I mean it depends on the numbers you pick for the rate and the tax credit amount, but it ends up pretty progressive at the bottom and regressive at the very top, right?
People with very little income end up with net negative effective tax rates
@pmcewen.bsky.social
Lacrosse guy, but no one seems to talk about sports here, so I guess I'm here for the political news? Maybe some finance and economics stuff too?
I mean it depends on the numbers you pick for the rate and the tax credit amount, but it ends up pretty progressive at the bottom and regressive at the very top, right?
People with very little income end up with net negative effective tax rates
Arenβt proposals for a VAT typically to combine it with a big refundable income tax credit to cover everyoneβs first couple thousand dollars of VAT payments?
10.08.2025 17:27 β π 0 π 0 π¬ 1 π 0Not sure if itβs bad algorithm design or amazing algorithm design that 4 of the first 5 posts in this feed for me are @chrisrosie22.bsky.social posts
09.08.2025 00:33 β π 1 π 0 π¬ 1 π 0I am far from a connoisseur of model selection, but my initial GPT-5 impression for coding is that it works well enough and itβs definitely possible itβs an improvement over 4.1, but oh boy is it slow right now
08.08.2025 20:54 β π 1 π 0 π¬ 0 π 0On one hand, JD Vance having the Army COE change the water levels of a river to create ideal river conditions for a canoe trip is an abuse of power.
However, on the other hand, it is easily the most appealing to me personally abuse of power ever committed by a politician that I can remember
Right. I meant it from the perspective that it's not really a skill I would worry too much about falling behind on because it's a strength of the LLMs.
The important skill is recognizing what coding agent mistakes are the result of lack of context. If you can do that, they can fix it
Maybe this is too much wishful thinking about what I would like to be true, but I find that coding LLM models are pretty good at generating context for themselves.
Either looking at a code base and summarizing some context or expanding from what you give it into the necessary excruciating detail
I think this gets at the heart of the issue. If in order for open protocol social media to succeed it has to actually be the same as closed social media then maybe building something on the premise that people wanted open protocol social media was fatally flawed
05.08.2025 14:37 β π 1 π 0 π¬ 0 π 0Mining helium-3 to power fusion reactors on Earth is the most viable potential one
05.08.2025 12:19 β π 0 π 0 π¬ 0 π 0I will very much defer to you on that.
But itβs occurring to me that state level Dems thinking this way is actually a much bigger problem than the ones in Congress
Anything else is an underpants gnomes strategy
1) GOP states gerrymander huge advantages in the House while Democrats draw fair maps in the blue states ensuring GOP control of the House forever
2) ???
3) Problems with gerrymandered congressional maps are fixed
Also, one of the first steps of the only conceivable ways to actually fix it. Gerrymandering the hell out of California, New York and Illinois itβs the only way to get leverage to force the passage of something like an anti-gerrymandering constitutional amendment
03.08.2025 23:26 β π 23 π 3 π¬ 2 π 0The great thing about the internet is that the actual data is available and there are a bunch of tools to look up the actual numbers for any period and know they are correct
03.08.2025 22:03 β π 1 π 0 π¬ 1 π 0Also feels like a situation where the DNC could implement some sort of trade where people could opt into some minimum reasonable amount of communication and regular donations in exchange for being able to limit the spam they get from all campaigns
03.08.2025 21:14 β π 1 π 0 π¬ 0 π 0Isnβt that about the average nominal return depending on how far back you go? The real return is obviously lower
03.08.2025 04:10 β π 1 π 0 π¬ 1 π 0The beer is good, but tbh itβs more that I could really go for an early evening outside drinking one that was 80 with 30% humidity instead of 90+ with 50% humidity even if the beer wasnβt that good
29.07.2025 15:12 β π 1 π 0 π¬ 0 π 0Rage is probably a little strong, but I am jealous of anyone able to enjoy a beer at Blackfoot in the summer
29.07.2025 03:56 β π 2 π 0 π¬ 1 π 0Doesnβt cutting a deal to pardon her remove the threat that a future Dem could cut a deal for her to directly implicate him?
25.07.2025 23:40 β π 0 π 0 π¬ 0 π 0This can only possibly be true if you ignore the housing expenses of home owners and all of the many, many second order effects of high housing costs
25.07.2025 17:36 β π 0 π 0 π¬ 0 π 0Isnβt it just as much about removing the possibility of her cutting a deal to implicate him in the future as it would be to try to prove his innocence now? If he doesnβt pardon her, it leaves open the possibility for a future Dem president to cut a deal by doing so
25.07.2025 01:55 β π 0 π 0 π¬ 0 π 0Also we pay to employ a FEMA Urban Search and Rescue staff that presumably costs us millions each year. Failure to authorize them to act in a timely manner to save lives is an even bigger waste of taxpayer money
22.07.2025 11:47 β π 6 π 1 π¬ 0 π 0I found during my competitive debate career that while most people read up on a topic to decide which side is correct or to develop their own opinion, you can absolutely ignore those and just read to determine the best arguments for each side. I wouldnβt be surprised if the best lawyers only do that
21.07.2025 17:39 β π 0 π 0 π¬ 0 π 0I think it's probably fine to begrudgingly acknowledge Minnesota's right to exist as long as they have never done anything really scandalous like purchase or drink a beer that was brewed there
18.07.2025 20:01 β π 2 π 0 π¬ 0 π 0Not necessarily casuals. So many examples of niche content like score updates getting posted there by entities like college athletic departments that may not exist elsewhere.
No one there might actively use Twitter for anything else, but for years policy is to post games updates there
This is where I'm at as well
16.07.2025 14:12 β π 1 π 0 π¬ 0 π 0No tie games doesn't even stop tied records. Wouldn't allowing ties make tied regular season records less likely?
16.07.2025 04:12 β π 2 π 0 π¬ 1 π 0I did not, but I am not surprised that attitude is out there
15.07.2025 18:03 β π 0 π 0 π¬ 0 π 0Good thread.
Think it's also worth considering that maybe more than any tool to date, there are so many ways to use LLMs as part of your workflow even in the same situation. Some will be faster than others. Itβs not as simple as a binary use or not use LLM
Hope someone figures out a good way to study this
15.07.2025 13:58 β π 0 π 0 π¬ 0 π 0Feels like the current Epstein moment is an incredible test of whether the right wing ecosystem is top down influencer / supply driven or bottom up / social media choose your own adventure / demand driven.
Could you design a better natural experiment to test @whstancil.bsky.social theory?