I don't necessarily disagree with where you're going here, but can you elaborate on this:

I've worked at government contractors that had real air-gaps for things like their databases, but that does not seem to be the norm for the rest of the world. How would ordinary businesses make use of their databases if they are not network accessible under any circumstances, printed reports? Some sort of unidirectional transmission? What sort of data ingress are they using?

I ask this because I have been involved in the transfer of data in highly regulated, air-gapped systems, and they are incredibly expensive. Are you really indicating that true air-gap databases will be ubiquitous (or at least commonplace) in the forseeable future?

It's a cool idea and they stand for a lot of great ideals, but laptops are incredibly hard to get right, drivers are hard to get right, and they are a small team trying to support a large number of possible configurations. Hardware gets more complicated by the year: forget the CPU and various GPUs, just look at how many other devices in a modern computer have a full-on processor, e.g. fancy touchbars, displays, even hard drives! Hell, your CPU probably has its own secondary general-purpose processors for things like security, and our CPUs themselves get firmware updates now to change how their instructions function. They are doing great work, but the deck is so stacked against them that it's not funny.

Most space launch companies are inefficient and ineffective. SpaceX has the margin to pay these taxes, those unfortunates don't. If you want to kill competition in an industry, tax it enough that only the large corporations can survive the loss, and add some complicated regulations in for extra effect. No one else has anything close to what Starship may become, and further reduction in margins will ensure that SpaceX will have a defacto monopoly on non-military space launches while their competitors are strangled paying for FAA services that is disproportionately benefit owners of private jets and charter flights for the rich.

The idea that software not signed by some entity in the country that they represent is installed any high-level official's phone is absurd. How did the world get this way?

I've been using this site for over twenty years, and it's a been most of a decade since I've commented. This is the best thing I've seen on here since then. Whatever you do, keep drumming up the fight against ignorance and propaganda, and the people who've fallen victims of it. I don't want to get personal, but lets just say that I know from intimate experience what brainwashing does to a person, and the tremendous cost of clawing one's way out of it. Division in modern society is inevitable--and we must fight against those who seek to destroy rational thought!--but without empathy for those infected by bad ideas, shortchanged by their personal experiences, we'll end up punishing and alientating those victimized by bad actors exploiting cognitive vulnerabilities that every one of us has, we will push them out of sheer self-defense into voting in the people who will undo us.

Not the person you're responding to, but I'm pretty sure their less-than-polite phrasing meant "biological female who has ever been involved in jetpack development or *jetpack* flying".

Everyone knows Amelia Earhart was a big part of aviation history in that era, but I strongly suspect that she didn't moonlight as the Rocketeer.

Indeed, people should take in the full context here, not just platitudes. They should look at how China has interacted with Europe and the US in the context of world wars, trade wars, cultural exchange and economic interdependence. ...then they should call Xi on his tantrums and fragility. He chooses to be this way.

I like the general sentiment here, but I think there are a lot of details to iron out, and we would have to be careful regarding unintended consequences.

For instance, there's the matter of how to treat trade secrets, which are common in computer code. In many cases, the creator of a work doesn't even have a right to distribute source code that they've purchased a license to (say, a game engine) and have modified, so this is untenable unless you are willing to make entire business models completely flat.

I suspect on re-reading your comment that you mean the portion of a work that is distributed--in this case the game client. There are less issues with that, but still licensed assets are a fairly reasonable part of the copyrighted works market. Perhaps unlimited duplication after a lapse time would be allowed, but derivative works would not be?

It's an interesting thought. It's not going to happen, but something like it's now on my wishlist.

Actually, at 30 mentions of owners and ownership in Title 17, Chapter 2 alone, you are dead wrong:
https://www.copyright.gov/titl...:

Read all of the laws there. You will find plenty more mentions. And in case you try to backpedal and amend your statement, since the term is used to describe the copyright itself, and not the work, you can find the term "owner of a work" and "ownership of a work" in multiple official documents associated with our government's various copyright bodies:
https://www.federalregister.go...
https://www.copyright.gov/docs...
https://www.copyright.gov/poli...

I'm not saying I agree with US copyright law, but lets get our facts straight. Your conclusions may (or may not) be valid, but that particular argument regarding legal wording is so wrong that I have to wonder if you've even read these laws.

Bonus: Contrary to your main argument, DCMA *does* in fact prohibit actions involving circumvention of copyright--many of which are actions taken for personal use, say, displaying a legitimate copy of a video from a computer by illegally circumventing HDCP or the like. This is absurd, but that's how the law was written, and I doubt it was put there by accident.

So, one problem is that there is not always more data. In my field, we have a surplus of some sorts of data, but other data requires hundreds of thousands of hours of human input, and we only have so much of that to go around. Processing all of that is easy enough, getting more is not.

Also, by "effective", I should have made it clear that I meant "an effective overall solution to the problem", which includes all costs of training a wider, lower-precision network. This includes input data collection, storage and processing, all of the custom software to handle this odd floating point format, including FP16-specific test code and documentation, run time server costs and latency, any increased risks introduced by using code paths in training and , etc.

I'm not saying that I don't believe it's possible, I've just seen absolutely no evidence that this is a significant win in most or even a sizable fraction of cases, or that it represents a "best practice" in the field. Our own experiments have shown a severe degradation in performance when using these nets w/out a complete retraining, the software engineering costs will be nontrivial, and much of the hardware we are forced to run on does not even support this functionality.

As an analog, when we use integer based nets and switch between 16-bit and 8-bit integers, we see an unacceptable level of degradation, even though there is a modest speedup and we can use slightly larger neural nets. I'm very wary of anything with a mantissa much smaller than 16 bits for that reason--those few bits seem to make a significant difference, at least for what we're doing. We're solving a very difficult constrained optimization problem using markov chains in real time, and if the observational features are lower fidelity, the optimization search will run out of time to explore the search space effectively before the result is returned to the rest of the system. It's possible that the sensitivity of our optimization algorithm to input quality is the issue here, not the fundamental usefulness of FP16, but I'm still quite skeptical. If this were a "slam dunk", I'd expect to see it move through the literature in a wave like the Restricted Boltzmann Machine did.

Oh, and thank you for the like (great reading) and the thoughtful reply. Not always easy to find on niche topics online.

This is by far the best comment I've ever seen regarding this sort of algorithmic labor management.

Normally I'm all for this sort of thing--my company is a client and uses it to handle large bursts of data processing quickly--but the information symmetry argument is a powerful one. Also, there doesn't seem to be a lot of competition in this space, which might otherwise ameliorate a lot of the problems induced by the "take it or leave it" bargaining approach.

The analysis provided by the article is absurd, but yours seems to lead to the inescapable conclusion that some kind of regulation is necessary to prevent blatant exploitation. Maybe just reducing information asymmetry in some way, or requiring transparency in reports available to the public on the website regarding effective wages paid to workers as a fraction of the minimum and average wages of employees in their respective countries. Surely someone can find an answer to this.

Accidentally posted as anonymous coward, reposting under my actual name.

There's a lot of rounding error with FP16. The neural networks I use are 16-bit integers, which work much, much better, at least for the work I'm doing. Also, do you have a good citation that FP16 neural networks are, overall, more effective than FP32 networks, as you've described?

The suggestion that a CS degree isn't worthwhile is preposterous. I lead a fairly large organization and I've hired dozens of software engineers over the years and hundreds of interns. With only a few exceptions, we find that self-taught programmers have some superficial skill in the languages or platforms they tinkered with but lack CS fundamentals that enable them to build well designed, maintainable, and performant systems. Their code doesn't adhere to patterns and standards that make it easy for other programmers to understand. They struggle to decompose complex problems and don't have a mathematical background to tackle the biggest challenges. They often haven't even explored the full capabilities of the languages they use. Yes, there are exceptions, but we've found that a CS degree from a good institution to be a very valuable indicator when selecting our employees. It's the difference between a home cook and a chef trained in a culinary institute.

I second this as well. I'm a computer scientist who spends almost all of his work time in front of a computer and a lot of my leisure time as well, but my two children (aged 7 and 5) never watch television or use computers (truly). What do they do instead? They read, climb, ride, draw, build, etc.. - they're constantly learning how to play, interact with others, and control their bodies. The time will come for them to learn how to use computers and I'll be there, ready with a series of great programming projects that we can work together on, simple robots - the works. I look forward to that, but it's not like there's any shortage of things to do with them in the meantime!

Numerous studies have shown detrimental effects to child brain development associated with the early introduction of television and computers. No, you can't get around these detrimental effects by using "age-appropriate" or "educational" shows and games. Apparently, something about the *medium*, not simply the message, is causing these negative impacts. Perhaps it's the pacing of the material, the quick transitions, or the sugary over-the-top positive feedback that they get for completing even the most simple task in an educational game. For me, I don't need to know exactly what is causing the negative impacts - I see no serious side effects associated with withholding computers and television for now.

I know you might thing that you're helping your child out by letting him get an early start on computers, but keep in mind that some things are much easier for someone to learn at a particular age. For example, I always have to bite my tongue when one of my friends tells me with great pride "It's AMAZING! My two year old already knows how to use a mouse!!!" I always *want* to say "Were you really under the impression that he was going to have a problem picking that up?" Honestly, people, it's a MOUSE. It's DESIGNED to be easy to use. I think it would take a 10, 20, or 30 year old with no training about a minute to figure it out and an hour to master it, and they'll understand the full context and purpose of the mouse and the computer - things that the 2 year old couldn't possibly comprehend. Controlling a mouse isn't something that your child needs a "head start" on.

You could attempt to teach your 2 year old the periodic table, but I think we can all agree that he'll pick that up just fine when it's time to study Chemistry in high school. Maybe it's time to focus on how to dig holes, stack blocks, and chase you around the house instead?

I think the Slashdot community would really get a kick out of this. We're inviting coders to help us solve hard problems that will have a tangible effect on the exploration of Mars. I can't wait to see what people come up with!