If you ask generic questions, it will give generic answers. The key is to include lots of relevant information and context in your questions. Yes a person might be able to provide more specific advice relating to your situation, but a lot of that is just that they have more context about what you're asking
Do LLMs actually store and reproduce facts? The longer I use LLMs, the lesser i trust facts as complex as "capital of australia" to be answered, truthfully.
LLMs could be described as storing and reproducing beliefs, but they're definitely not databases that can be easily maintained as things change or verified for perfect accuracy at any given moment.
I think the best way to use them is to have a minimal language model that is as small as possible while still being able to comprehend language; and that this then goes off to an actual knowledge base of some kind where all the factoids can be checked separately.
Humans have separate explicit/declarative memory and implicit/procedural memory memories. I think the Transformer architecture puts everything into what is really only suitable for implicit/procedural; I think RAG is trying to be a separate explicit/declarative memory system, but I'm always to busy to study this in more than a superficial level, so I'm not sure.
I think a lot of this was done with the hope that more immigrants would translate to higher economic productivity, but this simply hasn't been the case. Unemployment rate is 1-2% higher than in the US. GDP per capital is stagnating. Most of the highest skilled workers leave for the US. I don't think the problem is the immigrants per se, but rather our government's addiction to deficit spending and printing money. Not to mention the tacit refusal to enforce anti-trust legislation against the major companies in the country which seems totally corrupt. The number of public sector jobs has ballooned and none of the public services have improved--quite the opposite. Canada needs some major reforms to get out of this mess.
To their credit there has been a movement across the country to scrap single-family zoning restrictions and allow more housing development. Which should improve the housing situation but it's still unlikely the housing market can keep up
> rather our government's addiction to deficit spending and printing money
The deficit spending is clearly to try to buy votes but I think he reading the tea leaves wrong - the average person is way more concerned about inflation, their job, and the economy, and I think this spending is unpopular, even from those who tangentially benefit from the spending!
Capital doesn't place too much emphasis on borders. When a person can double or triple their salary just by moving a few hours south it's easy for all the capital to get concentrated in one place (namely the US)
"Highly skilled" people are highly sought after because, by and large, they are the people most likely to create new capital. Success in creating new capital begets a desire to create even newer capital and so long as the people continue to be successful in capital creation efforts, there is really no end.
Not a whole lot different as to why Silicon Valley hasn't stabilized with Detroit. Of course, Detroit is also interesting here because it provides a cautionary tale as to what happens when capital creation seizes up. That is what Canada is desperately trying to avoid by trying to bring in the "highest skilled" people from other countries.
the important caveat here is the TN visa, which makes it reasonably easy for certain skilled fields to move across the border.
if you're in STEM there is easy a 40% difference in pay, and considerably more on the high end (FAANGM, or similar large Enterprise tier).
by limiting it to high-end, educated work, it also skews the market a lot -- high earners leave, low-to-mid market stagnates. Canada then has to import both high-end talent via immigration, but also low-end Temporary Foreign Workers to do crappy jobs like slinging coffee at Tim Hortons.
Capitalism is not supposed to increase salaries forever. Certain regions are more developed in certain industries and can demand higher salaries in those industries. Thus people move where the jobs are. Brain drain is perfectly consistent with capitalism and an expected affect.
As far as I can tell the life imprisonment is for people advocating genocide. It is not clear to me that this just allows any so-called hate crime to be punished with life in prison, which seems pretty authoritarian. But the thing is it's pretty difficult to understand just what exactly the changes being proposed are. From a cursory reading of the bill it is not obvious what exactly the implications are here.
As far as how likely it is to pass, I would say not impossible. The liberals have an outsize influence on how the country is run and will until next year at least, unless Trudeau calls an early election (which by all accounts he would lose right now).
No, if you search through it it adds 'imprisonment for life' to both the 'advocating genocide' offence and also the 'Offence motivated by hatred' under the act (presumably meaning online speech) section separately.
It's seems to leave a lot of room for interpretation. Is _any_ hate crime punishable by life imprisonment? I would hope no judges in this country would agree with such an interpretation, but you never know
The problem with Canadian airlines (and Canadian companies in general), is that if you don't like the service there aren't really any meaningful alternatives. The two largest carriers have a near total stranglehold on the market (+75%) of flights. They practically have zero incentive to improve because they are propped up by protectionist Canadian policies that prevent foreign companies coming in with any meaningful competition. Their main competitor WestJet is similarly bad. Story of the Canadian economy really
Does reading have to be "productive" in order to be worthwhile? I would argue that reading is an enjoyable activity regardless of whether there are any tangible benefits. Similar to how it's enjoyable to go on a meandering walk. Even though there are more efficient forms of cardio and you could get to where your going faster by driving, that doesn't keep the activity from being enjoyable in and of itself.
I've probably forgotten 95% of everything I've ever read. But I certainly don't feel like all that time was wasted.
As a dyslexic person, I can assure you that reading is not a universally enjoyable activity. You should enjoy your gifts, though. I do enjoy a good walk.
Anecdotally I've heard the people living downwind of the big oilsands developments suffer a wide range of health problems from the toxic air quality. It's no secret that these developments are not good for human or animal health, although the big oil companies don't advertise it. I guess the upshot is that to even get to Fort McMurray you need to drive for hours through pure wilderness--no towns, no farms, nothing but endless forest. It's far out of the way of where most people would choose to live (were it not for high-paying oil jobs).
But even in the south of the province it feels like the air quality has been deteriorating. We seem to get air pollution warnings are long more than I ever remember growing up.
I wouldn’t just assume that. Yes fire smoke isn’t good, but different types of smoke can be exponentially worse depending on the chemicals of what’s burning. Just look at all the firefighters who got permanent lung damage for example
I was in Austin a couple months ago. If that city is dead, there are few cities in the US that could be considered alive. Honestly this whole business of proclaiming cities "dead" is a bit absurd. Despite all the predictions of doom SF has remained vibrant and it's clearly not going anywhere, nor does it need AI to save it. Neither are Austin or Miami. Cities can succeed together, it's not winner takes all.
One thing to note when making comparisons like this is that LLM output is not deterministic, in the sense that if you ask it the same question 10 times you will get 10 different answers. So the question to ask is not, “is GPT4 better on this one specific question?”, but rather “does GPT4 produce better results on average?”. I would bet that it does, for no other reason than that it is much larger, and LLM performance seems to just scale with size. Also worth noting is that the more detailed your prompt is the better the response will be. Sometimes you have to encourage GPT to get the best results. GPT4 should be able yo handle much more complex and detailed prompts than 3.5
An unintuitive consequence of this nondeterminism over millions of interactions is that different individuals will see different trends. IME the quality of response is accurately modeled by "luck", and people's luck can change.
So we have different population of GPT users.
An average experience might be to get a mixture of spot-on helpful responses and obvious bullshit^H^H^Hallucinations, this population might learn what questions to ask given the limitations of the model. This is really a best case scenario as people can actually get a feel for how to use the technology, strengths and weaknesses etc.
Personally my experience was the first few dozen times I used it I was amazed at the responses, I was on team superintelligence, anyone who is getting lackluster responses is just holding it wrong. But luck changes and over months of use I see now that on average the responses are just OK. But this is the case that leads to disappointment and bitter conspiracy (the superintelligence is being suppressed, give it back!)
Another population had rotten luck to begin with, and got dumb unhelpful response over and over. This population quickly determined that the AI was all hype and stopped exploring (you don't keep going back to the casino if you lose everything your first time...).
This divergence is destructive to the larger discourse, since we have fanboys flummoxed by naysayers and critics bamboozled by hype beasts.
Interestingly enough, I don't think this applies to the APIs as much.
What I've seen on indie hacker type website is that developers are fully on this train and not very critical of the outputs.
This is why you get very basic prompts sent by "wrapper apps", which might have given the developer a good result the only time it was tested before being put in production.
I think it might take a while before tools show up that can generate 100 test cases and test a given prompt with all 100 to report on the results... It seems to be a tough problem to crack.
IMHO front-end chat end-users have many many more "at-bats" and get to see more model results than devs do, which make them more critical of those results.
I’ve seen this stated a ton and it’s not really true. Once trained, the model (except for decoding) is deterministic, and you can enforce determinism fairly easily. ChatGPT is not deterministic at the chat window but that’s not inherent to the model.
I imagine both actual population and population density have something to do with it as well. The difference is massive, which probably also means a lot more green areas and a lot less concrete and metal in proportion.
a. Northern Hemisphere: The half that lies north of the Equator. This hemisphere contains approximately 68% of Earth's landmass and is home to about 90% of the global population. It includes North America, Europe, Asia, and most of Africa.
b. Southern Hemisphere: The half that lies south of the Equator. It contains only 32% of Earth's landmass and is home to about 10% of the global population. It includes South America, Australia, Antarctica, and the southern parts of Africa.
If you want a more serious answer, tectonic plates are the floating cooled solidified 'slag' on a spinning spherical molten furnace (that's been cooling for 4 billion years).
Distance from spin axis (equator Vs poles) plays more of a dynamic than "South" vs "North" (air quoted as that's merely a convention as to which hemisphere is the 'top' hemisphere).
Two major points are that:
* N vs S looks a bit different on a globe or non Mercator projection (chosen by European traders to maximise the parts of the world of interest to them).
* current positions are just that - they've steadily moved since formation of the planet.
North and South should not be in air-quotes. They're correct, by definition.
"Top" and "bottom", however, can be air-quoted. There's nothing that says north equals "top"; north and south are arbitrary, and are just terms chosen to designate the two poles. The only reason people think north=top is because most maps are oriented that way.
If you flip North can change from "top" to "bottom" or from "bottom" to "top" depending on your map. But it doesn't stop being North, which is a fairly well defined concept. Hence no air quotes.
> But it doesn't stop being North, which is a fairly well defined concept.
Can you cite the definition please?
My understanding, from 40 years of geodesy and cartography, was that it was a fairly recent (post 16th Century) convention and many many significant older maps orientate in quite different ways.
Huh? This doesn't make sense. Lots of modern maps even have north pointing somewhere other than up, though they use a marker to indicate which way is north.
It doesn't matter which way you point a piece of paper: the North Pole is always going to be the North Pole, and the South Pole is always going to be in Antarctica (until the continent drifts, though the pole will still be there). The names of these poles may have been originally chosen centuries ago because popular maps pointed that way, but now those names are fixed by definition.
So you link to Wikipedia. OK. You see that it defines North pole in terms of Northern hemisphere. Now if you follow the links through you will see that Wikipedia defines North as "where the up on the map is". Qed.
If I'm standing on the Moon, looking at the Earth, I can see which side is North because of the arrangement of the continents. I don't need to have a paper map with me. The whole thing doesn't make sense.
Finally, you're just plain wrong, and you're reading Wikipedia wrongly. According to the page for "North",
>The word north is related to the Old High German nord, both descending from the Proto-Indo-European unit ner-, meaning "left; below" as north is to left when facing the rising sun. Similarly, the other cardinal directions are also related to the sun's position.
So the definition is derived from the Sun's position, not some stupid paper map.
You're probably reading this part:
>By convention, the top or upward-facing side of a map is north.
That just says that maps are normally oriented with North pointing up. It doesn't say that "north" is defined* that way.
First you define North in terms of arrangements of continents (saying north is where northern continents are, cool) and then you use English etymology as definition? Lol.
> You're probably reading this part
Probably? I literally quoted it.
You claim I misinterpreted that phrase but you don't give a better option. All definitions of north are by convention or relative to space (so if it flips then "northern arrangement" of continents you see from the moon will be southern)
"North" IS by convention. I've been saying this all along. It has nothing to do with any map. The place we call "the north pole" is north because it's been defined that way for ages. Apparently the origins of the word have to do with Sun, but the effect is the same, because it points to the same place on this spinning ball.
Ah but that's my point, turns out north is not a well-defined concept contrary to what I thought first.
That quote is ambiguous and could be interpreted as "north is where the up on the map is". And actually there seems to be no better definition (that eg. would stand the flip of axis). What are we arguing about again?
North IS a well-defined concept. From the most ancient times, people knew where north was: it's 90 degrees left of east, and east is where the Sun rises. Go back in time 5000 years and ask anyone where North is, and they'll point you to it, long before any modern maps of the world were ever made.
Wikipedia having one poorly-written line about it doesn't change tens of thousands of years of human history and knowledge about where north is. Wikipedia isn't even an authoritative reference on anything.
I think it's pretty much a myth that a projection was chosen to accentuate certain countries, the simpler and less nefarious reason is that they're easier to use for navigation.
The land masses are floating plates pushing and shoving against each other over a bubling mass of heat energy pushing up to escape outwards, there are more forces at play than simple centrifuging.
The linked animation shows a relatively short sequence of movement, over the last few million years, that is likely the most we'll be able to reconstruct thanks to the arrow of time.
BTW, looking at northern hemisphere vs southern hemisphere doesn't show just how unevenly it is distributed. To really see how unevenly land is distributed go to Google Maps (or any other online map that switches to a 3D globe view when you zoom out and search for Bora Bora in French Polynesia.
Then zoom out to the globe view. Then tweak the position to get North America to just disappear over the horizon, along with most of South America.
With sufficient tweaking you can get it down to the half of the Earth you are looking at just being ocean, Australia, Papua New Guinea, Tasmania, about 2/3 of Antartica, and parts of Argentina, Chile, Indonesia, and Philippines, plus a handful of islands.
You're looking at a snapshot of a process that goes on for billions of years... there's no particular reason why any snapshot would look like any sort of way.
