GamingChairModel

You can reason from a few principles:

• At its core, the math functions being optimized by these AI tools and their specialized hardware is that they can perform inference and pattern recognition at huge scales across enormous data sets.
• Inferring a rule set for pattern also allows generation of new data that fits that pattern.
• Some portion of human cognitive work falls within the general framework of finding patterns or finding new data that fits an old pattern.

So when people start making claims about things with clear, objective definitions (a win condition in chess, the fastest route to take through a maze, a highest lossless compression algorithm for real world text), it’s reasonable to believe that the current AI infrastructure can lead to breakthroughs on that front. So image recognition, voice recognition, and things like that were largely solved a decade ago. Text generation with clear and simple definitions of good or bad (simple summaries, basic code that accomplishes a clearly defined goal) is what LLMs have been doing well.

On things that have much more fuzzy or even internally inconsistent definitions, the AI world gets much more controversial.

But I happen to believe that finding and exploiting bugs or security vulnerabilities falls more into the well defined problem with well defined successes and failures. So I take it seriously when people claim that AI tools are helpful for developing certain exploits.

but isn’t the memory on the Neo on the same die as the processor?

Not actually on the same die, but in the same package, stacked on top using TSMC’s Integrated Fan-Out Package on Package (InFO-PoP).

So the memory still needs to be sourced from memory manufacturers, sent to TSMC, and then have TSMC package it all together in a single package. It’s unclear whether they had locked up this supply at pre-AI prices, though. The underlying A18 Pro chip/package was annoinced and launched about 18 months ago, so if they had the manufacturing pipeline set up for that they might have kept the contractual rights to continue buying memory at the old prices.

The website could know whether the username actually exists on the system. But revealing that information is a security weakness because someone could at least learn who has an account at that site (especially if usernames are email addresses, as they often are).

God I wish democracy meant that we could vote on decisions like this You can! Only problem is that it’s one vote per dollar instead of one vote per person.

First of all, “Intellectual property[sic]” is a not a thing. There are copyrights, patents, trademarks, and trade secrets, but they are all significantly different from each other. Trying to lump them together under a single term is disingenuous at best, and using the word “property” in that term is biased loaded language. You don’t get to redefine words like “property” or “intellectual property” how you see fit, completely untethered to the way the legal system uses those terms with specific meaning. Intellectual property rights include all of those things, in the same way that copyright can include copyright over text or musical compositions or sound recordings or photographs or building architectures. But note that copyright over each of those types of media is subject to its own rights and rules, and you’ll need to apply the correct rules to the correct contexts. But it’s still useful to group similar concepts together, and have a name for the category. That’s why people refer to intellectual property. A property right is a thing the owner is entitled to, and a natural right. This is a naive take. Property rights are natural rights? No, property rights are defined by the legal system of whatever sovereign nation you’re in. And they’re limited by whatever rules of that legal system are. If I own land in the U.S., I’m still required to pay taxes on it, and to enforce my property rights against adverse possession, lest I lose that property to the state or to a squatter. If I don’t record my ownership with the county recorder I might lose the property to someone else who comes along and records them buying it from the guy who sold it to me (and fraudulently sold it twice). Property rights can be chopped up and distributed in different ways. I might own a house but rent it to a tenant and have a mortgage on it from the bank, each of whom will have certain rights over that land, despite me being the owner. And property can apply to tangible things (a painting, a car), intangible things (a checking account balance at the bank, a certificateless share of stock in a corporation, a domain name registered with ICANN), and all sorts of concepts in between (the right to use a particular mailbox in a post office, an easement to use a driveway over my neighbor’s land, the right to use my name and image in a commercial, a futures contract that entitles me to take delivery of a whole bunch of wheat on a particular day at a particular time in the future). All of those are property, and recognized as property rights in U.S. law. What copyright actually is, is a temporary monopoly granted at the whim of Congress. It’s a license, not a right. Licenses are a right to do something. In fact, copyright owners assign licenses to others to use that intellectual property all the time. And the copyright itself is not property over an idea. It’s the right to copy something specific that has already been fixed in a particular physical medium. If you come up with an idea for a melody, you don’t own the copyright until you write it down. You’re just pretty far off base because you don’t understand how broad the word “property” is, and you don’t seem to want to examine just how man-made other forms of property are, and think that copyright is something special and different.

I have no idea why a thin client would need a 2.5Gbps NIC. I know bandwidth isn’t latency but for a thin client having a rock solid network connection to the virtual desktop server is pretty important for the user interface. I’m guessing pushing video and animations can require pretty high data rates, too.

Yup. LTE can support something like 300-400 connections per band and there are 16 primary bands licensed in the US. 5G and mm wave open things up some more, including beam forming techniques that may allow an antenna array to communicate with two devices on the same frequency at the same time. But at the same time, each carrier only gets some of those bands, and they want to separate bands by physical space so that neighboring cells are using different bands, and in 3 dimensional space there can be a lot of neighbors. And 300 passive connections simply keeping the connection alive are different from 300 active users trying to actively transmit and receive significant data. Plus real world interference will always make devices come up short from the theoretical max performance. Temporary/mobile towers go a long way, though, for temporary surges in demand, like sporting events. Things have gotten a lot better on game days in certain places (especially small college towns whose populations basically double on game day, with everyone jammed into a single stadium for about 4 hours).

It’s a fancy marketing term for when AI confidently does something in error. How can the AI be confident? We anthropomorphize the behaviors of these technologies to analogize their outputs to other phenomena observed in humans. In many cases, the analogy helps people decide how to respond to the technology itself, and that class of error. Describing things in terms of “hallucinations” tell users that the output shouldn’t always be trusted, regardless of how “confident” the technology seems.

Writing 360 TB at 4 MB/s will take over 1000 days, almost 3 years. Retrieving 360 TB at a rate of 30 MB/s is about 138 days. That capacity to bitrate ratio that is going to be really hard to use in a practical way, and it’ll be critical to get that speed up. Their target of 500 MB/s is still more than 8 days to read or write the data from one storage platter.

I would argue, and I’m sure many historians and librarians and archivists would agree, that “general data backups” are essential human data. Storing the data allows for later analysis, which may provide important insights. Even things that seem trivial and unimportant today can provide very important insights later.

Honda won’t honor my 10-year powertrain warranty just because I yeeted my 2-year-old Civic off a bridge into salt water!

That’s why Research in Motion (the developer of the Blackberry) had to buy the domain “rim.jobs” when the .jobs tld was launched.

I don’t think it’d be that simple. Any given website URL could go viral at any moment. In the old days, that might look like a DDoS that brings down the site (aka the slashdot effect or hug of death), but these days many small sites are hosted on infrastructure that is protected against unexpectedly high traffic. So if someone hosts deceptive content on their server and it can be viewed by billions, there would be a disconnect between a website’s reach and its accountability (to paraphrase Spider-Man’s Uncle Ben).

The company describes this generator as a solid state device, but the diagrams show the reliance on fluid/flow of hydrogen between the hot side and the cold side for moving some protons around. That seems to be something in between the semiconductor-based solid state thermoelectric generators that are already commonly understood and some kind of generator with moving solid parts. It still seems like a low maintenance solution to have a closed loop of hydrogen, but that seems like a potential maintenance/failure point, as well, to rely on the chamber to remain filled with hydrogen gas.

The inventor/founder at the center of the article, Lonnie Johnson, was on the team at JPL that designed and implemented the thermoelectric generators (heated by radioactive decay from plutonium-238 pellets) on the Galileo spacecraft sent to Jupiter. So I would expect that he’s more familiar with the thermodynamic and engineering challenges than even a typical expert. The PR fluff put out by the company mentions that the theoretical basis for this specific type of generator was worked out a while ago but needed materials science to advance to the point where this type of generator can be thermodynamically and commercially feasible. Looking at how this generator is supposed to work, it’s interesting in that it does rely on the movement of fluid, but is supposed to be a totally closed loop, to be a bit different than the pure solid state, semiconductor-based Seebeck generators that are already well known. The other area talked about in this article is that they believe that it can be effective with lower temperature differentials than any previous technology, which might make a huge difference in whether it can be deployed to more useful places and thereby make it economically feasible more easily than prior concepts. In the end, if these generators can output some electric voltage/current, it might just take on similar generation characteristics as photovoltaics, which could mean that hooking these up to the grid could draw on some of the lessons learned from the rise of grid scale solar.

Financial analysts were sounding the alarm in October. On October 7, Bloomberg ran an influential article about the circular deals: bloomberg.com/…/openai-s-nvidia-amd-deals-boost-1… That built on earlier reporting where they described the deals as circular, as the deals were being announced. Each of these reports notes the financial analysts at different investment firms sounding the alarm. From there, a robust discussion happened all over the financial press about whether these circular deals were truly unstable. By the time Gamers Nexus ran that video the financial press was already kinda getting sick of the story. Whatever the hell these trading algorithms were doing on November 20, they definitely weren’t ahead of the curve on investor knowledge and belief.

it’s like AI companies went from buying 10 memories as usual to 1000000, I mean, they basically did. OpenAI announced a few months ago that they reached deals to buy 900,000 wafers of DRAM per month, representing 40% of global production capacity. For a single company. There are several other companies competing at those scales, too.