This underscores a major issue with commercial software like Steam and games obtained therein. You cannot trust that software not to maliciously scan your device for secrets such as crypto wallets or other information. Ideally, you want to run apps like Steam, Discord, Zoom and whatever else does not come from a trusted distribution repo as a separate user. This is not always convenient however, and the compromise I adopted on my gaming PC is to bubblewrap Steam. Do not mount things like /home and devices games should not be using in the sandbox.
In theory you could change the steam command line fpr each game and have them launch in bubblewrap. Though I'm not sure how well that would play with graphics drivers. It might be friendly with AMD, but my experience is that NVIDIA is a pain to have play nice with containers.
I sandbox Steam itself for simplicity. My experience with Nvidia and Wayland has been good. I simply exposed /dev/nvidia*, a few networking-related files in /etc and system binaries like /usr.
Dedicated hardware is definitely the most secure option, but realistically it is even less convenient than a separate user. You can't realistically play graphically intensive games on a Deck, and separate workstations are very unaffordable at the moment.
Android solves this problem by making it impossible for an app to access the storage of another app. No matter how much the Steam app wants to see your crypto wallet's key it is not possible (assuming no Linux exploit).
That's nice, but a cell phone isn't a viable gaming platform for anything but very basic games. And Android, at least in its typical Googled form, has its own set of problems; it's a privacy nightmare.
Modern macOS devices have a secure element just as good as iOS devices. Almost nobody makes use of it, though.
(I haven't seen anything yet that uses the secure element to control access to a wallet. Are there any technical obstacles to keeping the keys out of main memory?)
I heard Monero can help against blockchain analysis (though less so if you buy from adversaries). Unfortunately it looks like blockchain analysis was not the attack vector here, but rather buying gift cards through an incompetent intermediary.
Not really an "incompetent intermediary", Bitrefill, the company used for buying gift cards with Bitcoin received legal requests to identify the user. Bitrefill doesn't proclaim insane privacy protections, they are based out of Sweden but aren't immune to being requested to provide the information.
Monero could of helped, although its not easy to transfer BTC to Monero without P2P trades, as effectively every exchange requires KYC. Those source and destination wallet addresses are tainted, you wouldn't be able to deposit on any mainstream exchange and if they do, its for a honey pot purpose.
Making it possible to identify a user is incompetence, exactly because of legal issues like this. You don't run a service that accepts crypto for gift cards and then keep information on file that identifies the buyer personally, that's just stupid. But then again, you don't illegally obtain crypto and then give out your info when you go to use the illegal funds, that's just stupid as well.
Is P2P that big a deal? I swap coins on Bisq all the time. It's super easy to pick up and use if you already have the BTC. The biggest obstacle is obtaining some BTC in the first place when you don't have any yet. Once you have enough for a deposit, you can buy/sell or swap thousands of dollars in funds at a time.
What do they do, _exactly_? Because in the real world out of YC, the rest of us can barely top $200,000 a year unless we move into management or Executive level roles.
A silver lining is that at least a lot of these AI generated malware and phishing scams are easy to disrupt with mild white hat skills. Often the people behind them lack the knowledge on how to prevent themselves from being disrupted by said white hats.
The most laughable I've seen was a scam that put a captcha on the frontend to stop me from flooding their free email system on the backend and rate limit them. However, the captcha was clientside only, so I didn't even notice until I looked to see if they rotated emails, but also the captcha was a textbox with the code set as the placeholder.
This underscores a major issue with commercial software like Steam and games obtained therein. You cannot trust that software not to maliciously scan your device for secrets such as crypto wallets or other information. Ideally, you want to run apps like Steam, Discord, Zoom and whatever else does not come from a trusted distribution repo as a separate user. This is not always convenient however, and the compromise I adopted on my gaming PC is to bubblewrap Steam. Do not mount things like /home and devices games should not be using in the sandbox.
In theory you could change the steam command line fpr each game and have them launch in bubblewrap. Though I'm not sure how well that would play with graphics drivers. It might be friendly with AMD, but my experience is that NVIDIA is a pain to have play nice with containers.
I sandbox Steam itself for simplicity. My experience with Nvidia and Wayland has been good. I simply exposed /dev/nvidia*, a few networking-related files in /etc and system binaries like /usr.
This is what my Steam Deck is for. There's no gap like air gap.
Dedicated hardware is definitely the most secure option, but realistically it is even less convenient than a separate user. You can't realistically play graphically intensive games on a Deck, and separate workstations are very unaffordable at the moment.
Android solves this problem by making it impossible for an app to access the storage of another app. No matter how much the Steam app wants to see your crypto wallet's key it is not possible (assuming no Linux exploit).
That's nice, but a cell phone isn't a viable gaming platform for anything but very basic games. And Android, at least in its typical Googled form, has its own set of problems; it's a privacy nightmare.
Criminals should have used Monero.
Victims should have not keep crypto on desktops. They are much less secure than Mobile.
Modern macOS devices have a secure element just as good as iOS devices. Almost nobody makes use of it, though.
(I haven't seen anything yet that uses the secure element to control access to a wallet. Are there any technical obstacles to keeping the keys out of main memory?)
> Victims should have not keep crypto on desktops. They are much less secure than Mobile.
Unless it's Qubes OS.
True. How viable is steam on Qubes?
I heard Monero can help against blockchain analysis (though less so if you buy from adversaries). Unfortunately it looks like blockchain analysis was not the attack vector here, but rather buying gift cards through an incompetent intermediary.
Not really an "incompetent intermediary", Bitrefill, the company used for buying gift cards with Bitcoin received legal requests to identify the user. Bitrefill doesn't proclaim insane privacy protections, they are based out of Sweden but aren't immune to being requested to provide the information.
Monero could of helped, although its not easy to transfer BTC to Monero without P2P trades, as effectively every exchange requires KYC. Those source and destination wallet addresses are tainted, you wouldn't be able to deposit on any mainstream exchange and if they do, its for a honey pot purpose.
Making it possible to identify a user is incompetence, exactly because of legal issues like this. You don't run a service that accepts crypto for gift cards and then keep information on file that identifies the buyer personally, that's just stupid. But then again, you don't illegally obtain crypto and then give out your info when you go to use the illegal funds, that's just stupid as well.
Is P2P that big a deal? I swap coins on Bisq all the time. It's super easy to pick up and use if you already have the BTC. The biggest obstacle is obtaining some BTC in the first place when you don't have any yet. Once you have enough for a deposit, you can buy/sell or swap thousands of dollars in funds at a time.
He was smart/determined enough to risk it all. Why didn't he just try find a job that pays $200k/year ? Why play into the stereotypes ?
> He was smart/determined enough to risk it all.
Desperate people are willing to risk it all, and usually smart people aren't desperate.
>Why didn't he just try find a job that pays $200k/year ?
Probably was not able to get into MIT or another prestigious school that many of those companies look for on a resume.
None of my peers went to any school of note, nor did I. A fair bit of community college, some lower ranked state unis, and one guy did a bootcamp.
All well above 200k.
What do they do, _exactly_? Because in the real world out of YC, the rest of us can barely top $200,000 a year unless we move into management or Executive level roles.
It is honestly much easier to spread malware or hack companies than it is to make 200k/year from a company.
Now with AI, you can accelerate this so much. The only thing holding back thousands of new threat actors from doing the same is Opsec, its hard.
A silver lining is that at least a lot of these AI generated malware and phishing scams are easy to disrupt with mild white hat skills. Often the people behind them lack the knowledge on how to prevent themselves from being disrupted by said white hats.
The most laughable I've seen was a scam that put a captcha on the frontend to stop me from flooding their free email system on the backend and rate limit them. However, the captcha was clientside only, so I didn't even notice until I looked to see if they rotated emails, but also the captcha was a textbox with the code set as the placeholder.
Truly the pinnacle of vibecoding a scam.
> didn't he just try find a job that pays $200k/year ?
you really think it's that easy?
Claude get me a job that pays $200k/yr. Make no mistakes.
Claude get me a job that pays $200k/yr. Make a breakthrough.
Yeah, clearly risking it all wasn't the smart play here, as evidenced by the fact he got caught.
A lot of people are willing to tolerate risk, that hardly makes them useful. Frequently the opposite.