Who To Sue When a Robot Loses Your Fortune (bloomberg.com) 201
An anonymous reader shares a report: It all started over lunch at a Dubai restaurant on March 19, 2017. It was the first time 45-year-old Li, met Costa, the 49-year-old Italian who's often known by peers in the industry as "Captain Magic." During their meal, Costa described a robot hedge fund his company London-based Tyndaris Investments would soon offer to manage money entirely using AI, or artificial intelligence. Developed by Austria-based AI company 42.cx, the supercomputer named K1 would comb through online sources like real-time news and social media to gauge investor sentiment and make predictions on US stock futures. It would then send instructions to a broker to execute trades, adjusting its strategy over time based on what it had learned.
The idea of a fully automated money manager inspired Li instantly. He met Costa for dinner three days later, saying in an email beforehand that the AI fund "is exactly my kind of thing." Over the following months, Costa shared simulations with Li showing K1 making double-digit returns, although the two now dispute the thoroughness of the back-testing. Li eventually let K1 manage $2.5bn -- $250m of his own cash and the rest leverage from Citigroup. The plan was to double that over time. But Li's affection for K1 waned almost as soon as the computer started trading in late 2017. By February 2018, it was regularly losing money, including over $20m in a single day -- Feb. 14 -- due to a stop-loss order Li's lawyers argue wouldn't have been triggered if K1 was as sophisticated as Costa led him to believe.
The idea of a fully automated money manager inspired Li instantly. He met Costa for dinner three days later, saying in an email beforehand that the AI fund "is exactly my kind of thing." Over the following months, Costa shared simulations with Li showing K1 making double-digit returns, although the two now dispute the thoroughness of the back-testing. Li eventually let K1 manage $2.5bn -- $250m of his own cash and the rest leverage from Citigroup. The plan was to double that over time. But Li's affection for K1 waned almost as soon as the computer started trading in late 2017. By February 2018, it was regularly losing money, including over $20m in a single day -- Feb. 14 -- due to a stop-loss order Li's lawyers argue wouldn't have been triggered if K1 was as sophisticated as Costa led him to believe.
It's obvious (Score:4, Informative)
Sue the owner of the robot. It's not that hard.
Re: (Score:2)
No.... you sue the one who trusted that fortune to be used by the robot.
And in many cases, that would ultimately fall squarely on the person whose fortune was lost - unless they never consented to allowing their fortune to be used by the robot's owner in the first place.
Re: (Score:2)
The flaw was in their testing versus real life investing. In testing, only pretending to play stocks, the algorithm can not take into account it's own trades and their impact. Sure for the little guy, their trades still will not have much impact but with larger transactions so the greater the impact. So the algorithm ends up chasing it's own tail, chasing it's own trades, it's own trades and their impact creating targets to chase.
Why you do not trade on patterns, people are greedy and stupid and will alway
Re: (Score:3)
Sure. But the things they are mentioning, the AI not being as good as they thought, poor back testing, etc are all examples of reasons they shouldn't have invested and poor due diligence on their part. They made a poor investment and lost money, it happens millions of people every day and for those who don't lose, this is where the gains come from.
Should we for some reason have sympathy when they wealthy lose money on poor investment choices when nobody gives a damn if you we take losses in our retirement
Re: (Score:3)
the algorithm can not take into account it's own trades and their impact
Sounds like a very small market, then.
Re: (Score:2)
Or possibly, entering a market that has a standard deviation of over 10% with 10x leverage isn't really a calculated investment.
Re:It's obvious (Score:5, Insightful)
As it should. Guy makes bad investment, losses money, news at 11.
Nobody would give a damn if this were a poor working man convinced by a trustworthy seeming person at a church luncheon. Why should someone who supposedly has $2.5 billion times the merit not have to face the consequences of investment decisions?
Short answer, they shouldn't.
Re: (Score:2)
Friendly reminder, the investor didn't make his own money. The guy suing inherited his money from his daddy, who owns billions of real estate. But what are the odds someone inherited billions in real estate would make bad investment decisions?
Re:It's obvious (Score:4, Insightful)
Poorer than you'd think. The default would have just been to keep investing in real estate or even to hire a handful of qualified people to do it for him. He had to deliberately and foolishly change course for something like this.
Just about anyone more conservative than "day trader" would have a hell of time failing to make a profit on $250m that beats inflation plus any sort of sane living expense (which hanging in Dubai certainly doesn't indicate).
There are literally at least a half dozen people in any suburban block that could have made him millions a year with diversified low risk investments. It's easy to earn passive returns of 3-6% at reasonably low risk, 6-9% at medium risk, 9-15% at high risk, 15-30% at very high risk, anything greater is insanely high risk. What is challenging is making those returns at a lower level of risk than I just stated. The secret is nothing more than already having wealth and not assuming much risk.
If you only have $500 turning that into something substantial is high risk gambling because you need triple if not quadruple digit returns to gain any real advantage and the risk level is off the charts. At $250m averaging 5% annually gives $12.5m/yr, retaining half for inflation and growth that is still $6.25m/yr to blow on a ridiculous lifestyle. Like I said, at least a half dozen people on any block could accomplish that after 30 minutes of google fu. With actual intelligence and effort you'd have the same $6.25m/yr lifestyle and $37.5m/yr growth. It's a hell of a lot easier to keep $250m than make $250m, its also about as easy to turn it into $500m as it is to turn that $500 into $1000 via investment.... which is the only reason fools like this lose it at all. They get greedy and take too much risk. People who have $500 (or even $100,000) aren't so much being greedy as making a desperate gamble to turn a toehold into a foothold in the game... of course if they get it, most of them THEN get greedy because they got lucky the first time.
Re: (Score:2)
Re:It's obvious (Score:5, Insightful)
"*of course general rules of liability when investing also apply here. For example, if someone invested (for you) in a stock that crashed and he did so without any malice or sheer incompetence then he doesn't not owe you something. Unless he guaranteed an X% return or that he you want lose your investment (in writing/contract of course)."
Exactly. He is out the money and those who trusted him to manage their money are out the money. Nobody is liable for anything except their own losses.
Re:It's obvious (Score:4, Insightful)
You know someone who'd take my losses while I get to keep my profits? I mean, outside of corporations using governments as their fall guy?
Re: (Score:3)
of course general rules of liability when investing also apply here.
Which in this case (and all others) begins with the actual agreement between the parties involved, the specific details of the contract.
Investment agreements are able to disclaim an enormous amount of liability. The machine creator could be liable if the machine were shown to be defective and didn't operate as was described in the agreement.
From the bits I've read, the investor knew there was risk on experimental trading algorithms, the investor specified both an amount of money to invest with the algori
Re: (Score:2)
Re:It's obvious (Score:5, Insightful)
"For example the robot could have been used without the owners consent.
It could also be that the owner have been misled about the robots capabilities from the one he bought it from."
Consent only matters to the extent that the person deciding to invest had power of attorney and liability would rest with him if he abused that. There isn't necessarily weight to "well if I had known he was going to do THIS"
As for capabilities. Maybe a tech deficient judge will buy it, or an armchair technical judge get dazzled in buzzwords. Unless the bot objectively had some hard capability they explicitly said it had the answer is a solid no. An AI is not a simple rules based robot, it operates on a neural net and its behavior in response to not yet seen conditions can't be predicted, it's behavior "capabilities" can not ever be guaranteed except to the extent you impose hard limits on the impact of its output in the overall decisions.
Something like they are discussing with insufficient back testing certainly doesn't qualify... those are just reasons the bot sucked and the bot sucking is a reason they made a poor investment not a reason it was a scam. It's a given that in modeling he showed cherrypicked benchmark periods, etc. Everyone does that from Schwab to Lynch to guy trying to sell his vending machine company, to Intel trying to sell their latest graphics chip. It's called sales.
No sympathy here. If I make a bad investment I have to cover my own losses. Why shouldn't they? Both citigroup in trusting this man to invest their money and this man in trusting this AI peddler should have done better due diligence. As for the AI peddler, he certainly did nothing more wrong than a mortgage broker and is equally entitled to his commission.
Risk and Reward (Score:4, Insightful)
I doubt this will go anywhere as the normal blurb is that the value of your investments can go down as well as up, and risk is proportional to reward. It seems like they were expecting magic to happen with no expectation of risks.
Re:Risk and Reward (Score:5, Funny)
One would have to be a complete moron to trust an investment banker called "Captain Magic".
Re: (Score:2)
Or "Made off" (Bernie)
The red flag was obvious here (Score:5, Insightful)
Re: (Score:2)
Re: (Score:2)
Lots of very smart investors [sic?] didn't have $100M+, but can make lots of money for people with $100M+ and begin to amass a fortune for themselves as well.
Are you talking about inventors or investors? Because tomhath post is about inventors, not investors.
This is obvious (Score:3)
Before I get into it, you're assuming they're exclusive. I assume he also invested a large chunk of his own money in it. For obvious and good reasons, you wouldn't invest 100% of your money in a single strategy unless you had to, even if it had a 99% chance of doubling every day.
So he took outside money. He did so for the same reason that cash-positive companies take investments. B
Re: (Score:2)
Re: (Score:2)
If the robot worked that good he'd destabilize the market beta in about 6 hours and the AI's predictions would no longer function due to destabilizing the market. No one makes that kind of money in the market except the market making firms taking percentages of every trade.
When something is too good to be true, it generally is.
Re: (Score:2)
I assume he also invested a large chunk of his own money in it.
This wasn't investing, this was gambling -- it was more about how the AI would manage the money. Never gamble with your own money or money you can't afford to lose -- and certainly don't borrow it. The guy's an idiot and owes Citigroup $2.25bn + interest.
Re: (Score:2)
Well, the hedge fund owes Citigroup, the fund didn't go bust (it only lost ~25MM dollars) and what the investor got from his daddy has been ~$5billion in gifts so far and more when daddy dies. I think the investor will be fine.
Re: (Score:2)
Yes that is how it should work and how it does work for people who have lots to lose. For the very wealthy whose losses are just numbers on paper well I'd be surprised if $2.5b doesn't magically buy at least some retroactive hedging.
It sounds like they intend to dazzle a judge with technology much the way the AI seller dazzled them. Insufficient back-testing... that is evidence of a lack of due diligence on their part (it might be lack of due diligence on the part of the guy selling the AI depending on whet
Sue who? (Score:2)
Unless the deception is obvious (and based on fraudulently promoting objective features in a product that did not exist), there's no grounds to legitimately sue (at least not and win under most legal systems.)
In this case, I think it's more an effort to save face than it is to air a legit grievance over lost funds. Only an idiot would trust anything touted as "AI" this damned early in development...
Re: (Score:2)
Actually, I wouldn't be surprised if this was just an elaborate con and no AI or supercomputer ever existed.
A Fool and their Money... (Score:5, Insightful)
There is a reason "most" people are poor. Most people are going to see this as an attack but that is not the intention. Go and read about distribution and the Pareto principle and start applying it to everything.
Then look at the psychology people have and their "savior complexes". Most people are looking for saviors. This is why people are usually religious... not because they actually seek the truth of a religion but the "salvation" being offered. Now juxtapose that against Government... same thing... we promise salvation if you "follow us". As governments grow more nanny more people leave religion proving my earlier statement. They no longer feel a need to pray to a faceless entity that could be their own imagination.
The moment you subject yourself to the rule of others, they now control at least some if not a significant portion of your destiny/future. It is shocking the number of people that exist that fail to grasp this one simple fact. This guy must now rely on a savior, if one is willing (government and courts are fickle), to reclaim his lost money.
Re: (Score:2)
You just confirmed what the GP drew a generalization around and gave us insight to.
People are fools for trusting in religious or governmental salvation. They willingly hand over control of their lives and resources for vague hopes and promises. The GP explained "why" in a larger context.
The story is as long as human history and is not over. But, from your response, I'd say that you are prone to repeat the mistake.
Re: (Score:2)
Plus the aqueducts. And the wine. law and order, obviously. But apart from that...
Everyone can be rich! (Score:4, Insightful)
This robot really illustrates how ridiculous large (probably most) parts of the financial industry have become. I mean, if analysts and traders were actually producing useful economic value, then as this trader robot became better at its work and cheaper to run, we could all eventually own trader robots and sit back while our bank accounts are filled with money.
Of course this is not going to work, because someone must actually clean the toilet and make the coffee. So why do we then believe that the continued rampant growth of the human run financial sector is going to make us all rich? This is the insanity of deregulating financial markets, and then propping them up with endless cheap funds (ZIRP, QE), in the hope that it is going to lead to prosperity. It is just making a few people rich and everyone else into debt slaves.
Re: (Score:2)
Actually, Everyone CAN be rich. The problem is that most people do not want this and actively work against it. The "super rich" want to make sure they stay that way so the objective is to have as much money as they can get to keep the power that being super rich gives them.
"Of course this is not going to work, because someone must actually clean the toilet and make the coffee."
And this proves that you are part of the problem. Why do you consider someone cleaning the toilet and making the coffee so value
Re: (Score:2, Insightful)
Actually, Everyone CAN be rich. The problem is that most people do not want this and actively work against it. The "super rich" want to make sure they stay that way so the objective is to have as much money as they can get to keep the power that being super rich gives them.
The reality is that there are too many people on this planet, and the sustainability of the natural resources like clean water, etc, is rapidly degrading.
The economical games will play themselves out, and there will be winners and losers.
But at the end of the "the day", we will all pay the price for the blatant greed and mismanagement of the planet.
Everyone is rich! (Score:2, Insightful)
Re:Everyone can be rich! (Score:5, Insightful)
Because, by the standards of 1850, we are all rich?
And by the standards of today, everyone in 2200 will be rich?
Yeah, there's not going to be a "**poof***, you're rich" moment for most people, but standards of living are, in general, going up for everyone. And, barring someone gaining the power to kill the goose that laid the golden eggs, standards of living will continue to go up for everyone....
Re: (Score:3)
And by the standards of today, everyone in 2200 will be rich?
I see you are a fan of extrapolating.
Re: (Score:2)
https://xkcd.com/605/ [xkcd.com]
Re: (Score:3)
No, this is not extrapolating but pointing to two facts:
a) We experienced actual improvements of standard of living over last century
b) rich and poor is measured against what other people have. If everyone is destitute, a person with modicum is considered rich. If everyone has plenty, then the person with the smallest pile of plenty is poor.
Re: (Score:2)
a) We experienced actual improvements of standard of living over last century
Extrapolating is when you assume that these improvements will continue until 2200.
Re: (Score:2)
Re: (Score:2)
Note the part that I quoted: "and by the standards of today, everyone in 2200 will be rich?"
Re:Everyone can be rich! (Score:4, Informative)
Since the 1970's standards of living in the US have stagnated and/or declined for the middle class and poor depending on location. The only increase has been with the wealthy.
Re:Everyone can be rich! (Score:5, Informative)
> but standards of living are, in general, going up for everyone
Except they're not really. In terms of how many hours of work it costs us to get various "toys" - cars, computers, etc. we're all getting wealthier rapidly, but when you look at the cost of the things we *need* - homes, nutritious food, medical care - most of the population is getting poore at least in the US.
Example:
1970: Median household income was $9,870 (https://www.census.gov/library/publications/1971/demo/p60-78.html)
Median cost of a new home: ~$24,000 =2.4 years income (https://www.census.gov/const/uspricemon.pdf)
2016: Median household income $58,000 (https://www.census.gov/search-results.html?q=median+income&page=1&stateGeo=none&searchtype=web&cssp=SERP&_charset_=UTF-8)
Median cost of a new home: ~$220,000 = 3.8 years income
Re: (Score:2)
Re: (Score:2)
"Interesting, but have you adjusted this for average square footage of houses built in 70s and today?"
Houses have gotten markedly bigger since the 70s, but apartments have gotten markedly smaller. Home ownership is falling, apartment living is rising.
Re: (Score:2)
> but standards of living are, in general, going up for everyone Except they're not really. In terms of how many hours of work it costs us to get various "toys" - cars, computers, etc. we're all getting wealthier rapidly, but when you look at the cost of the things we *need* - homes, nutritious food, medical care - most of the population is getting poore at least in the US.
Example: 1970: Median household income was $9,870 (https://www.census.gov/library/publications/1971/demo/p60-78.html) Median cost of a new home: ~$24,000 =2.4 years income (https://www.census.gov/const/uspricemon.pdf)
2016: Median household income $58,000 (https://www.census.gov/search-results.html?q=median+income&page=1&stateGeo=none&searchtype=web&cssp=SERP&_charset_=UTF-8) Median cost of a new home: ~$220,000 = 3.8 years income
The metric you use is incorrect. The issue is not home price/income, it should be home payment/income.
1970: Mortgage rate in 1970 averaged 9%. The payment on a $24,000 home would have been $193/mo or $2,316/yr.
That is 23.5% of annual income.
2019: Mortgage rate in 2019 is currently 4%. The payment on a $220,000 home would be $1050/mo or $12,600/yr.
That is 21.7% of annual income.
We are the ones that have it easier.
Re: (Score:2)
1. No it can't. It is dependent primarily upon length of mortgage and interest rate. This statistic is more useful than the total price.
2. Pedantry. The metric was unfit for the purpose of determining if people were wealthier in 1970 or are wealthier in 2019.
Re: (Score:2)
The CPI - rate of inflation - is calculated based on the cost of those essentials [investopedia.com]. It's basically the cost of stuff most families are likely to spend money on. Certainly you can cherry-pick outliers which have risen faster than median income (like housing and cars), but others have risen slower (pretty much everything else except medical care and energy
You're implying that the financial sector (Score:2)
bitcoin (Score:2)
Bitcoin is the perfect example of what you say. Two people can keep trading it back and forth until both are millionaires.
Trading in the stock market is entirely different because a company's price won't go up unless there's a reason to believe it's profits will increase. The mistake this guy made was believing that yesterday's charts and today's rumors are good predictors of where a stock will go.
Re: (Score:2)
Of course they don't have any money until they cash out. But here's an example:
I find a pretty rock and offer it to you for $1M. You buy it for $1M and give me an IOU. Now if you can find a sucker to buy your $1M bitcoin - oops sorry, your pretty rock - we both get $500K.
A small number of people are controlling the price of bitcoin and trying to lure suckers into buying it when they run the price up.
Re: (Score:3)
This robot really illustrates how ridiculous large (probably most) parts of the financial industry have become. I mean, if analysts and traders were actually producing useful economic value, then as this trader robot became better at its work and cheaper to run, we could all eventually own trader robots and sit back while our bank accounts are filled with money.
The problem with the modern financial industry is that it has become a game where the only winning move it to get others, with access to less resources (financing, information, etc) than you, to play. HFT trading faster than any person could possibly react, siphoning value off of trades. Massively overpriced IPOs (I'm looking at you tech unicorns with no viable path to actual, regular profit in the next decade!) being pushed to retail investors, making the banks, trading houses, and initial stock holders
Re: (Score:3)
HFT trading faster than any person could possibly react, siphoning value off of trades
HFT is done for arbitrage. Arbitrage is a value-adding service. The value that is siphoned off, is only a part of that added value, so it works as a net benefit for all.
Re:Everyone can be rich! (Score:4, Informative)
HFT in particular has reduced the spread on purchase/sale substantially. When I started investing in the 90's it was typical to pay an arbitage of 5% or more for any sale or purchase, but with HFT that percentage is now much less than 1% and it's approaching zero because of the HFT's. This has benefited every investor.
The dirty little secret is analysts and traders (Score:3)
This is why the economy crashes every 10 years like clockwork. We let the wolves watch he chickens and then when they run out of chickens we borrow money to buy more for those wolves.
A more interesting question (Score:3)
Re: (Score:2)
If you can come up with a reason as to why an AI would do that, sure.
Robots do not have the same measure of materialistic needs as humans, nor do they experience any concept of pleasure by satisfying needs or desires, so there is no reason for a robot to want to embezzle money unless it was explicitly programmed to do something which required it to do so.
Re:A more interesting question (Score:4, Funny)
Re: (Score:2)
Again, remember that AI's are not human, nor are they driven by ambition or greed.
Re: (Score:3)
There's probably a story to be written where an AI/robot does this to wipe out humanity. Think Ultron, but instead of making thousands of copies of itself, it sets up a company (or series of shell companies), amasses a fortune, and then uses said fortune to influence global politics. Then, it's just a matter of lobbying the right politicians in the right directions and broadcasting the right reports (via the media organizations that the AI/robot owns) so that humanity wages war on itself. In the end, humani
Corollary: ... (Score:2)
Re: (Score:2)
The owner/operator of the car, obviously.
Re: (Score:2)
Re: (Score:2)
I meant it as "owner OR operator", depending on circumstance.
AI is magic (Score:2)
AI is the new magic for people who don't understand it. People really believe it is intelligent, and "AI researchers" are more than willing to let them believe it. That is why they use terms like "Neural nets" and "deep learning" even though they are nothing like the how the brain works. It is like the Theranos of Computer Science right now.
Games (Score:3)
If the AI worked, one would keep it to oneself and manage your own money into the stratosphere. This is just more pushing around someone else's money.
Get rich quick schemes would, for the inventor, be much more profitable to the inventor to start a business and hire people to do the scheme. Selling how to do the scheme is a hint it ain't all that and a bag of computer chips.
Re: (Score:2)
Only true if the AI cannot profitably manage more money. Otherwise, 10% of someone else's money is still free money. I mean, I'd rather invest my savings plus yours, and keep all the change in my savings and 10% of what your savings make than just manage my savings.
Re: (Score:2)
Sometimes you need a critical starting cash level to make it work efficiently. (The reverse can also be true—too much money could move the market in unanticipated ways.)
The problem is obvious (Score:3)
Re: (Score:2)
The AI needed more blockchain and quantum computing.
If only they went with Web 2.0 solution, none of this would have happened.
Rich people problems (Score:2)
Would he have shared profits? (Score:2)
Sometimes dumber is better (Score:2, Interesting)
A friend got me quite interested this field a few years back. I don't have the reference now but, there is an amusing paper about a contest for trading bots.
The TLDR; is that the best performing bot in their simulations was.... 'the dumbest'. It used no information other than price (including fees/taxes ofc), and simply calculated and watched the first and second order derivatives of the price, and then executed a simple iterative purchase/hold/dump strategy based on the current price movement. (it would cy
Citibank is the real victim here. (Score:2)
When I RFTA, nowhere did Li (the investor) do any kind of serious due diligence or even research. This idea caught his fancy and he was willing to drop a quarter of a billion of his own cash and added over two billion of Citibank's.
If anybody should be suing anybody it's Citibank suing Li for bringing them into a deal that wasn't appropriately vetted or researched. I'd be interested in seeing the contract there and what Li promised Citibank.
Re: (Score:2)
Hard to say that being leverage 10x into anything is not gambling pure and simple. Robot or not, only having 10% downside risk tolerance before you are bankrupted is hugely risky. With that said, Citibank should I am sure was collecting a decent interest rate on the money they lent, that is kind of what these banks do. Should they be allowed to be involved in such leveraged investing is a different question.
Only a story because "AI" is involved (Score:5, Insightful)
When I RFTA, if you take out the words "AI" and replace them with "New Approach" I don't see anything unusual about this swindle.
Rich guy who wants to get richer gets introduced to somebody his crowd calls "Captain Magic" and says he has a new approach to playing hedge funds. "Captain Magic" presents the rich guy with some numbers suggesting the new approach is better than the way people currently invest in hedge funds. Rich guys is infatuated with what he sees and drops a quarter billion of his own cash and brings in Citibank on his word. New approach is applied using rich guy's and Citibank's money and turns out to be a bust.
The story isn't about AI, it's about a rich guy who didn't do his homework and lost a tractor trailer load of cash on what should have been obviously a bogus investment.
PEBKAC (Score:2)
Robots have a pretty clearly defined role in today's stock market, and that is managing funds that already have a clear-cut investment strategy (ie; index funds that own equal amounts of everything in whatever index they track).
This is something robots can do very well - when the fund's only job is to divide all of its money equally between stocks, a robot can cut an ETF's expense ratio dramatically. Most ETFs have an expense ratio around 0.20-0.30.. but I've seen robot-managed funds with expense ratios as
No one (Score:2)
You can't sue a "Quantitative Analyst" either (Score:2)
If you trust your money to some hedge fund, and they just make the wrong decisions and lose it, you can't sue them either.
Why should that change, when the decisions are made by an AI instead of a human?
They probably should be transparent about what they're doing.
Also maybe we should be less concerned about such an AI failing than about such an AI being successful beyond all expectations, because that could pose a much larger risk to the world wide financial market than just some hedge fund going belly up.
Re: (Score:2)
He's claiming fraud. If you trust your money to a hedger fund, and they claim all the analysts are MIT trained PhDs in math, and it turns out all the analysts are high school dropouts, you can sue. He's claiming the AI was misrepresented.
Do I think he's right, no. But it's a legal reason to sue.
The real question here is a different one (Score:2)
You bet money, you lose money. That's not the question.
The question is, how did he get Citigroup to dump 2.25 billion bucks on him. Did he just tell them what he's going to do with the money?
And will we get to pay for Citigroup's gambling with yet another bailout?
Shouldn't sue anyone (Score:2)
LOL this is stupid (Score:2)
Nothing new
https://en.wikipedia.org/wiki/... [wikipedia.org]
Bunch of clueless people hear trigger words and they lose money. Back then it was "quantitative modelling", now it's "AI'.
It's just a bunch of algorithms and IF-THEN-ELSEs, guys.
This again (Score:2)
I dunno how many times I've read about someone with an algorithm or AI system that can beat the market. They tweak it until it can retroactively predict past markets, and then proceed to lose a fortune. Sometimes their own fortune, sometimes somebody else's.
Markets are fundamentally chaotic systems. There are too many factors, many of which are utterly unpredictable. Buy and hold for the long term - that seems to be the only strategy that works consistently.
Re: (Score:2)
Markets are fundamentally chaotic systems. There are too many factors, many of which are utterly unpredictable. Buy and hold for the long term - that seems to be the only strategy that works consistently.
Also, markets are almost zero sum, and there is no way that you can consistently beat other agents (which may be clones of the same AI) that have access to same information.
Re: (Score:3)
Basically, even with infinite penises, there are only so many assholes to be fucked.
AI's go negative? (Score:2)
It always seems that AI tend to favor dark holes. YouTube showing people disturbing stuff, FB showing sad stuff, MS Chatbot that turned racist.
We have here an investor AI that went negative. Down a dark hole, favoring the down-swing. Making poor investment recommendations.
What McAfee did for virus scanning, someday AI will need a kind of computer-Prozac.
Until then, I can't feel sorry for a poor investment where the investor didn't pull the plug. It's probably a gambling like addiction - throwing bad mo
Same as when you sue people with stupid employees (Score:2)
You can only really* sue someone if it does something negligably stupid, the sort of thing you sue people for.
If you wrote it/built it, you're shit-outta-luck. You can't sue yourself.
If you had someone else build it for you, you can sue them.
If you paid for it, like a plane ticket or a Tesla or a stock broker, then you sue the people you paid for using faulty equipment. Just like if their plane falls out of the sky for any other reason. They in turn might be able to sue whomever they had build the thing. Ag
Aand (Score:2)
Artificial "Intelligence"? (Score:2)
"Artificial Intelligence" will never be able to compete with human stupidity. And you should immediately discount anything said by anybody who thinks that "artificial intelligence" will have any significant influence in the real world.
Look, there's nothing artificial about intelligence. A computer program was written by HUMANS, and just acts in accordance with its programming. And EVERY program has bugs. No program will ever be able to make "decisions" other than the ones already designed in by the HUMAN pr
Something borrowed, someone's blue. (Score:2)
Li eventually let K1 manage $2.5bn -- $250m of his own cash and the rest leverage from Citigroup.
So he borrowed the majority of the money from Citigroup and bet machine would beat the market? Sounds like he was just gambling with other people's money and lost; now he's looking for a scapegoat instead of being on the hook for $2.25bn to Citi. Sucks to be him.
Correction (Score:2)
The title should read: "Who to sue when *greed* loses your fortune"
If you believe you can have double digits returns without significant effort, then I have some bridged to sell you, along with unicorns to ride on them.
Why would someone trust Raffaele Costa? (Score:2)
Re: (Score:2)
He just believed that an AI was going to magically make him money?
I thought that's what AIs are supposed to do, right?
Re: (Score:3)
Because playing with fake money doesn't account for the effect of real money on the market once it starts playing with real money. There is a real disconnect when playing with fake money, that disconnect doesn't allow the AI to learn its own impact on the market with its trades.
What ends up being created is an AI that is not self aware enough to realize that the trades it is making is affecting the market and you end up in a feedback loop that is based on an AI that hasn't spent enough time learning in the
Re: (Score:2)
We are already there. High frequency trading is impossible with a human in the loop. Milliseconds of network latency can affect trading outcomes. Automatically scanning news feeds or tweets for company news is already a thing.
The subject asks 'Who to sue'. That's an easier question to answer. The broker licensed to execute the trade. Even if that broker is sitting idly behind an HFT system s/he is still responsible for the trade. It's legally the same as a broker calculating an equity position for a custo
Re:Sue yourself (Score:4, Interesting)
Welcome to the world of the ultra-wealthy. They have so fucking much money you literally don't believe it. He did try it out small, investing less than 5% of the money he inherited from daddy. And he did promise to double his investment if it showed positive results after a year. And he did have his investment pulled out if it lost more than 0.5% of his total wealth (10% of what he invested.) And that's exactly what happened.
You have to remember that hedge funds that are primarily buying stocks, unlike those trading in derivatives or startups, are unlikely to go to zero. So an investment of this scale is unlikely to all get lost before you pull out your cash.
Re: (Score:2)
FTFY. Reason: fees.