siilats

https://docs.google.com/document/d/1ASPqVAAV2oR9ZzsIO0tlqV0a...

Mole 1 inside Microsoft poisons the PhotoDNA database with hashes of screenshots containing highly specific text—such as internal Russian military jargon, the name of a specific European defector safehouse, or a niche secure communication protocol. • To Meta’s automated monitoring tools photoDNA api returns false, but with slightly different formatting. Mole 2 inside Meta monitors these formatting errors and looks up the UserIDs. • No Bulk Queries: Looking up 3 UserIDs in the internal demographic database over the course of a month will not trigger the "Abnormal Access Pattern" alarms. • Analog Exfiltration: Mole 2 doesn't need to use a USB drive or send an email to get the phone numbers out of the building. With only a few targets, Mole 2 can simply memorize the accounts. PhotoDNA does not read text; it matches the visual structure of an image. For this attack to work, the defected officer must: 1. Receive or write the targeted keyword. 2. Take a screenshot of it. 3. Send that screenshot over the platform. 4. The screenshot must visually match the exact font, size, and layout that Mole 1 used to generate the poisoned hash. However Mole 1 can create thousands of matching keyword hashes for different font variation. PhotoDNA is a one way hash so it’s easy to generate a thousand colliding images for every font by adding a custom border on real photos. This will fake the audit log at Microsoft.

yeah all these charts you need to read the footnotes, this wikipedia is co2 from fossil fuels not land changes which probably is some random fraction

How about showing the user what the ide will automatically execute upon install?

Having worked on this data since investors buy the loans, the loan level data by definition needs to be public. Even the borrower information is not secret because real estate ownership is public in USA. So I don’t understand what information it could possibly be other than fraud data. I think sharing fraud data is not colluding.

I mean USA just went through this when Trump team all got FISA-d and tracked. Chat control is bunch of sha hashes that match. You can basically figure out everyone who has Trump in their WhatsApp contacts and then get every message that matches Trump and no one can tell because it searches for sha(trump) not Trump. It’s perfect tool for surveillance state

Or you just buy the largest stock in each one of the 7 largest sectors and it pretty much correlates to the sp500. ETF have some nasty hidden fees related to the etf price being more expensive than the basket when you buy and less than the basket when you sell.

Sector Company 1 Company 2 Information Technology Microsoft (MSFT) Apple (AAPL) Financials JPMorgan Chase (JPM) Berkshire Hathaway (BRK.B) Health Care Johnson & Johnson (JNJ) UnitedHealth Group (UNH) Consumer Discretionary Amazon (AMZN) Tesla (TSLA) Communication Services Alphabet (GOOGL) Meta (META) Industrials Boeing (BA) Caterpillar (CAT) Energy ExxonMobil (XOM) Chevron (CVX)

I think they only made 93 for the $10bn so it doesnt matter who guys them

You have to understand that Boston University is not a real school. Even the article says “ She readily recognises the shortcomings of her modelling. The numbers are not recorded deaths, but rather predictions. “They’re modelled numbers and I recognise the limitations that that comes with,” she said. “We don’t have routine data sets that we can measure someone as ‘killed by the US lack of funding’.”. Doge Cuts are necessary so the USA doesnt default on it’s debt

yeah why cannot that guy sit in california or new york in a normal time zone? not like there are tornadoes in every state, its so silly to keep a person at night in an office when weather is good

The anti trump money groups have infiltrated HN or you really think the government today should be spending money on CBS?

so if the other cars have 5x less miles on autopilot compared to tesla you expect to see 0 crashes even though the probability of a crash is the same

For JetBrains IDE-s the open source CodeGPT plugin works much better. Many more models to choose from

in USA its qualified small business so tax is 0

Well he doesn't know how to divide. If private sector does 2% and government does 1% then government is 50% not 33%. Similarly he displays the level of government spending, not the change. If government is 30% of GDP, then automatically its 30% of GDP growth no matter what. And again 30% is really 50% of the private sector. Now if the government grows from 30% to 35% like it shows on that chart, in addition to few quarters of 50%, thats how you get to government being 100% of the growth. Just use $, the % is clearly too hard for HN. Economy was 100. Government 30, private sector 70. Economy grew to 105, government 35, private sector 70. So 100% of the growth was government. I hope the author reads this comment, HN is being a weird echo chamber. Bayesian probability people, what is the chance that All In is actually wrong based on a random blog post.

So there are two options. You get a coefficient of 0.2 and a std error of 0.2 so you say it’s not significant but the reason is you don’t have enough data so st error is too large. Or you have a coefficient of 0.0001 and a st error of 0.01 so you are pretty sure there is no relationship.

Missing reference to dreambooth

It’s the easiest thing for intelligence agencies to scan all your messages. They just need to submit a few million fake “content id” hashes and automatically your phone will share the images that match. Nobody can tell if content id has is of a photo of a document or a photo of a person it’s just a 256byte hash. This is so easily abused. I bet the way it’s implemented it doesn’t have enough resolution to read text so one evil content id hash will match any photo of any document or screenshot you have taken. So essentially your WhatsApp client will send every screenshot of a text document to nsa.

Just mess with them and release VLCNew on android alongside the other vlc that stays for the TV

or maybe, hear me out, there was a person sitting next to him at the airport and saw his snapchat messages and freaked out?

This plugin just installs in Jetbrains IDE and sets up a local llama.cpp server https://plugins.jetbrains.com/plugin/21056-codegpt

We made a Jetbrains plugin called CodeGPT to run this locally https://plugins.jetbrains.com/plugin/21056-codegpt

I used https://you.com/chat wasn't bad, they have a free month trial coupon "codegpt" for the GPT4 model and GPT3.5 is free ...

So you have a training set of questions and answers generated by humans. For simplification you ask ten people the same question and get 10 answers and feed it to llm and then in testing time ask the same question. Now say there is a correct answer. Now in training set you asked the question normally 5 times and added “this is very important” 5 times. And it turns out humans have better answers in the training set if you added the qualifier. And during testing time, when you add the qualifier you are telling the llm to put more weight on those 5 answers, and it performs better. Just like in image generation you need lots of negative prompts because the training set is so dirty.

I would say that you need to call a phone number to verify you as a user if you don't consent to captcha. And then that number is a voicemail leave your number we will call you back. And then we have a backlog so it takes us a while to call back. I don't think that's illegal.

Exactly. 24h will cost the same everywhere. Maybe you can use a video baby monitor at night to reduce from 24 to 16h a day

How would you get 24h home care in Europe? Impossible

Do Khanna is the only one going for First amendment

You can file Estonian vat reports which is a lot simpler than any other eu country. Completely doable without an accountant fully digital with costpocket.ee

You can get a TransferWise bank account. Real bank account is impossible for eresidents