A very good point. For anyone not familiar with anterograde amnesia, the classical case is patient H.M. (https://en.wikipedia.org/wiki/Henry_Molaison), whose condition was researched by Brenda Milner.
> Near the end of his life, Molaison regularly filled in crossword puzzles.[16] He was able to fill in answers to clues that referred to pre-1953 knowledge. As for post-1953 information, he was able to modify old memories with new informations. For instance, he could add a memory about Jonas Salk by modifying his memory of polio.[2]
The nature of memory is so cool, the idea that there are completely different systems governing the creation of wholesale "new" memories and the modification of existing concepts is fascinating to me because those things really do "feel" different in a qualitative sense, but having evidence that you're physically doing something different in those cases is really cool.
I thought maybe people would be curious to read about how we came to understand the condition and the history behind it, as well as any associated information. Forgive me for such a deep transgression as this assumption.
All that needed to be conveyed was that there are humans who cannot create new memories. That is enough to pose the philosophical question about these models having intelligence. Anything more is just adding an anecdote that isn't necessary.
lol, as if pointing at a wikipedia article (without any relevant discussion of the contents therein) is some kind of conversational excellence.
Or perhaps you were referring to the impact of the two in that the "sledgehammer" of "they can't make new memories" is a lot more effective than the tiny scalpel of "if you do a wikipedia search this is a single one of the relevant articles"
The extra information is that he is the canonical case which defined our clinical understanding of the condition. Not just a "single relevant article."
I pulled it up because I was familiar with this fact.
Nobody else in the thread is making an argument that relies on the distinction.
"Intelligence" is used most commonly to refer to a class or collection of cognitive abilities. I don't think there is a consensus on an exact collection or specific class that the word covers, even if you consider specific scientific domains.
LLMs have honestly been a fun way to explore that. They obviously have a "kind" of intelligence, namely pattern recall. Wrap them in an agent and you get another kind: pattern composition. Those kinds of intelligences have been applied to mathematics for decades, but LLMs have allowed use to apply them to a semantic text domain.
I wonder if you could wrap image diffusion models in an agent set up the same way and get some new ability as well.
The problem I see regarding LLMs is they are the extreme edge of what humans have created. They are trained on the outputs of intelligence and thought and its representation in language is this like parallel stream to intelligence that has pointers back to the underlying machine and semantics. The fact that LLMs are able to take that output and reverse engineer something that mimics the underlying machine that created that output is fascinating. But you can still see this machinery for what it is.
LLMs falls apart on really simple reasoning tasks because when there is no statistical mapping to a problem in its network it has to generate a massive amount of tokens to maybe find the right statistical match to this new concept. It is so slow. It is not something you or I would recognize as a process of logical reasoning. It is more like statistically brute forcing reason by way of its statistical echo.
So, I guess pattern recall is the right words. Or statistical pattern matching. Recall works if you view a trained model as memories, which is how I often model what they store in my own mind. So, it is... something. Maybe intelligence. Maybe just a really convincing simulation of the outputs of intelligence. Is there a difference? Fundamentally I think so.
There may be some demographic groups located between people who were young during the 1980s and people who are young during the 2020s, time periods which are 40 years apart.
The behind the scenes on deciding when to release these models has got to be pretty insanely stressful if they're coming out within 30 minutes-ish of each other.
I wonder if their "5.3" was continuously being updated, with regenerated benchmarks with each improvement, and they just stayed ready to release it when claude released
This seems plausible. It would be shocking if these companies didn't have an automated testing suite which is recomputing these benchmarks on a regular basis, and uploading to a dashboard for supervision.
Given that they already pre-approved various language and marketing materials beforehand there's no real reason they couldn't just leave it lined up with a function call to go live once the key players make the call.
Could be, could also be situations where things are lined up to launch in the near future and then a mad dash happens upon receiving outside news of another launch happening.
I suppose coincidences happen too but that just seems too unlikely to believe honestly. Some sort of knowledge leakage does seem like the most likely reason.
Heritability is a pretty funky concept because it's contextual to a certain point in time, environment, and population, effectively.
An example I like is that if you measured the heritability of depression in 2015, and then you measured the heritability of depression in 2021, you would likely see changes due to environmental effects (namely, there's the pandemic/lockdowns and this could conceivably cause more people to experience depressive symptoms). Let's assume we make those measurements and the rate of depression did increase, and we could tie it causally to the pandemic or related events.
In that scenario, the heritability of depression would have decreased. I don't think anyone would argue there were massive genetic changes in that 6 year time period on a population scale, but the environment changed in a way that affected the population as a whole, so the proportion of the effect on the trait which is genetically explained decreased.
For something like lifespan in the above example, you can imagine that in a period of wartime, famine, or widespread disease the heritability would also decrease in many scenarios (if random chance is ending a lot of lives early, how long the tail of lifespan is influenced genetically is much less important).
Given that note, it's generally tricky to talk about whether heritability increases or decreases, but with more accurate estimates of how genetic predispositions form you could see the heritability of certain traits increase with the environment held stable, as there's certainly ones that may be underestimated or genetic factors that aren't currently accounted for in many traits.
*edit: I realized I never mentioned the other thing I wanted to mention writing this! since you mentioned what the percent heritability means here, I think the best way to think of it is just "the proportion of phenotypic variation for this trait in a measured population which is explained by genetic variation." So it's dependent on the amount of variation in several aspects (environmental, genetic, phenotypic).
Some epigenetic effects are semi-heritable too, eg maternal exposure can be transmitted. That's in addition to environmental effects like you mentioned. Two otherwise identical cohorts can inherit the same genetic predisposition for depression where one manifests and another does not entirely due to their circumstances.
This is wild, what were the effects like for you? I imagine your eyes and hands would start to see physical effects from that level of use for such a consistent time.
What sort of social changes did you notice after that period of time?
I've never used TikTok, but the techniques they employ sounds seriously addictive.
The main one is a deep sense of defeat. The app would keep me there longer than I want. I would waste money ordering doordash, or lose sleep, or get drunk, or all of the above. Each time, I'd feel like I set myself back a little too far. I'd try to ignore the feeling, but wouldn't know what to do. What's the default action? Keep scrolling. And of course I'd just keep missing more such personal deadlines. Then I'd feel more defeated, and keep scrolling. It just spirals further and further down, far past rock bottom. Maybe it's what advancing the Kola Superdeep Borehole felt like.
Thank you for the reply. I get that feeling, I've felt it in smaller doses in other contexts. There's a sense of resignation there where you keep doing something that you know is hurting you when you feel like you don't know another option and just can't muster the fortitude to escape that cycle of guilt.
For the US, it's not too out of the norm historically speaking. Up until relatively recently tariffs were very popular in the US despite the clear understanding by academics that they were incredibly damaging to the economy. Political movements based around protectionist economic arguments have a long history in the US.
For an example, take a look at the 1888 US presidential election, which largely revolved around tariffs. Grover Cleveland lost re-election due to being part of the pro-business wing of the Democrats, and he came to the conclusion that tariffs were a negative to the economy overall, while his opponents were strongly protectionist. After McKinley's Republicans won the election on a protectionist platform, he instituted the McKinley tariffs (average import duties of around ~50%), which were devastating to the economy despite being extremely popular with the nation in the election. It led to massive price increases which led to the re-election of Grover Cleveland in 1892 (only other non-consecutive term president aside from Trump). Despite expert opinion being fairly solidified against tariffs even at the time, the idea of "protecting American business" and "punishing other countries for their unequal trade deficits with the US" was pretty popular with specific interest groups!
Parts of this sound rather familiar, do they not? I would then argue that it points to a cultural element, out of the two options of a failure of education or a culture of rejecting it. History certainly rhymes on this point.
Serious question, if tariffs are so terrible, why did so many countries have tariffs on US (and other country's) goods?
Also, during the period you describe the US was a major export economy. Now the US economy is far more insular (even before Trump) than it was during that period (foreign trade was more than 50% in the late 19th century vs 7% today). Why would you assume that doesn't impact the effects of tariffs?
2. Quick and easy reaction to other nation's tariffs, which we saw this year when Trump announced all his tariffs.
3. Targeted at specific industries to influence politics in other nations. IIRC, the EU is actually doing this to the US, specific states that have a lot of support for Trump, in the hope those voters will make the connection and get Trump to back off.
4. Targeted at specific industries to protect domestic industries from being undermined. The USA has accused China of this in various cases, any "anti-dumping tariffs" would be perfectly reasonable where this happens. China was accused of trying basically the same thing Uber was accused of, spending (VC|tax) money to corner the market then raising prices when all the old (taxi drivers|PV makers) were gone.
5. Sin taxes. Singapore doesn't have their own car industry to protect, they make it really difficult to get a car just because they don't want lots of cars. I mean, it's more of a registration fee, but the effect there is much the same given the lack of local industry.
There's definitely debate over which specific scenarios tariffs could be beneficial, as I understand it, but the general theme is that any benefits are highly concentrated (one or two companies or industries will benefit) and the negatives are felt in a diffuse fashion(every consumer pays the tax). They are broadly protectionist and the ones that do exist usually are implemented for pretty specific reasons like the following:
1: A government wants to protect domestic industries over ones outside of the country by applying price increases to the foreign ones, with the idea being that the domestic industries just need to grow into being able to compete with the industries in other areas. This is called the infant industries argument. A central problem with this is that the industries will always benefit from the protectionist policy, and are unlikely to ever admit that they have "grown up" so to say. My general view on this is that groups will of course lobby to have benefits specific to their industry, and that there are probably scenarios where we would prefer to have things handled domestically rather than abroad, but I would generally want this to be highly targeted and time-gated(Once the industries are mature enough to compete, we wouldn't want to keep subsidizing them), and that other tools are probably more efficient for this purpose.
2: Some sort of national security argument, where production being cut off during war would be a serious concern. My general thought on this one is that if something is specifically important for national security, broad reaching taxes on all imports probably aren't as useful as targeted government interventions in those specific industries. The government can build whatever factories it wants or contract people to do specific things if it passes a law to do so. If we're worried that we need a domestic supply of beets(randomly selected example) and the government is willing to produce them at a loss for national security reasons, they should probably just do that rather than tax imports of coffee, chocolate, bananas, beets, beef, and cars in order to encourage domestic production of beets. The broad spectrum nature of across the board tariffs doesn't specifically protect any given industry, unless the specific protection desired is "nothing should be imported, only ever produced domestically."
3: Historically speaking tariffs were a major source of government revenue. There was no income tax in the very early US (and this was the case in many places), and tariffs were seen as an efficient way to raise a lot of money for the government. At the time it was also something that was a lot easier to measure than things like property value, sales, or individual income, because all the goods had to come in through a port. Pretty easy to check the majority of the things coming in, compared to other taxation methods. A major argument in the time period was actually that the government was making too much revenue, such that it was constricting the growth of the private economy! A huge debate in the 1880s and 1890s was on how the share of government revenue could be lowered, and the growth of the economy could be encouraged. Republicans argued that implementing more tariffs would actually reduce imports and lead to lower revenues, which was the stated goal of the McKinley tariffs.
The general reason some people oppose tariffs overall is that they represent an approach to economic growth based on zero-sum thinking, i.e. an idea that if another country experiences economic growth, ours must suffer economic decline. There tends to be more support from many people behind the idea that international trade allows multiple economies to grow in tandem, as I understand it, but I'm definitely not an expert in this stuff, haha. I just find the historical aspect interesting.
On your second point, describing it as a major export economy in the period I describe is maybe not capturing the scenario, because we were in the middle of a major change in manufacturing. We were major importers of manufactured goods in the preceding time period, and we exported agricultural goods! The period from 1890 to 1910 roughly(depending on when you draw the cutoff) is when the US mainly started exporting manufactured goods more than importing them, and it was a massive transition. The period we're talking about is probably best understood as when we were in the process of industrializing more.
It's fair to point out that the economy was pretty different at the time, but it was different in a bigger way.
If advertising is no longer financially rewarding, is there not an argument that labor could transition into a different sector of the economy?
Companies based around advertising would die, yes, but they only exist in the first place because of how lucrative the activity is. Nobody is sitting around dreaming of how they could sell ads better than anyone else while not thinking of the financial compensation. At least I hope they aren't.
If someone was saying "many people have jobs in running offshore internet sports betting companies, if we put regulations on offshore internet sports betting, it would remove jobs" wouldn't the natural question be whether those industries are actually productive to have people employed in, or if it's a harmful industry overall? Generally in my view its somewhat sad that the system as a whole optimizes for advertising work rather than orienting in a way that everyone could be putting their work towards something they see as more fulfilling.
There is certainly more need for product discoverability broadly than something like online gambling, but I think the more relevant conversation is if the current advertising model is more like a local minima preventing progress towards a more economically viable method of handling product discoverability.
Currently I think it is difficult to argue that advertising in its most visible forms have any serious benefit to people looking to obtain a service.
How often does an actual random advertisement shown on a billboard or a preroll youtube ad actually lead to a quality product? I think it is fairly common for people who are acquiring the best versions of things to do so primarily through research in forums or reviews, which is coming from the user looking from the product, rather than the product forcing itself into the mind of a given user to convince them to consume it.
reply