The more I read about this, the more Im absolutely dumbfounded that Sequoia poured money into them. It took an outside executive 1 day to find out how effed their balance sheet was. Did Sequoia not perform any due diligence?
One interesting story that's gotten a little lost in the chaos is that after Sequoia invested in FTX, Bankman-Fried also invested $200M into two Sequoia funds:
doesn't make sense. if this guy invested 200M USD into Sequoia and then Sequoia invested 150M into FTX. This is not even breaking even and also washtrading.
Exactly lol. Way too many people are getting hung up on how sophisticated VC / funds could have invested in FTX/SBF. We often forget these "investors" want above all an exit plan where they get their money back and more. They couldn't give a shit if the company is run with a balance sheet drawn in mspaint with the spraypaint tool by teenage ponzi schemers as long as they know they get the exit before it topples down.
Giving them the benefit of doubt, this is not a contradiction. The statement means that they have enough illiquid assets to cover the withdrawal that they are working on converting into liquidity.
FTX/Alameda holds tons of illiquid FTT tokens that they cannot sell and which Binance was dumping. Thus, they might have not technically lied. But they were still wrong from the accounting perspective - they surely understood that FTT token cannot be used to cover gaps in large scale.
It was the question that matters "how fast you can process user withdrawals and with what risk"
Sam owns 8% of Robin Hood that is worth around ~$1B - he could sell that and cover some of the gap. But what we do not know yet is the size of the gap in time and space. FTX had $6B withdrawals pending on Tuesday.
> Sam owns 8% of Robin Hood that is worth around ~$1B - he could sell that and cover some of the gap.
Not knowing too much about this space have you ever seen anything like this happen? A CEO using their personal wealth to cover their customers funds seems unlikely.
That can't happen directly (I think).
Mingling breaks the limited liability mantle.
He will have to buy equity or other product from FTX to infuse with cash. And if it is going bankrupt he has to stop running it to preserve the liability "shield" -- details of course depend on the country/state of organization/incorporation. I know nothing about the Bahamas.
Example: Elon for instance infused his own cash into Tesla, when it was going bankrupt. But he practically bought equity to my understanding.
Covering customer accounts is different, in a financial firm. I do not know what tools the Bahamas give to FTX. This is all uncharted territory. (Also there is legal exposure to other countries.)
This isn’t a thing, limited liability (which is the fundamental principle that distinguishes corporations from partnerships) prevents this. The only way would be if SBF was charged with defrauding FTX
But they're meant to store the customer assets in a cold wallet. They're not meant to invest them in illiquid assets that would need to be liquidated to give people their money. If it's not a contradiction, it's an intentionally misleading statement to avoid admitting they let Alameda Research invest the money when the entities are supposed to be completely separated.
In that scenario they could point to some of the wallets to help calm the fears the money isn't available. Or they could approach a number of different lenders who would be comfortable lending at high interest rates if the money is there but slow to access. They only sell if we're in the non-charitable case.
I thought FTX Alemeda (the trading side) invested into their own token, which is tanking thus causing problems.
Alemeda has like $14b assets and $8b in liabilities. But of that $14b, $5b are in their own token (FTT) which is kinda?? worth nothing at this very moment. So now the assets and liabilities are more equally matched, but less margin for shifting values of tokens.
I don’t know, but the derivative of their assets looks scary the last 24hr.
Illiquid just means that you can't cash it in quickly.
Cash is 100% liquid while a house is illiquid you might have millions US$ parked there but only if you manage to sell it, then you convert it into liquid cash.
Liquidity is a measure of how easy it'd be to trade a thing for another thing you want.
Agreed, but in this case these tokens have suddenly become 'illiquid' because they have just become worthless. It seems that here the term 'illiquid' is being used as a euphemism.
But cold storage and locked funds lead to you taking out a loan or just saying "sorry, it's going to be a while, our funds are locked". You know what you don't do if your funds are illiquid? Sell to your competitor over night.
There's nothing decentralized about FTX or Binance. They operate in an opaque manner like any traditional business, transparency comes from forced audits
& regulation.
Decentralized finance is built on chain where all assets are publicly auditable at all times.
EDIT: parent comment talked about decentralized finance, then edited to remove mentions of defi
>literally centralized companies as "not-decentralized"
So defi is only 100% decentralized everything, even if the financial tools are decentralized? That feels like an appeal to purity if ever there were one.
Ok, my previous comment is a bit unclear about what I mean. In my opinion if there is a group of people, other than the participants themselves, who can control the operation of the financial service then it is not decentralized. There can absolutely be an organization that builds the service, but participants should not be forced to adopt updates and should be free to transfer their entire balance to any other service at any time.
I realize I’m on the fringe a bit with this but I think it’s not because I have an extreme idea of what defi is, it’s that there have been so many grifters in the last 5 or so years that have used the buzzword “defi” to sell their shitty reincarnation of long-outlawed shady centralized financial schemes as something revolutionary that it’s shifted the public perception of the term. I’d even agree with you that it’s an appeal to purity.
Note that FTX.us is regulated under some US licenses and is unaffected. What was blown up was FTX.com operation that is licensed and regulated in Bahamas.
The FDIC is just a ruse to let "useful idiots" think that everything is okay. In reality, the FDIC charges banks 90% less than the actuarial value of the risk they take on, and banks make wildly risky loans/bets all the time, knowing it's "heads I win, tails the taxpayer loses."
Insofar as you can call US Finance any better than crypto, it's because of socialized losses. IMO, bank failures are a much more appropriate solution.
I can tell you a bank like say JP Morgan Chase, who is charged 5bp a year (i.e. 5 cents for every $100 dollars), has a much higher chance of catastrophic failure than 1 in 2000. Many banks just like them fail every few decades, and it was generous of me to only say they're undercharged by 90% (i.e. 1 in 200 odds), when the reality is probably more within a range like 1 in 20 to 1 in 100.
The insurer is the United States government. They take losses on things all the time. It's called "socialized losses." I referred to it before, and it sounds like you don't even understand these finance 101 (or even basic high school civics) topics, so why are you insulting anyone?
The insurer is a corporation with its own financial statements, so it's pretty easy to see if it's operating at a loss (and thus subsidising the banking industry) or at a profit (not subsidising it). I guess you didn't know that either.
In case anyone is curious, here is what some of my research has found:
The empirical rate of bank failure in the last couple decades has been slightly over 1 in 250 banks per year (that is, ~0.4%/bank/year, or "40 basis points"). This is from these two sources: https://www.fdic.gov/bank/historical/bank/ says that on average 27.3 banks per year have failed, while https://banks.data.fdic.gov/explore/historical?displayFields... says that there have been ~6500 banks covered. (I think that the probability of a massive bank failure is in fact higher than the empirical rate, due to the tail risk of catastrophic failures.)
I have not been able to find what rates JP Morgan Chase pays for their deposit insurance, but I think this page https://www.fdic.gov/deposit/insurance/historical.html suggests that the rate is between 1.5 and 40 basis points per year. Some other sources I've found do suggest that the average rate is around 5 bps/year.
Already we see that the empirical failure rate is higher than the assessment rate. (Although note that the probability was not weighted by dollars, whereas the rate is.) This is perhaps surprising, because the FDIC claims that "The FDIC receives no Congressional appropriations - it is funded by premiums that banks and savings associations pay for deposit insurance coverage." https://www.fdic.gov/about/what-we-do/index.html Perhaps this is part of the point of this comment I am replying to.
As a result, it's plausible to predict: (a) the deposit insurance fund might go negative again (ie, the insurance rate is incorrect), (b) the deposit insurance fund will definitely go negative in a situation like the S&L crisis or the 2008 financial crisis (thus requiring tricks like the borrowing mentioned above), and (c) in the event of a more catastrophic failure, the insurance fund will go so far negative that it might be explicitly bailed out by the broader federal government.
I’m a little ignorant to the whole crypto ecosystem, so can someone give me a quick rundown on the chain of events that led to this? Seems a little out of left field.
BTX, from the outside looking in, looked to be one of the more well run, stable crypto exchanges. $1.02 billion in revenue with $388M in net income in 2021. They didn’t go on any crazy hiring spree when they didn’t have to. Liquidity crisis implies that people are withdrawing cash they do not have, but if so, where did it go?
FTX bet customer funds through the CEO’s hedge fund on an FTX token [1]. The token price fell when this was revealed [2].
The hedge fund, and thus FTX, had less money than they owed lenders and customers. FTX found a bail-out in Binance; otherwise everyone would have lost their money.
It isn't officially a bailout. They just said they intend to acquire FTX. But, once they look at the books they may backout out the deal, especially if most customers want to take their money out. What is the point of buying an exchange that has no customers?
FTX's ceo also runs a prop trading firm which is the real source of his wealth. FTX loaned the trading firm billions of its own token FTT. FTX also gave binance billions of dollars of FTT because binance invested in them. Over the weekend FTX's ceo and Binance's ceo got in a fight on twitter and binance sold all of their FTT which collapsed the price. FTX's assets are tied up in the loans to their trading firm which are denominated in the now essentially worthless FTT and so they can't convert the FTT to cash which means can't process withdrawals.
If history is any indicator, starting a company during or shortly after a recession has yielded some massively successful results. Not only does it crowd out parties that would enter easily during good times, you also have access to incredible talent on the market due to a high supply. Yes, access to capital is definitely more strained, however PG has said that technology progresses more or less independent of the stock market and economic cycles.
All this to say, I’ll probably look to start something. Anyone else down?
Terraform (as you may well have already found out) is a blessing and a curse. It's great for instantiating infrastructure quickly and repeatedly, but what it's not great for is managing it going forward and streaming changes (it seems to break A LOT when there is drift from managed infrastructure vs. real life infrastructure).
Do you guys plan to handle managing instances created via terraform code via Gallery, or is it strictly a tool to instantiate environments and add triggers around it?
Like I'm thinking of cases where you might want to edit your existing env created by Gallery with other stuff that other teams may have worked on and merged onto prod. If I edit the config, there could be a lot breaking changes
Absolutely - managing and syncing the state of IaC tools is core to the value of Gallery. Infrastructure templates can be edited without changing currently-active environments, and this is handled gracefully right now. Propagating template updates to running environments seamlessly is also on our roadmap.
Thanks for the reply. It certainly will be an interesting technical challenge to try and propagate changes efficiently. Speaking from experience (I've worked with Terraform a lot the past 2 years), it can be a tough nut to crack, so good luck!
I would add one more thing (and this is maybe for down the line): There's a use case for duplicating envs beyond just development environments. There's a lot of value in cases where the SaaS product itself requires on-demand environment generation. Where customers of it need staging instances so they can see what config changes are before they merge to production. Obviously in that case, you're treating an actual product like a terraform provider, and the API's of that product as resources. I would explore that down the line because there's a lot of value in a service like that.
FYI, I would love to get involved any way I can to help you guys grow this. Let me know how I can help! mailto:ramish@ualberta.ca if you're interested :)
As you may already know, integrations are the heart and soul for products like this. I'm assuming you're already being bombarded by potential/current users asking "when will you have integration X?".
What is your strategy to scale out & maintain integrations? Speaking from experience, it's not something that is easy to scale out unless you have a dedicated team whose job is to build them out, or you have some third-party provider like CData providing OOTB connectors for your product.
(On a side note, this looks fantastic. Are you hiring any product folks per chance? I have significant experience tackling this same problem).
You're spot on. One thing we're doing to deal with the long tail of integrations is opening up an API to customers.
We're also considering open sourcing that part of the product, but haven't made a firm decision on that yet. Would love to chat if you're open to it. We're definitely looking for people in product. Feel free to send me an email to etai@secoda.co
Uber I can still believe due to their complex app (real-time tracking of drivers and riders, matching algorithms, route mapping, pooling, fare calculation, surges). Not to mention other divisions like food delivery, autonomous driving, freight etc.
AirBnb is...a database of static listings and a search engine on top. Their top technology problems were all solved by Expedia like 20 years ago.
Uber is not that difficult. There's work for hundreds of developers between the few products, the localization for tens of countries with different language/currency/culture, and maintaining the main applications in working condition 24/7 at global scale. That's about it.
They have thousands of engineers sitting idle or making pet projects, struggling to justify their existence, as regularly attested by internal employees when an article pops up.
