FTX – The fraud was in the code
newsletter.mollywhite.net
newsletter.mollywhite.net
And just what in the world were any of these people thinking? Sam Bankman-Fried isn't remotely charismatic, so I just don't understand why people would follow his demands and actively participate in such obvious fraud. How do you reason with yourself about randomly generating an insurance fund's balance that has been testified, under oath to Congress and various other binding contracts I'm sure?
Sam is highly charismatic, especially when all you want to hear is that you send 2000$ and will receive 100k$ if only you listen to him and use his product. He's part of the mumbo jumbo guru crowd.
I'm a programmer in finance: I understand nothing when SBF talks, it's like complete random BS, totally different from the traders I work with who would use complex concepts but with always a visible thread: utility -> transformation adding value -> profit, with SBF it's all so fuzzy and unclear, it's like he pretends to talk like we're idiots to hide the fact he's the idiot who forgot that you need to sacrifice potential upside for downside insurance...
I honestly struggle to understand how anyone would listen to him for any length of time.
Picked an interview ar random. He is not charismatic, but he can talk like so many banking VPs talk ( almost no substance ). This may have helped him make people part with money.
It appears that there is a sizable audience who will hear that and conclude he's a subject matter super genius and that you're a total idiot.
Many of them the titans of SV no less, Sequoia being chief among them; honestly, I called this imbecile being in jail in mid-late '21 when he announced he was going to start influencing US politics. He was an idiot then and he has proven himself and even greater one with time. Likely mentally ill, as well.
I do want to underscore that this scammer would have NEVER gotten the chance he had were it not for the connections his family has at Stanford, and thus VC, and the litany of connections that followed. This article is a must read [0], as were the testimonies given by gyus like Scarmocci (I'm not going to correct his name), because it all boils down to the fact that he was the child of privilisge and connected to all the power-brokers in SV.
SBF, FTX and Alameda is and will always remain a a Silicon Valley VC backed cluster***, and it must be noted that this has nothing to do with us and everything to do with THEM.
> Silicon Valley could be said to be in the business of reality distortion. Fundraising for startups can be as much about narrative as about economic fundamentals. Most venture capital portfolios are filled with companies that will fail because their model is wrong, their product won’t land, their vision of the future won’t pan out. The high dropout rate means that everyone is in search of the one thing that will reach escape velocity. Everyone is looking for an epochal success—a Steve Jobs, a Jeff Bezos. That creates a degree of hunger—even desperation—that can be exploited by someone who arrives with a great story at the right moment.
0: https://www.wired.com/story/why-silicon-valley-falls-for-fra...
It seems like if you were committing massive fraud, you'd want to avoid leaving permanent records of it on someone else's systems.
But I think this is generational thing. If you grew up when web browsers were already prevalent, a "text file on your hard drive" probably means almost nothing. It's just not the way you use computers.
I also like this bit from the article:
Note to self: if you’re going to write code to do fraud, make it messy and unreadable to reduce the chances it’s later put in front of a jury as evidence.
Of course, the other possibility is that at least having this amount of visible paper trail was intentional - the CTO probably had a little voice in the back of his head going "Umm, maybe committing multi-billion dollar fraud is a bad idea...", but the GitHub history probably helps him more than it hurts now, at the expense of SBF.
I'm not sure of that. When you need an account's credit limit set to $65 billion because the account (Alameda Research) kept hitting the credit limit, something is seriously, seriously wrong irrespective of the state of crypto. It seems like they were inevitably going to fail hard. FTX was literally allowing Alameda Research to trade with make believe money.
Some of the major traditional brokers will just up and close your account permanently if you abuse, even within their already strict rulesets, their margin accounts.
Yes, I had to give my professional opinion on why some healthcare software here was breached so I asked the company to deliver the source: this software is written from begin 2022. They don’t use any normal (git, cvs, sourcesafe etc) versioning system; just zip files with dates in them. So they sent the ‘latest version’ and we found it’s not the same version as the breached version on the server. So they are saying they don’t know who or what ‘changed the deployed version’.
He said he got a spreadsheet from a company, and also insisted he didn't doctor the spreadsheet. But he also admits he was the only person who handled the spreadsheet (presumably to not through his collaborators under the bus).
He simply says he doesn't know how to reconcile those facts -- that's "your problem" if you want to prove him guilty.
But since there's no version control, there's no smoking gun. His defense is basically "I don't know what happened but I'm innocent"
And plausibly some of the FTX defendants could have tried this, had they not left permanent records all over the place.
They could have worked with zip files with no provenance :-P
After all, the burden of proof is on the prosecution.
(though of course I'm glad they did leave records, and wish Ariely would have)
And of course this is the "quality software" that is charged an arm and a leg for
Would easily bet that was his decision. He's obviously made insanely stupid choices already, like leaking Caroline Ellison's journal and getting his bail revoked.
I'm sure he was counseled against it, but his lawyers were probably like "well, he wants to swing the bat, and we get paid more if it goes to trial anyway."
>Would easily bet that was his decision.
That's a tautology. His decision to plead not-guilty is literally his decision.
and decent lawyers would have told him to STFU and not do a bunch of things he did. and would have likely told him to plead out.
Prosecution have stated that they haven’t offered SBF a plea deal, so what incentive does he have to make their job any easier?
This was the supposed crypto wonder-platform!
Their actions brought in more victims, and therefore investors ought to be wary of listening to them in future. They ought to be also punished by a hit to their reputation i.e. by being held partially responsible.
They were lazy because they relied on the reputation of who was pushing for FTX, now their reputations won't be so rock solid. That's some justice.
Just because they stand to lose millions doesn't excuse their responsibility and influence. Venture capitalists have no one to blame but themselves for the risky companies out there. They are effectively snake oil salesmen, just looking to siphon off relatively short-term profits under the guise of "changing the world".
I see what you're getting at, but this is a bit of a silly statement. The obvious downside of "throwing millions at people" is that you lose millions of dollars.
The startup culture that has fetishized making it big at all costs (move fast and break things) has lead to customer (and non-customer) deaths amongst several other massive failures. The investors should have some responsibility and accountability assigned. Right now, they have zero.
But since it was the "wonder kid of crypto" (read, very well connected person) it was a reason for praise
Yes please explain me how he really merited that
They won’t be though, because everyone would be afraid for future repercussions. That said, they’re on a list of VC firms I would hesitate to take calls from going forward.
I know this will just get me labelled a conspiracy theorist but I refuse to believe his well-connected parents had nothing to do with the praise FTX received in the mainstream media.
Also known as a company.
> web2 company
Are web3 companies less fraudulent?
Completely untrue, of course. There's always a real-world link(because otherwise nobody would be profiting(people want stuff)/nothing would actually happen)
Crypto scams are always off chain (like FTX and many more) or have smart contracts that can be upgraded by anyone (but again, anyone can check that the contracts can be upgrade and by whom).
Hmm in fact dont use git at all.
Law enforcement only need to find a single developer pc/laptop/whatever and they will have the entire history, comments, who did it, and code.
Having it all in a centralized system that is configured to be easily nuked is safer, than trying to nuke every laptop/pc/whatever that has a copy.
Of course, the entire code base will probably be available on any laptop/pc/whatever that LE can seize, but it is less data than with git.
I wonder if you could do the development all on ram drives on the development machine, that way if you shut down all the code is gone. Just make damn sure to stow/check in your code before doing so.
I have partially done this myself a several years ago, but that was only to try and speed up compile times.
However, it seems that the FTX people were not really bothered to cover their tracks.
or if you have any downtime you lose all of your shit
May have to be a bit more dilligent but not a big change.
https://www.ycombinator.com/ Is every number on that page real time accurate? Is the combined value $600bn - if it is 599bn or even 601bn then ding ding ding we have ourselves fraud.
https://en.wikipedia.org/wiki/National_Debt_Clock - Can I get any and all people behind this (and Similar Projectes) jailed for fraud?
Certain customers (including Alameda) - I understang other customers had similar arrangements, and why wouldnt they? Every account was trading on margin and high leverage - the reason you auto liquidate is so they dont run off owing you money.
Is it fraud for a company to extend credit to another company? Any company that issues a invoice and allows say 14 days to settle is extending credit to the other party.
I will at least say, "goes to character", is absoultely an excellent and as I see it the only logical answer. If however, they have to play that game and that is all they have then they are going to struggle to make a legitimate case.
Maximum entertainment value from the sidelines will be not guilty with all the other FTX employees having plead out!
We are not even one week into a six week trial. This particular fraud is only one small part of their overall case.
I haven't been following the trial though, and I'm unsure of how much testimony was given. What's reported on this blog does seem like it is a little small given Gary Wang's large role.
He did not testify that the number on the website was real. He testified to the existence of the insurance fund and the single biggest daily drawdown from the fund.
"which by itself is a felony" - wouldnt you mention that as the prosecution??? Or are they aiming for a mistrial on the grounds of incompetence?
> Prosecutors haven’t mentioned it, but Sam Bankman-Fried would go on to testify under oath in front of the U.S. Congress in May 2022 that “the insurance fund has paid out a net total of $9.5 million” in the preceding three years, and that “the single biggest daily drawdown from the FTX.com insurance fund was $4.7 million.”
Actual testimony. https://www.congress.gov/event/117th-congress/house-event/11...
> Finally, prosecutors questioned Wang about the FTX “insurance fund”, which was ostensibly supposed to protect both FTX and its customers from trades that went badly even more quickly than the exchange’s risk engine could account for.
> One such example was in 2021, when a trader was able to exploit a bug in FTX’s margin system that allowed them to take out a massive position in the MobileCoin cryptocurrency. They were eventually liquidated, and FTX suffered a loss of “several hundred million dollars,” according to Wang.
The losses to the exchange from a failure of the code should be borne by the exchange not from any "insurance fund" whos purpose is the make good slippage losses for customers. There are grounds to ask why they passed the loss off to Alameda (in which case FTX co was defrauding Alemada co for an FTX mistake).
I see the narative that the prosecution is trying to push, but unless SBFs laywers are idiots they should be pointing out that they are trying to connect two distinctly different dots.
The "listen to the article" feature on this blog post ... is it using an AI voice trained specifically on the author's real voice? Or is it a real recording by any chance?
if not account.allow_negative:
if (balance.available_ignoring_collateral if ignore_collateral else balance.available) < 0:
raise BadRequest('Account does not have enough balances')
ignore_collateral ? balance.available_ignoring_collateral : balance.available
https://docs.python.org/3/reference/expressions.html#conditi...
1) even before the speedups done now on CPython, PyPy has always existed (and NumPy was HW-accelerated)
and more importantly 2) It has a native arbitrary-precision decimal type. Even other languages seriously used in financial institutions like OCaml (https://discuss.ocaml.org/t/ocaml-needs-an-arbitrary-precisi...) doesn't even provide it natively.
The benchmark would need to be Java or C++, which are the mainstay of the industry.
There’s always some process a firm must use to make digital records available to an auditor - it’s at this point in the step code must be trusted even if it’s supposed to be as simple as “make everything available “.
The "step code" here is a group of humans adding documents to some type of storage available to the auditor.
There is no question of fraud in that code because it's generic. Microsoft, Box, Dropbox, etc. aren't building in custom code to commit fraud.
When publicly-audited firms commit fraud, it's not an issue of untrustworthy code. It's because humans are going to some trouble to commit fraud. There are even "checksums" (balances) that tell you if something is accidentally missing, so it's rare that financial fraud can be conflated with an innocent mistake.
int((2^16 - 180) * 999999.9999 + 1) = 65,355,999,994