whoami_nr

8 ms·

61.
by whoami_nr 3y ago
Author here. Thanks for your list. Every paper I read on this topic has Carlini or has roots to his work. Looks like he has been doing this for a while. I shall check out your links though some of them have been linked in the post (at the b
62.
by whoami_nr 3y ago
Author here. Some of them are black box attacks (like the one where they get the training data out of the model) and it was done on Amazon cloud classifier which big companies regularly use. So, I wouldn’t say that these attacks are entirel
63.
by whoami_nr 3y ago
Author here. I get what you mean and I remember the incident happening when I was in college. However, I also remember that they were reproduced across multiple publications which means you are implying some sort of data poisoning attack wh
64.
Attacks on machine learning models(rnikhil.com)
103 points by whoami_nr 3y ago|38 comments
65.
AI Alignment – Weak-to-strong generalization (W2SG) explained(rnikhil.com)
1 points by whoami_nr 3y ago|0 comments
66.
by whoami_nr 3y ago
Yes I have seen the thread. Catching a player using just the winrate is hard. 9k hands is peanuts and some of the assumptions in that thread (like a 53% VPIP player must have 0bb/100 winrate) are slightly far fetched. I know for a fact
67.
by whoami_nr 3y ago
You are sort of correct. This used to be the case until 2 years back but now there are solvers like Monker which solve multiway spots fairly fast.
68.
by whoami_nr 3y ago
It’s the latter. All the best engines are closed source. It makes more sense to capitalise on it by deploying it for coaching/playing.
69.
by whoami_nr 3y ago
Ah then, just go buy any of the training tools like GTO+, GTOBase, Vision(by run it once), GTOWiz, PLOMastermind tool by Jnandez etc. They are essentially coaching software to learn the game(powered by a solver) and a ton of them have a pra
70.
by whoami_nr 3y ago
Author here. There are CFR implementations and even open source implementations of Pluribus. A google search shows me many such implementations. The tough part is not implementing the bot or replicating the research(Noam Brown mentions in t
71.
by whoami_nr 3y ago
Author here. Yes, you can still make money in online Poker. Its not as juicy as it was and rake is very high these days making micro stakes unbeatable. Moreover, highest stakes usually run live and access is gated. The old 2000s dream of gr
72.
Counterfactual Regret Minimisation or How I won any money in Poker?(rnikhil.com)
117 points by whoami_nr 3y ago|26 comments
73.
by whoami_nr 3y ago
Why next-token prediction is enough for AGI - Ilya Sutskever - https://www.youtube.com/watch?v=YEUclZdj_Sc
74.
Intro to LLM Security(rnikhil.com)
2 points by whoami_nr 3y ago|0 comments
75.
Chinchilla Paper explained(rnikhil.com)
4 points by whoami_nr 3y ago|1 comments
76.
by whoami_nr 3y ago
yeah, its hilarious asking it "who are you" questions? Every time (for me) it makes some new Phd student/AI researcher names.
77.
by whoami_nr 3y ago
Pretty useless as an online retrieval tool especially given that the knowledge base is outdated (capped to April 2023)
78.
The Novice's LLM Training Guide(rentry.co)
2 points by whoami_nr 3y ago|0 comments
79.
by whoami_nr 3y ago
The new tax regulations effectively broke the industry. Currently the websites are bearing the tax burden in effect sacrificing CM1 to keep the revenue constant. This is not sustainable to small players and lot of small RMG companies alread
80.
by whoami_nr 3y ago
Yes. Its as you describe it. One of my first coaches and probably the best MTT player in India (Danish Shaikh) quit poker for the exact same reason. He got so good at the game that he pitied the people he won money from. After a point, he j
81.
by whoami_nr 3y ago
No problem. Please replace the above link with the original: https://rnikhil.com/2023/11/12/quitting-fulltime-poker.html
82.
by whoami_nr 3y ago
Author of the post here. Yes RTA are common. Botting is common. RTA based on simple GTO solvers is common. I have built bot/fraud/collusion detection for the websites along with being on the other side as well. I recently built th
83.
by whoami_nr 3y ago
Author here. Thanks for the kind words and happy to answer any questions. If you like my writing, checkout my blog at rnikhil.com where this was posted originally.
84.
by whoami_nr 3y ago
Author here. I mentioned Pio in a different comment. You are literally the reason I won any money in Poker and I must have studied using Pio/Monker for 1000hrs+ easily. On a relevant note, I built a preflop tool for PLO folks. Check it
85.
by whoami_nr 3y ago
Author here. I originally wrote it on my blog (www.rnikhil.com) and cross-posted on team-bhp.com . Its not just a car forum but just enthusiastic folks discussing anything and everything. Like HN but based on cars and not startups. Somebody
86.
by whoami_nr 3y ago
Author of the post here. Your software is literally the reason I made any money in Poker. Must have studied using your software for more than a 1000hrs haha. On a related note, I made a small tool for PLO folks to study their preflop game.
87.
by whoami_nr 3y ago
Author here. Yes, the grind is monotonous and borderline depressing because you are playing a zero sum game. Regarding the money, after you do that for a couple years, money has very little meaning and value.
88.
by whoami_nr 3y ago
Author here. I also played PLO for a living. BTW, I made a small tool for professionals to study PLO. Check it out here: https://rnikhil.com/2022/06/15/gtoinspector-startup.html
89.
by whoami_nr 3y ago
Author here. You are right. Doing hard things just for the sake it builds your confidence. Relevant article: https://blog.nateliason.com/p/proof-you-can-do-hard-things
90.
by whoami_nr 3y ago
Author here. You are spot on about the career break. Luckily for me, during my time I played poker professionally, I was also consulting with poker website to build out their matchmaking and fraud detection algorithms. I also built a small

More ›