11 Matching Annotations
  1. Mar 2024
    1. The White House immediately blasted the proposal for trying to raise the retirement age and reiterated the promise Biden made at the State of the Union address to block any proposed cuts to Social Security.

      We need to rephrase this issueto be; "The white house IS planning a 23% cut in ALL social security benefits in 2033". THAT is actually the PLAN, look at the social security site, trustees reports.

      Taking no action means we are continuing as PLANNED, cut benefits 23% in 10 years.

      We need a new plan, actually just pull up the plans from 10, 20, 30 years ago to address this very issue of addressing social security solvency. Nothing has changed, except for the urgency.

    1. The electric vehicle market, while growing, is in flux, due to softening consumer demand and persistently high interest rates. Just last month, Apple—the first company in history to ever record a $3 trillion valuation—canceled its decade-long quest to develop an electric car. General Motors and F

      Typical series of steps to 'help'. New industry gets hot (EVs), state/feds jump in with subsidies (money, laws), business investment doubles down with all the free money, plans and promises are made, financial expectations are increased even more, 6 months later doesn't look quite as good anymore, government subsidies/laws take a while to kick in due to regulations and physics, everyone gets scared, crash

    1. I wear this debt clock that I built in Congress to remind people of it. One side effect of me wearing this is that I've noticed the rate at which the debt is increasing is going up. For the math nerds, that's the second derivative. Today, the debt per second is an average of $78,000.

      thats $78K increase per second. So if the current new debt per second is $62,000,000 (yes that's millions), then the following second we will borrow $62,078,000.

      this site is entertaining, and scary: https://www.usdebtclock.org/

  2. Feb 2024
    1. But it's really absurd that seniors making $1 million a year after retirement are still getting generous benefits.

      how many $1million seniors are there?

    1. If fraud "is insignificant," Engorion concedes, "then, like most things in life, it just does not matter." But that "is not what we have here," he add

      i hope this forgiving attitude is applied equally to all defendents, especially poor. But I'm afraid to look at Engoron's record.

    1. And in woeful en banc news, the Fifth Circuit will not reconsider its ruling that a SWAT raid that caused $60k in damages to an innocent woman’s home is not a taking that requires just compensation. Dissental: “By placing the onus on [plaintiff] to ground her right to compensation in a historical analogue—rather than requiring the City to establish some historically based exception to the compensation requirement—the panel flipped the burden that typically governs in cases involving individual rights.”

      and bad news

    2. And in marvelous en banc news, the Fifth Circuit will reconsider its ruling that a woman falsely convicted by a Midland County, Tex. prosecutor who was simultaneously working as a law clerk for the presiding judge cannot bring constitutional claims against the prosecutor and his supervisors. The original three-judge panel called the situation “utterly bonkers,” but held that circuit precedent required her to successfully petition for habeas corpus before bringing suit, even though she cannot so petition because she never went to jail.

      good news

    1. Additionally, you can provide the url of a YouTube playlist and the app will load the transcriptions of the videos in the playlist, enabling you to query the content they cover.

      this seems super useful!

    1. Allegation: U.S. Marshals sneak up on suspect, kick him in the head, and take turns stomping on him while he’s unconscious. Unconstitutional excessive force? Tenth Circuit: You can’t sue federal officers for violating the Constitution. (IJ filed an amicus brief urging a different course.)

      follow case, bivens, federal officer immunity

    1. The GPT Tree Index first takes in a large dataset of unprocessed text data as input. It then builds up a tree-index in a bottom-up fashion; each parent node is able to summarize the children nodes using a general summarization prompt; each intermediate node containing summary text summarizing the components below. Once the index is built, it can be saved to disk and loaded for future use.Then, say the user wants to use GPT-3 to answer a question. Using a query prompt template, GPT-3 will be able to recursively perform tree traversal in a top-down fashion in order to answer a question. For example, in the very beginning GPT-3 is tasked with selecting between *n* top-level nodes which best answers a provided query, by outputting a number as a multiple-choice problem. The GPT Tree Index then uses the number to select the corresponding node, and the process repeats recursively among the children nodes until a leaf node is reached.

      Interesting, this explains how indexing and searching is very different, it's building a tree of nodes using the LLM to summarize children in a parent, during indexing, and then using the LLM and vector index to traverse top down and select the best response.