The model is too big for casual use (IE running inference at home). LLama can just barely fit on either one 48 GB GPU, or two 24 GB GPUs. The number of parameters in Grok is much, much greater, really placing it outside the realistic realm of those wanting to run inference on their own hardware in a more hobbyist fashion. Sure, you can always run it on a normal CPU with a ton of RAM, but you'd be looking at multiple seconds per token.
In manufacturing, everything little thing is documented
I'm curious if it is a normal part of production to remove and reinstall the door plug. It sounds like this was some kind of exception during the manufacturing that it needed to be removed after it was already installed. That could have been done by someone not qualified to work on that specific aspect of the plane, and not properly documented as well since it had already been assembled.
Ironically this could be related to quality control and inspection, where an inspector had the door removed for some reason, but then it was not correctly reinstalled.
Yes, exactly, this is a very slippery slope. Then you have presidents banning foreign platforms because they have political leanings that hinder their party or chances at being reelected, etc.
Global capital markets do operate on a set of established rules and norms
This is extremely common and typically handled via tariffs. If China won't allow the USA to sell Product A in their country, then the US slaps a tariff on Product B that China sells in the US. The issue here is that with these kinds of online services, the user is the commodity being monetized, so there really isn't much of anything that a tariff can be placed on with TikTok that is direct enough.
Government just hasn't adapted fast enough to figure out how to regulate these newer technologies, so their default action is to ban it entirely.
The point is that this is done with every other type of commodity that exists (IE no one is allowed to import wool into our country, to protect our local production), so it seems reasonable to do it with online services as well.
What would you do if your local currency was down 250% (vs USD) over the last year?
Um, exactly what they've done? Blame outsiders and don't take any internal responsibility for your country's issues?
The Toronto Sun is part of Postmedia, an American-owned chain that exercises strong editorial control and is basically Fox North.
Canadian Civil Liberties Association seems legit, but the Canadian Constitution Foundation is part of the Koch Atlas network, slipping far-right American money into Canadian politics.
That doesn't mean that they're wrong, but what they say should be examined carefully for fish hooks.
LLMs will make this sort of thing much, much easier. Use something like Google Translate to translate from Russian to English. Then run that output through ChatGPT asking it to format it like a well written major news outlet story, and there you go. That will clean up all the typical translation and grammatical errors that make it easy to spot these kinds of fake sites.
I can see at least one issue in this particular case. Some messaging platforms, like FB Messenger, will default the Send button to an emoji if no text has been typed into the outgoing message box. If it's someone you're in a relationship with (as determined by your FB status) it is a kiss. Otherwise it's typically a thumbs up. So what happens is when sending a message, if you inadvertently press send twice, the second time results in the emoji being sent.
When you type a word "Okay" or some other confirmation, it's pretty clear you didn't send it by accident. The UI for emojis is complex and easy to make mistakes or inadvertently send something you didn't intend to. So there may be a legal argument on those grounds still, which is an entirely different argument than the meaning of whatever emoji that was sent.
This is just a superficial fix to underlying problems in LLMs. If you have to engineer your question such a very specific manner that the AI produces a correct answer, then you're merely adding another layer of crap on top to try and patch your system to work correctly.
Imagine a company produces a math processor chip that always gets the operation 4 + 4 wrong and produces 7. So to fix that the UI on top converts 4 + 4 into 4 + 4 + 1 before feeding it to the processor. That's pretty much what prompt engineering is doing but in a vastly more complex, nebulous and often biased way.
Writable, which is billed as a time-saving tool for teachers, was purchased last month by education giant Houghton Mifflin Harcourt, whose materials are used in 90% of K-12 schools. Teachers use it to run students' essays through ChatGPT, then evaluate the AI-generated feedback and return it to the students.
A teacher gives the class a writing assignment -- say, "What I did over my summer vacation" -- and the students send in their work electronically. The teacher submits the essays to Writable, which in turn runs them through ChatGPT. ChatGPT offers comments and observations to the teacher, who is supposed to review and tweak them before sending the feedback to the students. Writable "tokenizes" students' information so that no personally identifying details are submitted to the AI program.
Taking into account the clarity, specificity, organization, engagement, and terminology of the text, I would assign it a grade of B.
The text effectively communicates the purpose of Writable as a time-saving tool for teachers and provides insight into its functionality with ChatGPT. It offers clear examples of how teachers use the tool and briefly mentions the privacy measures taken by Writable. However, there are areas where the text could be improved for clarity and engagement, such as providing more specific examples and explanations, smoothing transitions between points, and offering quotes or testimonials for added context. Additionally, the explanation of tokenization could be clearer for readers unfamiliar with the term. Overall, while the text effectively conveys its message, there is room for enhancement to elevate its quality and readability to a higher level.
The FAA and friends need to quit crippling SpaceX with never ending red tape. You'd better believe the Chinese space programs do not have this kind of drag slowing down their development cadence.
To clarify, Morse code is still alive and well in the Amateur Radio community (although it is no longer required for any license levels at all now - even the "Extra" highest class license doesn't require it). It is still used regularly by hams, and there are also Morse code beacons that transmit automatically for various propagation purposes.
What this article refers to is the last commercial Morse code radio station.
Our company has some business presence in India, and we have found that it is part of the standard business ethics / practice to basically try to get away with whatever you can. Fees, expenses, agreements, etc, are never perceived to be set in stone. Whatever one can get by with is considered acceptable and moral from a business standpoint, and when "caught" breaking the terms of the agreement it is expected there is no harm / no foul and they simply will pay up since they must.
Wikipedia is exactly as reliable as its sources.
Wikipedia is, at best, as reliable as its sources
FTFY
We're not talking about layers of stone or ice here, but muddy sediment at the bottom of a lake. I'm sure there are many other things that have encroached down into that sediment as well, especially various chemicals, radioactive isotopes, etc. So they simply cannot use that particular type of layering as an indicator for this particular kind of pollution.
"Experience has proved that some people indeed know everything." -- Russell Baker
