As stated in other comments, it would be actually easy to implement as it doesn't require a change in the constitution.
Anyway, as I told in another comment, I'm not from the US, so my intention wasn't really to say how to implement the proposal. I just had this idea and wrote it down in a blog post.
Yours is a very good point of view actually. I initially hadn't thought about the tragedy of the commons situation where all the states agree to implement it and later one state decides to go back to winner take all system.
Ok yes maybe "easy" isn't exactly the correct term, but the way to appoint electors in each state doesn't need a 2/3 majority in every house + 3/4 of the states as it happens with Constitutional amendaments. So it's definitely an easier approach
- California has 38,940,231 inhabitants and 54 electors. That means 1 elector every 721,115 inhabitants
- Oklahoma has 4,053,824 inhabitants and 7 electors, so 1 elector every 579,117 people.
Yeah, spoiler vote brings to 2 parties (like the US). Although the current electoral college doesn't work well, I'm not sure the Americans would prefer to remove it completely and rather use directly popular vote because in the rural areas they don't want to be "governed" by the decisions of big cities.
What do you think about keeping the electoral college and use IRV instead?
> Although the current electoral college doesn't work well, I'm not sure the Americans would prefer to remove it completely
I think the electoral college is working as intended to balance out power between larger more powerful states and less powerful states. Though I also don’t believe the popular vote is necessarily ideal from the perspective of maintaining more diversity of thoughts and opinions which is useful for a healthy republic.
As someone who’s lived in mostly rural states I’m glad because it prevents places like California with large populations with pretty homogeneous opinions and world views to dominate the nations elections. I get folks in those larger states would also feel equally left out as well.
I view the electoral college as a means of accounting for the “entropy” of votes among states. There’s more diversity of thought and opinion between say a rural Wyoming farmer and a worker in Hawaii than there is among most Californians living in say the 50 square miles of Hollywood.
In a fashion it’s similar to how network effects dominate large in markets where companies lucky to get a market first or to grow first get an unfair network advantage. Antitrust and tax laws ideally gently balance out this winner-takes-effects to provide more diverse and robust markets. Most hyper successful companies aren’t necessarily better at what they do and there’s a large amount of luck in success. Of course they still have to work hard and be effective to capitalize on the opportunities. Similar things happen among states in the USA.
One thing to consider is that the electoral college system also accounts for varying levels of economic density.
While I often hear “land doesn’t vote”, people don’t consider that a farm will never have the same economic output as an office building with the same footprint. This includes the population that comes with all that employment. There needs to be a way to represent the less economically dense sectors of the country because otherwise cities will dominate everything.
Yeah good point. Its sort of hidden in my argument, but economic diversity (including density) is part of the overall diversity and it's critical to "the system as a whole".
That office building might have more economic outpiut, but at the end of the day it's the farm that feeds people and mines that provide materials to build the offices. We can do without Meta or Google, but without farms people start starving and revolting pretty quickly.
Hello, here is my proposal to improve the Electoral College. The basic idea is to allocate electors proportionally in each state with the Jefferson method (aka d'Hondt method) rather than using a "winner-take-all" system.
What do you think about it?
PS: I'm not from the US, so an American perspective on that would be extremely appreciated :)
I think that it would be a way to get the US closer to popular vote aligning with the election result, and in a way that allows states to make the change themselves instead of requiring a constitutional amendment or somrthing. You'll never get exactly the same as the popular vote because you're rounding to integers, but it gets reliably close.
More interestingly, this scheme reduces party power inside states, so the incentive is for each individual state not to want to do it even though on the whole it's better, and instead the current status quo is the stable configuration, so getting states to want to do this is basically impossible. Think about it: if every state did this, then one state said "nope, we are winner take all again" that state could decide elections. So this is a system that is easy to implement by states, requires no US constitution amendments or anything like that, but works in such a way that no state would for fear others wouldn't. Interesting game theory here, it's very similar to a tragedy of the commons.
Alternatively, there's a proposed amendment to the US constitution called the equal apportionment amendment, that was passed 200 years ago but never ratified by the states, that changes the way the house of representatives is apportioned, such that among many other improvements, will change the way electors are apportioned in presidential elections. You don't need every state to ratify it because you only need 3/4ths of states to do so, many of which already have, it's binding on all so no worry about any one backing out, and you don't need congress to vote on it because they already did and voted yes centuries ago. It has other benefits too, like reducing the prevalence of 2 parties in the house and therefore elsewhere potentially, and increasing the fair distribution of representation in the house, which suffers from a similar problem as the electoral college.
I believe you're right, even if all the states agreed to make a similar change, this "equilibrium" state would be unstable. It would require an amendment to the US constitution to make it stable and that would require a huge majority.
Even if they decided to ament the constitution, you would still face another issue: now electoral system is written in the constitution, so it becomes even more difficult to change in the future.
What do you think about the system adopted by Maine and Nebraska instead?
You wind up with one representative for every 50,000 people, minimum. Currently, the number of representatives is decided by congress, and the districts are drawn up by them as well. So you've got districts with overrepresented constituents, just like states with the electoral college. It is capped at 435 currently. The parties of course collude to keep it this way, they agree to trade power when redistricting and stuff like that.
With one representative per 50,000 minimum, you wind up with a house that would be, today, about 7000 members. And most of what they do can be passed by a simple majority. House representatives have to campaign directly to their constituents, having that few per means they have to get closer to what they want, which means that, as far as direct legislative representation goes, the pressure to run on overarching political football platforms wanes and the pressure to run on niche and local concerns dominates. With simple majority in the house for most things, that means you have to deliver on a lot more of those local concerns to get anything passed, because half of 7000 people is going to be hard to whip up for a vote on anything.
So you may get some aligned groups caucusing together, you may get them nominally under the same umbrella, you'll get coalitions, all just like happens in Europe, but people will get more granular, close to home representation.
Our problem isn't that there's a lack of ideas on how to reform. It's that said reform requires amending the constitution, which in US is contingent on 3/4 of states (specifically, their legislatures) ratifying the amendment. It is a very high bar in general, but is especially hard to break through when what you propose is virtually certain to change the balance of political power in the country.
In this case isn't it up to states to decide how they select electors? So in sense, nothing stops any state from assigning them in some proportional manner.
And one could even ask, if it is electors who decide why not get them campaign? Not voting for president, but voting for person who votes for president. Isn't that the intention of EC anyway?
The states can assign them proportionally, but why would they? Again, when we say "the state", this really means state legislature. And, generally speaking, control of state legislature correlates pretty strongly with popular presidential vote in that state. So why would the party that controls the state, and which currently gets all of its electoral votes, effectively volunteer to surrender some of them to their political enemies?
Now, if one party generally expects to win the national popular vote more often than not, it might sign up for such a scheme on the condition that other states do - this is what the National Popular Vote Interstate Compact is. But, by the same token, it means that the other party would generally expect to lose the national popular vote, and thus would do the best it can to block such a proposal from going forward by preventing states under its control from joining (or by other means; e.g. North Dakota actually passed a law to block release of its popular vote data until after EC vote specifically to disrupt NPVIC).
Thing is, nobody in position of power today actually cares what the "real intention" of EC is or ever was. The only thing that matters today is winning elections, and so a bipartisan agreement on any electoral reform that upsets the existing balance of power is extremely unlikely.
In the EU there is already a similar system because, in the parliament, the number of MEPs is not really proportional to the population (Germany has 84,48 million inhabitants and 96 MEPs, so 1 every 880,000 people, while Malta has 553,000 people and 6 MEPs so 1 MEP every 92,000 people).
The difference is MEPs are appointed proportionally within each state (so if a party gets 40% in a state, it will get ca. 40% of the state's MEPs)
Sorry, bad phrasing. What I mean is, LLMs are more useful when they're asked to parse user input, into a more structured and narrow output, that is then fed to some business logic. Like, there's countless different ways the user could ask about their missing shipment; unlike earlier methods, LLMs are able to correctly classify almost all of them as "missing shipment". So do this, and then stop at that. The conversational output you send to the user does not need to be generated by an LLM. There's not much value added for the massive risk of hallucinations LLMs introduce.
Are there any ways to show the source of the information retrieved by the model? For instance, the LLM forms a sentence and it points to a stackoverflow answer with the same or similar content.
As I understand it, pretty sure that is impossible. When it is input a single datum, sure, trivial. As soon as it is fed a second one though the weights are already a kind of blend of the two tokens (so to speak).
Its not impossible, but its definitely difficult. There is some overlap in the methods used to detect benchmark data contamination, though its not entirely the same thing. For the detection use case, you already know the text you're looking for and you are just trying to demonstrate that the model has "seen" the data in its training set. The challenge is proving that it is statistically improbable that the model could stochastically generate the same tokens without having seen them during training.
Some great research exists in this area [1] and I expect much of it may be repurposed for black box attribution in the future (in addition to all the work being done in the mechanistic interpretability field)
If your dense housing leads to higher housing costs that's actually great news; it means that your city is enabling higher-productivity arrangements that boost local incomes
Take Milan as a counterexample. A single room in a shared flat costs around €700/month, a single-room studio €1200/month. Yet, people on average earn €30,000 a year, which means €24000 after taxes, or around €1800 per month. That's not what I would say increased salaries and productivity.
In Italy, there are serious problems with stagnating productivity and national, collective bargaining. Nevertheless, that doesn't mean rents do not skyrocket in big cities
Anectodically speaking, it has always been extremely helpful not only reading texts and solving problems/exercises, but also:
1) repeating topics out loud and, if there were some formulas, writing them down while doing that
2) explaining something to your friend if they didn't understand it
Anyway, as I told in another comment, I'm not from the US, so my intention wasn't really to say how to implement the proposal. I just had this idea and wrote it down in a blog post.