The Phantom of Heilbronn (or how for 16 years the German police hunted a serial killer who didn't exist)
(Thread, non sweary version here: https://t.co/iDOIhaFgIT)

In 2005, two brothers got into a domestic dispute and one brother fired on the other, and then the police *found her DNA on the bullet*.
More from Tech
The story doesn\u2019t say you were told not to... it says you did so without approval and they tried to obfuscate what you found. Is that true?
— Sarah Frier (@sarahfrier) November 15, 2018
In the spring and summer of 2016, as reported by the Times, activity we traced to GRU was reported to the FBI. This was the standard model of interaction companies used for nation-state attacks against likely US targeted.
In the Spring of 2017, after a deep dive into the Fake News phenomena, the security team wanted to publish an update that covered what we had learned. At this point, we didn’t have any advertising content or the big IRA cluster, but we did know about the GRU model.
This report when through dozens of edits as different equities were represented. I did not have any meetings with Sheryl on the paper, but I can’t speak to whether she was in the loop with my higher-ups.
In the end, the difficult question of attribution was settled by us pointing to the DNI report instead of saying Russia or GRU directly. In my pre-briefs with members of Congress, I made it clear that we believed this action was GRU.
At the heart of this lies the most important technique in modern deep learning - transfer learning.
Let's analyze how it
THREAD: Can you start learning cutting-edge deep learning without specialized hardware? \U0001f916
— Radek Osmulski (@radekosmulski) February 11, 2021
In this thread, we will train an advanced Computer Vision model on a challenging dataset. \U0001f415\U0001f408 Training completes in 25 minutes on my 3yrs old Ryzen 5 CPU.
Let me show you how...
2/ For starters, let's look at what a neural network (NN for short) does.
An NN is like a stack of pancakes, with computation flowing up when we make predictions.
How does it all work?

3/ We show an image to our model.
An image is a collection of pixels. Each pixel is just a bunch of numbers describing its color.
Here is what it might look like for a black and white image

4/ The picture goes into the layer at the bottom.
Each layer performs computation on the image, transforming it and passing it upwards.

5/ By the time the image reaches the uppermost layer, it has been transformed to the point that it now consists of two numbers only.
The outputs of a layer are called activations, and the outputs of the last layer have a special meaning... they are the predictions!

You May Also Like
As someone\u2019s who\u2019s read the book, this review strikes me as tremendously unfair. It mostly faults Adler for not writing the book the reviewer wishes he had! https://t.co/pqpt5Ziivj
— Teresa M. Bejan (@tmbejan) January 12, 2021
The meat of the criticism is that the history Adler gives is insufficiently critical. Adler describes a few figures who had a great influence on how the modern US university was formed. It's certainly critical: it focuses on the social Darwinism of these figures. 2/x
Other insinuations and suggestions in the review seem wildly off the mark, distorted, or inappropriate-- for example, that the book is clickbaity (it is scholarly) or conservative (hardly) or connected to the events at the Capitol (give me a break). 3/x
The core question: in what sense is classics inherently racist? Classics is old. On Adler's account, it begins in ancient Rome and is revived in the Renaissance. Slavery (Christiansen's primary concern) is also very old. Let's say classics is an education for slaveowners. 4/x
It's worth remembering that literacy itself is elite throughout most of this history. Literacy is, then, also the education of slaveowners. We can honor oral and musical traditions without denying that literacy is, generally, good. 5/x

Just added Telegram links to https://t.co/lDdqjtKTZL too! Now you can provide a nice easy way for people to message you :)

Less than 1 hour since I started adding stuff to https://t.co/lDdqjtKTZL again, and profile pages are now responsive!!! 🥳 Check it out -> https://t.co/fVkEL4fu0L

Accounts page is now also responsive!! 📱✨

💪 I managed to make the whole site responsive in about an hour. On my roadmap I had it down as 4-5 hours!!! 🤘🤠🤘
To me, the most important aspect of the 2018 midterms wasn't even about partisan control, but about democracy and voting rights. That's the real battle.
2/The good news: It's now an issue that everyone's talking about, and that everyone cares about.
3/More good news: Florida's proposition to give felons voting rights won. But it didn't just win - it won with substantial support from Republican voters.
That suggests there is still SOME grassroots support for democracy that transcends
4/Yet more good news: Michigan made it easier to vote. Again, by plebiscite, showing broad support for voting rights as an
5/OK, now the bad news.
We seem to have accepted electoral dysfunction in Florida as a permanent thing. The 2000 election has never really
Bad ballot design led to a lot of undervotes for Bill Nelson in Broward Co., possibly even enough to cost him his Senate seat. They do appear to be real undervotes, though, instead of tabulation errors. He doesn't really seem to have a path to victory. https://t.co/utUhY2KTaR
— Nate Silver (@NateSilver538) November 16, 2018