Elie Bursztein
elie.net · Personal blog · on HN since 2014
Elie Bursztein leads anti-abuse research at Google and writes about security research, including machine learning attacks, poker cheating devices and side channels.
What HN saysHN links it for the poker cheating device write-up and the Defcon talks on deep-learning side-channel attacks, which commenters call fascinating.
38HN people shared it
9linked it in comments
67stories submitted
1.9Kstory points
1 year agolast shared
2014people per year2026
Most-shared posts
- 1Exposing high-end poker cheating devices6 people · 739 points · first shared 2016 · 140 comments on HN
- 2Hacking Hearthstone with machine learning – Defcon talk wrap-up4 people · 121 points · first shared 2014 · 31 comments on HN
- 3Attacks against machine learning – an overview3 people · 336 points · first shared 2018 · 64 comments on HN
- 4The bleak picture of two-factor authentication adoption in the wild3 people · 87 points · first shared 2018 · 95 comments on HN
- 5Deep-learning based side-channel attacks with TensorFlow tutorial3 people · 7 points · first shared 2021 · discussion on HN
- 6Understanding the safety and privacy challenges faced by South Asian women3 people · 6 points · first shared 2019 · discussion on HN
- 7Insights on the First Three Years of the Right to Be Forgotten at Google2 people · 112 points · first shared 2018 · 83 comments on HN
- 8Deep-learning side-channel attacks: the theory2 people · 42 points · first shared 2021 · 2 comments on HN
- 9A Hacker Guide to Deep-Learning Based AES Side Channel Attacks2 people · 4 points · first shared 2019 · discussion on HN
- 10Homepage2 people · first shared 2014
- 11What are malicious USB keys and how to create a realistic one?1 people · 124 points · first shared 2016 · 77 comments on HN
- 12Taking down Gooligan, an Android botnet1 people · 75 points · first shared 2018 · 2 comments on HN
- 13Password checkup: from 0 to 650k users in 20 days1 people · 74 points · first shared 2019 · 15 comments on HN
- 14Quantifying the impact of the Twitter fake accounts purge – a technical analysis1 people · 60 points · first shared 2018 · 20 comments on HN
- 15Account Security – A Divided User Perception1 people · 21 points · first shared 2019 · 3 comments on HN
What HN people wrote when they linked it
chrisco255 · 2023-09-26
It doesn't need to be a sex toy, obviously. $5 worth of parts from Ali Baba could be assembled into a signaling tool:
https://www.alibaba.com/product-detail/Type-C-Bluetooth-5-0-Audio_1600…
There's even prebuilt solutions for things like poker:
https://www.cardslenses.com/remote-contorl-poker-game-vibrator.shtml
https://elie.net/blog/security/royal-flush-an-in-depth-look-at-poker-c…
ebursztein · 2023-01-06
https://elie.net mostly about the research we do at Google. Generated statically and hosted on firebase. The design was custom made for it.
declan · 2014-02-17
Elie Bursztein leads Google's anti-abuse research, and recently mentioned on Google+ that he moved his personal web site to GAE: http://www.elie.net/
I can load his GAE-hosted web site in ~300ms, and there's a useful performance box on the right sidebar of his site that breaks down how long each component takes to load on GAE:
$ time curl http://www.elie.net/ > /dev/null
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 36146 0 36146 0 0 113k 0 --:--:-- --:--:-- --:--:-- 114k
real 0m0.316s
user 0m0.005s
sys 0m0.005s
I've been thinking of doing the same with my personal web site but haven't yet.
raphman · 2022-06-29
Here's an article from 2017 talking about this class of devices:
https://elie.net/blog/security/fuller-house-exposing-high-end-poker-ch…
joatmon-snoo · 2019-08-16
Google's Elie Burzstein and Jean Michel Picod gave a fascinating talk at Defcon last week about silicon-level SCAs: https://elie.net/talk/a-hackerguide-to-deep-learning-based-side-channe…
Disclosure: also work at G.
ralfd · 2018-08-06
Already done, sort of. Blizzard was not amused, as botting makes the game less fun, so it wasnt released.
https://elie.net/blog/hearthstone/i-am-a-legend-hacking-hearthstone-wi…
michaelt · 2017-11-28
You can also get cheating devices [1] where the edges of all the cards are printed with IR barcodes, then a special camera (concealed in a modified phone or car key fob) can tell you upcoming cards by looking at the edge of the deck.
Of course, you need to convince people to play with your deck of cards first, so you can't just walk into a casino with these things.
[1] https://www.elie.net/blog/security/fuller-house-exposing-high-end-poke…
bengtan · 2021-06-07
Hi,
I put together a newsletter called Interesting Things that curates 40+ interesting stories and links every week about tech with a splash of business and science. Looking through the past few editions, there has been an average of one AI/deep-learning/neural-network related article per week:
Issue #001
Shedding light on fairness in AI with a new data set — Facebook publishes a dataset of of 45,186 human face-to-face videos with an emphasis on fairness and diversity across age, gender, skin tone, & ambient lighting. — https://ai.facebook.com/blog/shedding-light-on-fairness-in-ai-with-a-n…
Issue #002
Geometric foundations of Deep Learning — An attempt to unify a broad class of ML problems from the perspectives of symmetry and invariance. — https://towardsdatascience.com/geometric-foundations-of-deep-learning-…
Latest Neural Nets Solve World’s Hardest Equations Faster Than Ever Before — New approaches to solve entire families of partial differential equations and make it easier to model complicated systems. — https://www.quantamagazine.org/new-neural-networks-solve-hardest-equat…
Issue #004
Superhistory, Not Superintelligence — Simpler algorithms and more data do better than more complex algorithms with less data. What are the implications? — https://breakingsmart.substack.com/p/superhistory-not-superintelligenc…
Reimagining Databases for Vector Embeddings — How is a vector database different from a conventional…
jonathankoren · 2016-02-17
But this all hinges on the naive assumptions that this is a one off and will never happen again, and that later generation devices are immune from any circumvention attempt. History says this isn't how this plays out.
If you crack the encryption once you'll get orders to crack it again and again, and in much lower profile and lower stake cases. Look at the prevalence of espionage tactics such as Stingrays and "parallel construction" by law enforcement. There may not always be someone you can pump up into an crack international terrorist, but there's always some low level drug courier, or a "quality of life" criminal to use your new toys on.
Also you're saying hat this doesn't set a precedent, but it does. Sure there's not a court case to point to, but it's a precedent none the less. It's that the company not only has the means, but the will to do it. What's stopping the government from coming back a second time, or a third time about this? What argument do you have on either a legal court or the court of public opinion to make when you stand up and say, "That first time was an exigent situation, and so was the second, and the third... But this time, the fourteenth time, THIS TIME we really mean no more!"
Finally, I don't think this trick only works on older devices. The FBI wants them to be able to brute force the passcode through a USB connection instead of making some sort of robot to tap the screen a bunch of times. Also presumably the FBI wants the the two many…