go left" and "go right"; and these actions are learned from the pixels that moved. This means that learned actions may not map exactly to the game actions available to the user. That means we (humans) cannot necessarily use this world model to play the game.

I suspect the inferred actions probably directly correspond to human-understandable actions; and after playing with the action tokens, a reasonable human can probably guess what, say, the third action token in the dictionary corresponds to ("jump"). This is likely as game actions are sparse (in both time and action spaces) and often independent/orthogonal (in action space).

The latest education report from the Organisation for Economic Co-operation and Development (OECD) raised alarm in Denmark when it found 24% of Danish 15-year-olds cannot understand a simple text, up four percentage points in a decade.

So, in 2015, 20% of 15-yo could not understand a simple text. Isn't that unbelievably high?

I had a colleague, an architect deeply soaked in the Design Thinking cool aid, telling me that architecting a single-family house is humankind's most intellectually challenging endeavor because he has to worry about the end users and construction materials that go in ("holistic[TM]"). I asked him if building a space shuttle is easier than building a house, and he genuinely believed that engineers have a far easier (and dumber) time than designers like himself.

This take of designers being superior being to engineers is something I consistently observed among designers over the decade.

Here is a light-hearted video: https://www.youtube.com/watch?v=uvU5dmu4sl8

The last time I checked about a decade ago, a full license with multiple users and on a server was around 100k USD. I don't recall exact number of seats or server count restrictions.

I want to add that, for many in the industry, it is well worth the price.

I would like to see evidence behind this claim on two fronts: 1. all parts are produced in other countries 2. slightly higher prices

There are countries (including Ukraine) that produce on-board flight controllers, but the controllers themselves often rely on components from China. I actually do think it is feasible to create a passable quadrotor using non-Chinese components only but I do not know of a rigorous study or a manufacturer that does this.

while we might learn from history - that that is not the reason to "do" history.

If we don't learn from history, why do we do history? Is it a form of pure entertainment, i.e. of arts? If so, does that give more credence to White's argument?

I have had a PhD colleague who genuinely believed that history ought to be (in the philosophical normative sense) contributing to national propaganda, thus of national interest. By extension, this is why history departments should be funded by tax dollars.

Today, engineers working on AI systems also need to think deeply and critically about the relationship between language and culture and the history and philosophy of technology. When they fail to do so, their systems literally start to break down.

Perhaps so. But not in the (quasi-)academic sense that the author is thinking. It's not the lack of an engineer's academic knowledge in history and philosophy that makes an AI system fail.

Then there’s the newfound ability of non-technical people in the humanities to write their own code. This is a bigger deal than many in my field seem to recognize. I suspect this will change soon. The emerging generation of historians will simply take it for granted that they can create their own custom research and teaching tools and deploy them at will, more or less for free.

This is the lede buried deep inside the article. When the basic coding skill (or any skill) is commoditized, it's the people with complementary skills that benefit the most.

A genuine intellectual question: what's driving the (hardware) cost of these humanoid robots? Dynamixel smart servo motors? Exotic sensors?

I do robotics research but not humanoid. Almost every component used in production came down drastically in price in recent years. Lidar sensors are a few hundred dollars now.