teraflop

That's an excellent reason to change the page design occasionally, so that people learn not to rely on it (the hard way if necessary).

Ethernet flow control doesn't really fix congestion except in very special cases.

Consider a very simple topology:

    A         C
     \       /
      S1===S2
     /       \
    B         D

Say hosts A and B are both sending data to C, as fast as they can, via switches S1 and S2 (which are connected via a high-speed link). And say the sum of these two flows is more than the capacity of the link to C.

S2 is receiving packets destined for C faster than it can forward them, but sending an Ethernet pause frame from S2 to S1 is not a very productive way to alleviate the situation, because it also disrupts any traffic that would be bound for D. It just moves the bottleneck elsewhere and causes collateral damage.

I'm no biologist, but it's pretty easy to imagine a plausible mechanism.

Stress doesn't cause your body to "immediately grenade itself", it causes gradual problems over an extended period. But you can imagine that evolutionarily speaking, a stressful situation would frequently be a situation involving imminent danger. A response that prioritizes short-term survival (e.g. releasing adrenaline and cortisol) regardless of the long-term consequences could easily improve evolutionary fitness.

The lunar near side is not shielded from impact craters in any significant way. Impactors can cross the earth's orbit and hit the near side of the moon just as easily as they can hit the far side.

The fact that the near side appears to have large areas with few craters, and the far side doesn't, is a result of many near-side craters having been erased by lava flows. It's not because it received fewer impacts.

The article says a new crater this size is expected about once every 132 years. Presumably, smaller craters are more frequent. So what's the approximate size of the largest crater that you'd expect, say, in an average year? And would it be scientifically useful to design a mission around sending a rover to explore one of those brand-new craters?

It's at the top of the article:

> but for the overview and demo of it working, watch the following youtube video.

If you don't care about the overview and just want to see demos, they start around 9 minutes in.

The "normalization of inexplicability" is indeed infuriating. It has always been bad when it comes to computer software, and it's increasingly creeping into other consumer products that depend on embedded software.

I bought a new electric car recently. For the most part I've been quite happy with it. Shortly after I bought it, it started popping up a warning message saying "check EV system" every time I started it. By the time I brought it into the dealership, the warning had gone away, and the technician just told me something to the effect of "eh, I guess it just does that sometimes, let us know if it happens again." Hardware fault? Software bug? Who can say?

Like most modern cars, it has connectivity and Google Maps built into the infotainment system. The vast majority of the time, it works fine. Sometimes it says it has no connectivity (meaning no traffic data and suboptimal routes) for the duration of a drive, even in areas with a strong cell signal where it normally works fine. Sometimes the car says it has connectivity, but Google Maps still thinks it's offline. Sometimes Maps will actually load and display a route, but the "start navigation" button just spins forever as though it's still waiting for something. Are these related issues? Is there a common cause that might be fixable? Who can say?

(Conveniently enough, the warranty specifically does not cover any failures of software or firmware to operate correctly.)

The flip side of this laziness is that now, when my elected representatives tell me "these are just license-plate readers that don't record video", I have evidence to show them that's false.

If Flock had done a more competent job of securing their system, it would be harder to demonstrate this in a compelling way. To a technically-inclined person, it's obvious from the get-go that somewhere in Flock's pipeline, video is being recorded and archived, and is therefore vulnerable to misuse. But the more they're allowed to keep the implementation proprietary, the easier this is to sweep under the rug.

Separately from how you present the number, the very concept of "uptime" as a single number is a bit muddy in the context of a distributed system, where different components can be differently available for different users.

Also, 0.1% downtime in the form of a 45-minute outage per month is very different from 0.1% of requests failing in brief bursts. You often see downtime reported as "increased error rates" which is so vague as to be meaningless.

Google's "windowed user-uptime" attempts to deal with this a bit better, by exposing different views of the data instead of trying to condense uptime into a single number: https://www.usenix.org/system/files/nsdi20-paper-hauer.pdf

Beyond "not working", concurrent editing is likely to silently lose data, or corrupt data, or possibly just fill up your drive with lots of "conflicted copies" that you have to manually reconcile.

For anyone who finds themselves in the vicinity of Atlanta. I can highly recommend the Mimms Museum of Technology and Art. Which was also formerly called the Computer Museum of America, but AFAIK has no relation to the one in San Diego.

Anyway, it's clearly a labor of love. They have some very cool artifacts from the space program (e.g. a some of the Space Shuttle and Apollo main computers). And a large fraction of the floor space is devoted to a new, very comprehensive exhibit on the history of Apple, with just about every product they ever made plus some internal prototypes.

The blog post doesn't seem to mention what strikes me as the most interesting application of a model like this, namely extracting semantic information from its latent space. It mentions robotics applications, but only in the context of generating realistic world models for simulation.

If you have a robot deployed in an environment, generating synthetic views of the environment you're in doesn't have any obvious value. What does have obvious value is the latent knowledge that the model could have used to generate those synthetic views.

For instance, the fact that Atlas is capable of identifying regions of the input images that look like "floors", and smoothly interpolating them and filling in gaps with more floor, suggests that it has a concept of "floor-like walkability" which it's learned from the examples in its training data. And being able to identify the regions of 3D space that correspond to that semantic label would obviously be useful for robot path planning.

There's plenty of literature about e.g. using neural networks to estimate walkable areas from a point cloud. And you could imagine just bolting one of those methods to the front of Atlas, using the synthesized point cloud (instead of traditional photogrammetry or LIDAR) as input. But that seems like it's throwing away a lot of potentially useful semantic information, on top of being needlessly inefficient.

Surely there are all kinds of interesting and constructive classroom discussions to be had about lawbreaking.

Classrooms teach about reality (among other things) and there are many things in reality that don't involve obedience to the law, as it exists at a particular time and place.

See also: https://en.wikipedia.org/wiki/The_purpose_of_a_system_is_wha...

Clearly these products are being named by topologists, who have no objection to things being simultaneously open and closed.

Another interesting non-hypothetical example: many US states have so-called "blue alerts" which are used to send cell phone alerts whenever a police officer is attacked or injured. I used to live in Texas and blue alerts were a fairly frequent occurrence.

The stated justification for blue alerts is to warn the public of dangerous criminals. But they are only issued in response to attacks on cops, regardless of the level of danger to any other people. Also, at least in Texas, they're broadcast statewide. Texas is an awfully big state, and so the alerts are guaranteed to be irrelevant and non-actionable for the vast majority of people who receive them.

So it's fairly obvious to me that blue alerts are first and foremost a propaganda operation, aimed at perpetuating the stereotype that policing is a uniquely dangerous job. (It is indeed more dangerous than an average office job, but it has a much lower rate of fatal injuries than construction, or farming, or driving garbage trucks.)

> The workflow had an if: condition that appeared protective:

> if: (github.event_name == 'issues' && github.event.pull_request.user.login != 'whitesource-for-github-com[bot]')

> However, on issues events, github.event.pull_request is always null.

This is extra dumb because even if you thought this condition was correctly testing the user's identity, it shouldn't have "appeared protective" upon even a moment's thought. If it worked correctly, it would obviously just exclude one bot user while allowing all other users, so it wouldn't provide any protection at all.

But more likely, this condition was never intended to be "protective" at all, and it's only being described that way because the writeup is LLM slop.

Because this particular sensor network is only monitoring earthquakes in the Pacific Northwest of the USA, not worldwide.

I guess the headline has been changed? The HN title says "illegally drills wells" but the article now says "prepares to drill".

Reading the article and the linked government statement, it sounds like the issue is that the company did receive permits to ship their equipment to one location in Greenland, and then moved that equipment to a second location without permission, while the overall drilling operation is still waiting for approval.

Still bad in the sense that they're presumably trying to push boundaries, but it's a far cry from actually carrying out "illegal drilling".

Agreed, but I'll just add that annular eclipses occupy an intermediate point on that scale.

With a partial eclipse, shadows look odd and the sunlight "feels" noticeably dimmer and more muted, but if you're not paying attention you might not notice much out of the ordinary.

An annular eclipse is much more dramatic and unmistakable. It's not as striking as totality, and of course you can't directly observe it without eye protection, but the "ring of fire" is beautiful in its own way.

True, but predictability cuts both ways.

We're all used to having to constantly update our browsers and phones to keep up with the security arms race. If a frozen model can't be updated, it will predictably remain vulnerable to any "exploits" or idiosyncratic quirks that people discover over time.

Let's say, as somebody suggested in another comment, that you buy 100,000 of these chips and deploy them to run fast-food drive-thrus. And then somebody discovers the model has a fondness for goblins[1], and if you role-play convincingly enough, you can get it to accept payment in shiny buttons and rodent skulls instead of cash.

What do you do then? I guess your options are to try and fix the behavior with a better prompt, or put some kind of filter in front of the model to catch attempted exploits. If the filter is cheap and dumb it probably won't work well enough, and if you use another model as a filter, you've negated the cost and speed benefits of putting the first model in hardware.

Of course the real answer is to just never expose the model to situations where an adversarial input could possibly lead to an undesired output. But that drastically limits what you can do with it.

[1]: https://openai.com/index/where-the-goblins-came-from/

This is indeed a very elegant fastening method.

I wouldn't be surprised if the designer of that CNC machine borrowed it from the popular Prusa MK2 3D printer, which uses a very similar design for its Y-axis belt: https://help.prusa3d.com/guide/2-y-axis-assembly_79992#81120

I mean back in the day, those kinds of titles involved film editing, not video editing. As in, physically filming title cards with a camera (and then maybe optically compositing them on top of other elements). It's easy to see how things could get slightly misaligned.

Looks like the misalignment is less than 1 degree. It's the kind of thing that stands out very noticeably in an image on a web page, when there are lots of perfectly horizontal edges to compare it against (e.g. the edge of the viewport). But I think it would be much harder to notice if projected onto a movie screen in a dark theater.

Look up "long arm statutes". State courts can have jurisdiction over out-of-state entities, subject to limitations established by federal precedent. Doing business with customers who reside in a state generally puts you under that state's jurisdiction, at least for purposes related to that business.

> therefore I write (hard?) sci-fi.

This is not at all what Le Guin is saying! It's right there at the end of the essay: "I agree with my reviewer that I don’t write hard science fiction."

There was a bit of a fad for stuff like this, maybe 20 years or so ago, as soon as browser JS/CSS support started getting good enough to support it.

All of a sudden, we had the ability to build complex interactive UIs in a web page, and we had to figure out how to use that ability. So naturally a bunch of people had the idea of just cloning existing desktop apps and styles, up to and including the full Windows desktop.

There were a few people who thought cloning the desktop UX into a browser would be "the future of computing" (see e.g. YouOS, a YC startup: https://www.benh.org/techblog/2006/07/youos-a-new-kind-of-co...) but I think most people just saw it as a fun way to show off your web dev skills.

I think it persists out of nostalgia, not just for the operating systems themselves, but for a particular era of web development history. And you still see this kind of UI occasionally in real products, e.g. Synology DSM.

The official website (https://www.staraxis.org/) has this to say:

> Fundraising is currently underway to complete construction this year and plan for the future of Star Axis. Star Axis will not open to the public for several years. Thank you for your patience.

> Your contribution to our fundraising campaign will provide opportunities for you to visit Star Axis between June and September 2026.

Which seems to hint that they're not planning to keep the current pricing model indefinitely. Hopefully it becomes more affordable once the construction has been fully paid for.

On the other hand, just looking at the pictures, it seems like the site might get pretty crowded if there were a significant number of visitors per day. So I guess there needs to be some kind of gatekeeping to prevent congestion. But maybe the remote location is enough to do that.

The coefficients are stored in writable memory as input data. Only the structure and degree of the polynomial are hard-coded in the program. So if the launch is delayed you just change the data.

Of course, the justification is that compute and memory were at an extreme premium on the spacecraft. Accurately predicting the position of the moon at arbitrary times requires more complicated calculations. You can dramatically optimize the calculation by fitting a polynomial approximation which is valid for the length of the mission. And there's no reason to waste resources fitting that polynomial in space when it could be done on the ground, using much more powerful machines.

The AGC is what tends to get the press. It's easy to forget that the Apollo program relied on a cutting-edge distributed hardware and software stack, most of which was in Mission Control at the Real Time Computing Center.

What does "no cash settlement" mean? If I sell a reservation, do I not get cash?

The short-term goal of a tool like this is to sell products. The more ambitious long-term goal is to shift cultural norms, blurring the lines between advertising and reality until the question you're asking is no longer consciously asked. At least, not by average people, and not at the point of purchase.

I find it easy to envision a world, maybe 50 years from now, in which the very concept of "truth in advertising" is viewed as a lost, idyllic fantasy. Something people are nostalgic for, but feel powerless to regain.