This needs to be changed. This is gamed by bad actors to limit the reach of accounts they don't like
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π1
This media is not supported in your browser
VIEW IN TELEGRAM
A retired Air Force lieutenant colonel warns that exposing sensitive security details used to protect Trump on Air Force One could endanger future presidents and military personnel for years
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π₯1
This media is not supported in your browser
VIEW IN TELEGRAM
A tiny home village has been unveiled in Sun Valley, featuring small white and yellow units expected to house more than 200 people, with shared bathrooms and basic air conditioning
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π12
BREAKING: Tether just got its first full independent financial statement audit by KPMG.
And it was unqualified, the cleanest audit opinion possible.
KPMG verified Tetherβs assets, liabilities, transactions, systems, counterparties, valuations AND physically inspected every gold bar.
Tether also had $6.814B more reserves than liabilities.
Years of βTether is a fraudβ FUD just got a lot harder to defend.
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
And it was unqualified, the cleanest audit opinion possible.
KPMG verified Tetherβs assets, liabilities, transactions, systems, counterparties, valuations AND physically inspected every gold bar.
Tether also had $6.814B more reserves than liabilities.
Years of βTether is a fraudβ FUD just got a lot harder to defend.
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π₯6
BREAKING: π is rolling out a new βUnder the Hoodβ page that shows whether your account or posts have labels that may limit their reach.
You can see what each label means, how it affects visibility, and download a monthly report.
Check yours:
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
You can see what each label means, how it affects visibility, and download a monthly report.
Check yours:
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
This media is not supported in your browser
VIEW IN TELEGRAM
Iβm here trying everything to make enough money for my family, killing myself 12β14 hours a day, while some low-life women are selling feet toast.
Fvck my life.
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
Fvck my life.
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π4π3
JUST IN : Stock Market hits highest level of Greed in 3 months
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π₯1
This media is not supported in your browser
VIEW IN TELEGRAM
Today, 3 men in Poland catch a toddler that fell from a 5th story balcony
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π«‘18π₯°3
A rich Chinese woman cast herself as the lead in a drama and requested 60 kissing scenes with a handpicked male actor
More than 20 intimate scenes had to be removed for being too explicit
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
More than 20 intimate scenes had to be removed for being too explicit
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π₯°3π2
X just open-sourced its new algorithm. How to crush it:
> post original content that gets more replies, especially from mutual followers
> predicted quotes carry a 5Γ weight and predicted follows carry 4Γ
> multiple posts from the same author for one feed compete with each other, meaning: don't flood the tl with many posts
> avoid repeating exact text. X has rules that label duplicate content as COPYPASTA_SPAM
> predicted βNot interested,β blocks, mutes, and reports crush a postβs score, reports carry the strongest negative weight
> likes contribute to a SimClusters discovery score, helping your post reach people with similar interests to the person who liked them, and it decays with an 8-hour half-life
> likes, quotes, reposts, bookmarks affect your positive-interest embedding, giving you a better feed. so if you see too many irrelevant posts in your feed, engage more with the ones you like.
X keeps an account-reputation score called UserCredV2, which is more similar to a trust/authority measure. This is your account credibility. What makes it go up:
> follower quality, not just counts. a follow from a credible account contributes more than a follow from a low-cred account.
> likes and reposts received in the last 7 days, weighted by the credibility of the people engaging. It's more like a momentum effect.
> a credible account that follows 1000 people transfers more reputation per follower than another account that follows 10,000. the same is true about their likes/reposts
NOTE: the values you see in the image are for PREDICTED EVENTS, so a like doesn't literally add +0.5 weight as it happens.
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
> post original content that gets more replies, especially from mutual followers
> predicted quotes carry a 5Γ weight and predicted follows carry 4Γ
> multiple posts from the same author for one feed compete with each other, meaning: don't flood the tl with many posts
> avoid repeating exact text. X has rules that label duplicate content as COPYPASTA_SPAM
> predicted βNot interested,β blocks, mutes, and reports crush a postβs score, reports carry the strongest negative weight
> likes contribute to a SimClusters discovery score, helping your post reach people with similar interests to the person who liked them, and it decays with an 8-hour half-life
> likes, quotes, reposts, bookmarks affect your positive-interest embedding, giving you a better feed. so if you see too many irrelevant posts in your feed, engage more with the ones you like.
X keeps an account-reputation score called UserCredV2, which is more similar to a trust/authority measure. This is your account credibility. What makes it go up:
> follower quality, not just counts. a follow from a credible account contributes more than a follow from a low-cred account.
> likes and reposts received in the last 7 days, weighted by the credibility of the people engaging. It's more like a momentum effect.
> a credible account that follows 1000 people transfers more reputation per follower than another account that follows 10,000. the same is true about their likes/reposts
NOTE: the values you see in the image are for PREDICTED EVENTS, so a like doesn't literally add +0.5 weight as it happens.
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π4
JUST IN: Dozens of Andrew Tate supporters stage a βFree Top Gβ protest in Miami demanding he & brother Tristan be released from federal custody
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π±3π3
With open model token prices rising, one can guess what is going on at Anthropic and ChatGPT.
What happens when China floods the world with cheap Ascend chips to power the Chinese open models
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
What happens when China floods the world with cheap Ascend chips to power the Chinese open models
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π3
US-based investors are dumping Chinese equities:
US-listed China ETFs recorded -$3.4 billion in combined outflows across May, June, and July.
This includes the China Large-Cap ETF, $FXI, the MSCI China ETF, $MCHI, the CSI 300 China ETF, $ASHR, the China Internet ETF, $KWEB, the China Technology ETF, $CQQQ, and the China Technology and Semiconductor ETF, $KSTR.
As a result, the 12-month cumulative outflow fell to -$1.0 billion through July, the largest withdrawal in at least a year.
Year-to-date, the China Large-Cap ETF, $FXI, has seen the largest outflow at -$1.6 billion.
This is followed by the MSCI China ETF, $MCHI, at -$868 million, and the China Internet ETF, $KWEB, at -$833 million.
Investor demand for Chinese equities is weak.
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
US-listed China ETFs recorded -$3.4 billion in combined outflows across May, June, and July.
This includes the China Large-Cap ETF, $FXI, the MSCI China ETF, $MCHI, the CSI 300 China ETF, $ASHR, the China Internet ETF, $KWEB, the China Technology ETF, $CQQQ, and the China Technology and Semiconductor ETF, $KSTR.
As a result, the 12-month cumulative outflow fell to -$1.0 billion through July, the largest withdrawal in at least a year.
Year-to-date, the China Large-Cap ETF, $FXI, has seen the largest outflow at -$1.6 billion.
This is followed by the MSCI China ETF, $MCHI, at -$868 million, and the China Internet ETF, $KWEB, at -$833 million.
Investor demand for Chinese equities is weak.
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
BREAKING: Reddit stock, $RDDT, surges over +11% on news that the company is being added to the S&P 500
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π4
A 27-year-old Spanish woman who worked for the Red Cross to teach young girls about how to spot and avoid gender-based violence has been murdered by her 39-year-old Moroccan ex-boyfriend in Santander, Spain.
Paula Solanas de Miguel was found stabbed dead in her apartment.
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
Paula Solanas de Miguel was found stabbed dead in her apartment.
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π6π5β‘1
X just open sourced their algo.
If you want to help accounts out, these are the interactions that matter.
The best easy thing you can do is reply but all these things are good.
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
If you want to help accounts out, these are the interactions that matter.
The best easy thing you can do is reply but all these things are good.
π³πΎπΎπΌπΏπ€π π πΈπ½πΆ
π1