The death and life of prediction markets at Google
asteriskmag.com
asteriskmag.com
Famously, Microsoft and others pioneered dogfooding decades before the events described in this article and approximately a decade (at least) before Google came into existence.
And I’m 99% certain company cafes existed at least a half century before Google invented the concept.
But dogfooding, yeah that had been around for a while. Originally from Alpo, iirc. The first tech company to adopt the term as well as the practice was Microsoft in 1988.
A/B existed before Google but 2000s era A/B testing and user research were unparalleled until Facebook also started putting serious capital into it. Nowadays it's considered table stakes but it was revolutionized in the beginning of the millennium. Maybe not entirely by Google but substantially so, and driven heavily by their product launch review process.
Google pioneered information retrieval and ranking innovations, not behavior optimization.
It would have better to say Google popularized these practices rather than pioneered them. Or at least, that these practices became much more widespread among tech companies after Google's IPO in 2004 than beforehand.
I think it's also safe to say that Google's culture was strikingly different from other tech companies of its era, as has been well documented in a few books.
This also seems incorrect. Before Google, it was common to have company-provided before Google. IBM and Motorola had cafeterias. I don't know when AMD installed their cafeterias, but if it was post-Google, it would've been inspired by IBM and Moto's cafeteria and not Google's. In Austin, the Moto cafeteria was known for having very good food and IBM was moderately subsidized and pretty good until the 2010s, which doesn't line up with Google being influential at all. And Centaur had great, free, food. This is an old idea that predates tech companies that a lot of tech companies have picked up that Google also happened to pick up.
As a term, dogfooding spread through Microsoft after Paul Maritz wrote an emailed titled "Eating our own Dogfood" in 1988. If the term was popularized by anyone, it was probably Joel Spolsky who took the practice from Microsoft and blogged about it when he was the most widely read programming blogger. But there are a lot of examples of people doing this before Martiz's email (they just called it something else) and before tech companies even existed; this is another practice that predates tech companies that tech companies picked up.
I don't know about the history of A/B testing in tech, but Capital One was doing A/B tests at scale before they would've been influenced by Google and that's another idea that was used outside of tech.
As far as I know, Google was one of the first to offer food that was tasty enough, healthy enough, and cheap enough (free!) that nearly everyone ate at the company cafes on a daily basis.
I've heard from people who stayed at IBM that the food declined to cafeteria food quality over the next ten years, which led to the cafeteria basically being abandoned because people ate out so much. But that's actually counter to the narrative in the post — IBM had decent food before Google, and then some time after Google's IPO, the food declined to became standard cafeteria food.
For a while, another building was notorious for serving sushi but only admitted their Android developers, because Andy Rubin was paying for that himself.
I've never had free food routinely except customer briefing center or some other lunchtime work function. Rarely went out unless it were a short walk. (The brief time I worked in downtown Boston with no cafeteria is pretty much the only time I went out for lunch routinely.)
Per another comment, my sense is that brown bag lunches used to be more common and most people stopped doing that.
Some hardcore eaters in Austin.
I'm not sure that was popular, though (as in something the majority of the population believed in). Grandma in Poducksville almost certainly had no idea. She would have known about Google doing the same, though, as it was blasted all over the news constantly for a while.
Edit: I don’t want to be harsh on you, but the fundamental problem of credibility, especially in online writing, is that it takes one mistake to lose an amount that takes hundreds of correct decision in a row to regain…
Google's consumer-facing systems all tend to be very focused. Things like search, maps, gmail etc. are not the same kind of system as Amazon's store.
While these systems do presumably give Google something to exercise their cloud systems on, the sense I have (as a longtime user of both GCP and AWS) is that it doesn't give them a realistic sense of what other companies, that don't just sell advertising and consumer data via focused products, do. Amazon's store is more representative of typical businesses in that sense.
Basically, it seems to me that Google Cloud has continually learned lessons the hard way about what customers need, rather than getting that information from its own internal usage.
Lots of other Google products suffer from similar issues because of an apparent lack of dogfooding. I bought a Pixel phone not so long ago and I had to install all updates, one by one, to bring it to the latest Android version. It took several days.
Let's use Google's Colossus (their datacenter-scale virtual filesystem) as an example. Due to the underlying architecture of Colossus, GCP can turn around and give you:
• GCE shared read-only zonal PDs
• near-instantaneous snapshots for GCE and BigTable
• async and guaranteed-durable logging (for GCE and otherwise) and Queues (as Pub/Sub and otherwise)
• zero-migration autoclassed GCS Objects, and no per-operation slowdown on GCS Buckets as bucket size increases
• BigQuery being entirely serverless (vs e.g. Redshift needing to operate on a provisioned-storage model)
But Google can't just sell you "Colossus as a service" — because Colossus doesn't have a "multitenant with usage-cost-based backpressure to disincentivize misuse" architecture; and you can't add that without destroying the per-operation computational-complexity guarantees that make Colossus what it is. Colossus only works in a basically-trusted environment. (A non-trust-requiring version of Colossus would look like Apple's FoundationDB.)
(And yeah, you could in theory have a "little Colossus" unique to your deployment... but that'd be rather useless, since the datacenter scale of Colossus is rather what makes many of its QoS guarantees possible. Though I suppose it could make sense if you could fund entire GCP datacenters for your own use, ala AWS GovCloud.)
For what I understand as an outsider, Google is much more monolithic, having a platform where each team can do their things independently is not really their culture, so if they build one, it is only for their customers, because they don't work like this internally. Whereas for Amazon, an AWS customer is not that different from one of their own teams.
I now work for a large customer and you would be shocked at the household names that basically put all their infrastructure in a single Account and Region. Or they have multi region but it’s basically an afterthought and wouldn’t serve any purpose in a disaster.
And that doesn’t even mention the comic “Moving to AWS” platform that technically consumed AWS resources, but was a wholly different developer experience to native.
It had an on-site cafeteria, which aside from being "out in the middle of no where", we all wondered "why?". Why didn't everyone just bring their lunch?
Even at the time, we just explained it as a "just a thing that Americans did" and wrote it off as because of the presence and involvement of NASA.
So it's good to hear that Google invented it sometime later...
/Never listen to tech people commenting on history or economics, lol
The CASE products were intended to automate away coding, and Oracle's HR products, with their gazillions of lines of code, and with the developers located in the same building and with a common senior management team, were the best possible opportunity for dogfooding.
However while I was in the HR product team we never tried in earnest to use CASE (except for the ER modelling tools). There was no real enthusiasm from the CASE team to support us and the tools as they were then fell far short of what was required.
Later, when I joined the CASE team, I learned that their narrative was that the company's ERP products (like HR) were so complex that they were not realistic targets for CASE (* cough cough bullshit * - and perhaps the sort of attitude that doomed the CASE products in the long term).
My learning was that dogfooding is an awesome strategy, but sometimes much harder to embed in a development team than one might think.
that's a very interesting tidbit.
It looks to me that the management structure and incentive at this department is too conservative, because failure is seen as bad and is probably punished (somehow - might not be overt).
Therefore, leadership is incentivized to target realistic use cases, which means simple use cases. This basically "guarantees" success as described by the objective.
This is the same as revenue forecasts being overly conservative, and the market sees through the lies.
As a vendor, I (well, my company) had to pay for meals at SGI, I have no idea if the employees got free meals.
I believe Charlie's (the main onsite cafe) has been renovated a few times although the basic layout was constant throughout my tenure (2007-2019) and in fact if you looked behind the curtains (literally), there was basically the equivalent of an archeological trash heap with generations of Google and SGI documents.
Over time Charlie's got worse and worse; the food quality dropped significantly and became quite monotonous (true for the other cafes as well), and Noname (eventually named Yoshka's) did too. In fact, every great cafe I remember attending was eventually replaced with a worse version of itself.
Crucially, "predictions markets" do not and cannot exist in any real sense. A pure predictions market would be completely isolated, causality-wise, from the event they are trying to predict. But the two are not and cannot be isolated, except for some degenerate cases like trying to guess the output of a true random number generator (and even then I'm not so sure sufficiently motivated people wouldn't try to game the system anyway). This is why we have problems with our current predictions markets, e.g. the stock market (insider trading, etc.) and sports betting (match-fixing, etc.).
Every prediction with a stake is an incentive to alter the outcome of an event. Once the weight of the stake outweighs the resources being used to ensure the impartiality of the outcome, the wheels fully come off the cart and the prediction stops being about the underlying event and starts self-referentially predicting the impact of the prediction itself. The snake eats its own tail and the market becomes useless. You cannot scale up a predictions market without this eventually coming to pass. See also the famous example of how a predictions market for when public figures will die is just an assassination market with extra steps.
E.g. if before the election you think that a certain candidate winning would cause the markets to react in a certain direction, you could "bet" on the other candidate so that if your portfolio value goes down, you earn the proceeds from the bet to recoup some of your portfolio losses. Or if the "good for stock market" candidate wins, you loose the money you bet but the gains in your portfolio balances it out.
In that case, you're not really betting on who you think will win. You're just betting as a hedge just in case that person wins.
But this itself is a form of market distortion. It calls to question what, precisely, people think the market is supposed to be measuring, both in theory and in practice.
If it's an extremely dry year, you profit from the weather futures instead of your crops (and vice versa). Buying weather futures isn't necessarily a prediction of what you think the weather will be.
If you have a higher risk tolerance, you will buy fewer futures. If you believe the next year will be dryer than normal, you will buy more futures than normal. If you believe your crop is likely to be better/more reliable than normal, you will buy fewer futures.
The point is that you, the farmer, don't need to take a view on whether the next year will be drier than normal. You just buy $X worth of rainfall futures.
The same way you shouldn't buy more flood insurance if you think the next year will be exceptionally wet. You can't really predict that, after all. You should buy flood insurance roughly up to the value of restoring your house after a flood, and you should hope the insurance market is healthy enough that the cheapest provider of that insurance offers you a price that reflects the expected value of the insurance plus a small markup.
And I'll reiterate, this is a function of your risk-aversion/efficiency. One would expect, for example, climate change to increase the price of weather futures as extreme/problematic weather events become more likely. It's often difficult to see the impact of these changes on the scale of a single farmer, but in aggregate lots of farmers do a market make.
> You should buy flood insurance roughly up to the value of restoring your house after a flood, and you should hope the insurance market is healthy enough that the cheapest provider of that insurance offers you a price that reflects the expected value of the insurance plus a small markup.
And the insurance companies have a small army of actuaries who make sure that the prices they provide take into account conditions like the relevant risk factors of where your home is. This is instead of a betting market style concept, where you could instead imagine every individual actuary as a potential insurer.
The cost of that varies though. If you have to pay $95 to get a $100 payout that’s a very different calculus from $50 for $100.
Sure, but if I, a non-farmer market player that couldn't give two fucks what the market is even about, can predict that the next year will be dryer than normal, and to what degree, better than anyone, I can make money buying up however many of these futures I can afford. It works even better if I can actually make the weather more dry somehow.
This, I believe, is called "providing liquidity to the market", but curiously, if I tried that with flood insurance, I'd just be guilty of insurance fraud.
Apparently it's onions and box office returns? What weird corner cases. Why not strawberries too?
Can't the onions futures market be regulated the same way as all the others?
If anything, this makes me think all the rules are arbitrary.
Nothing (at least for other perishable foodstuff); law often doesn't even in theory have a broad universal theory behind it, but instead responds narrowly to observed or perceived immediate problems.
You're saying the answer to the above question is "because there was an immediate problem with onion futures in the 50s". I don't think that's what they meant. That would be unrelated to "the fact that futures markets are so heavily regulated".
I guess if everyone has a different opinion, and every reply comes from a different person, there's no "discussion" as I understand it.
We saw Trading Places.
Calling some of those effects "distortions" is a tricky business at best.
Large trading firms exist on finding and exploiting small arbitrages between various correlated assets. If you assume a perfect market with infinitely many participants and infinite liquidity, then this “works” - there is no distortion at scale.
I'm starting to think that the answer is what mhh__ wrote: who cares? Markets aren't there to measure anything. Markets are there to make money for participants. Any measurement that can be attributed to the markets under some conditions is, at best, an incidental side effect.
The main arbitrage opportunity was in finding ways to place illega bets in the bettor's jurisdiction.
Some people did exactly this back in 2016, and just ended up feeling bad, because they were profiting off a "bad" (in their eyes) event.
Can't let perfect be the enemy of the good!
If you want a fuller critique of prediction markets in the corporate setting, see the Dec 2021 article linked near the end [1].
[1] https://forum.effectivealtruism.org/posts/dQhjwHA7LhfE8YpYF/...
OK but strong participation isn't necessarily a positive for the accuracy of a market. The problem is the prediction market with lots of participants can be just outlet for partisans to put forward their opinions. There's a tax on wrong opinions but someone is spending a bucks, the markets won't be a powerful force for changing those opinions.
What you want for a prediction market is for the major participants to actively researching the problems - expend money and effort to have a well founded reason for their positions. Markets for random real-world events have the problem that many events don't occur often to weed out arbitrary biases and there may not be any easy or cost effect way to attain a well-founded opinion on the subject.
> A senior executive saw Prophit give a very low probability that the company would complete the hire of a new senior executive on time (filling the position had been a quarterly objective for the past six quarters). “The betting on this goal was extremely harsh. I am shocked and outraged by the lack of brown-nosing at this company,” the executive said to laughter in a company-wide meeting. But the market was the nudge the execs needed. They subsequently “made some hard decisions” to complete the hire on time.
Indeed. The whole point of the prediction markets espoused is to alter the decisions being made. That means the prediction itself can have an impact on the outcome intentionally or otherwise.
> This turns out to be a general lesson from running a corporate prediction market. Forecasting internal progress, and acting on that information, requires solving complex operational problems and understanding the moral mazes that managers face. Forecasting competitors’ progress has almost none of these problems.
Forecasts on competitors (or, say, regulators) avoids this problem... unless employees are manipulating the outside world too!
Basically, betting markets have all the problems and risks of traditional public markets (insider trading) without any of the regulation or ability to enforce the law.
Which is great when the impact of the prediction market is "people making hard decisions" and not so great when they're studiously slowing down the process because they've got a bet on something not happening on time...
I think one criticism that the linked post misses but the OP article touches on is that most forecasts (whether they be prediction markets or super-forecasting style) is that they often predict the wrong things.
> We asked questions of the type “Will Google integrate LLMs into Gmail by Spring 2023?” and “How many parameters will the next LaMDA model have?” Yet what executives would have wanted to know was “Will Microsoft integrate LLMs into Outlook by Spring 2023?” and “How many parameters will the next GPT model have?”
It's really hard to build a prediction market for these things and I'm not sure "forecasting" is the right way of thinking about them.
(Following some links led to https://www.lesswrong.com/posts/uGkRcHqatmPkvpGLq/contra-pap... which has some interesting points too)
You’re describing endogeneity. It’s unlikely prediction markets affect natural disaster odds.
They don't need to. Predictions markets for natural disasters exist, we call them insurance companies, and the concept of moral hazard when it comes to insurance is a well-studied topic: https://en.wikipedia.org/wiki/Moral_hazard
Would be very interested to see further discussion about this.
They key thing to remember is that these markets can have utility beyond zero-sum betting aspect. For example, the stock market isnt just gamblers betting against each other, but is also a tool for auctioning corporate ownership and raising corporate funds.
A unit of stock represents a legal claim on a company's assets even in the absence of a market for that stock.
In many situations, we do want people to influence the value of their stock. Any company that grants stock is doing so because they expect employees to work harder. There are many cases in which employees with stock grants might make short-sighted decisions for quick profits, but on the whole stock grants make employees into better workers.
> A senior executive saw Prophit give a very low probability that the company would complete the hire of a new senior executive on time (filling the position had been a quarterly objective for the past six quarters). “The betting on this goal was extremely harsh. I am shocked and outraged by the lack of brown-nosing at this company,” the executive said to laughter in a company-wide meeting. But the market was the nudge the execs needed. They subsequently “made some hard decisions” to complete the hire on time.
In this case, the predictions market would've been useless had the executive team not used it in their decision-making.
A pure predictions market would be completely isolated from the event they're trying to predict. But I would say that's not an ideal predictions market.
I have plenty of internal metrics to demonstrate my contribution, but nothing I do measurably affects the stock price.
Could $10,000 equivalent of the yearly prediction market awards have actually moved a Google executive-level job search if they were spent directly on that process? That's less than a month's salary for a sourcer, recruiter, and interviewer.
The execs will then end up with lightly-traded, inaccurate markets on events they control, but probably still reasonably accurate predictions on events that they can't. That is maybe still useful, but it means you will have to think hard about the nature of each market before offering contracts on it, which may not really be easier than just doing the forecasting some other way.
A prediction market that is a dance between observer and market and agent doesn't "just work". You can't rely on its predictions because they are conditional on your choices, it's a complex feedback loop.
> Every prediction with a stake is an incentive to alter the outcome of an event
...is a feature not a bug.
For instance, imagine if PR's on core infrastructure (like xz-utils, for instance) were first reviewed by a set of trusted maintainers, and--supposing they pass that gate--were later put into a sort of limbo where people can bet on whether they'll be merged. The maintainers then make a policy around betting that the PR will be merged. They're in charge of whether it actually gets merged or not, so most of the time this is a pretty good bet and they just recycle that money by winning and then re-betting it on the next PR.
Of course third parties can also bet in the same way--these would be stakeholders which are not maintainers, but who wish to "sweeten the pot" and encourage people to spend enough time with the pending code that they might find a reason to "bet against the house". No prior coordination between the maintainers and the stakeholders is necessary.
Suppose I notice some malicious code in one of these commits which is in the betting phase. (Maybe I do my own testing on pre-release versions as a stakeholder, or maybe I'm a bug bounty hunter.) I can bet that the PR won't be merged. The monetary value of that bet makes it clear that even though I'm a stranger, I'm not a spammer. That "buys" me the attention of the maintainers, and if I'm right about the code being malicious, the maintainers will decide not to merge the commit: I'll have successfully altered the course of events (bad commit doesn't get merged) such that I get paid for having been right.
Why not just write it so that non-natural causes of death don't pay out? More generally, make it so you can't wager on illegal events / outcomes, or ones where a crime materially affected the outcome. Anyway, if someone bets big on a public figure being assassinated, and then that public figure gets assassinated, it would seem like a good place to start investigating would be to look at the people who made a lot of money from that bet.
There were 2-3 assassination attempts on Trump this year.
It would be valuable to incentive people to share knowledge of assassination vulnerability, to guide security efforts. Banning that defeats the purpose.
The question is whether this is improper influence? Is it too influential, for bad reasons?
I think a more reasonable critique of prediction markets is that it's guesswork laundering. We are given a number, but we don't know why that number was chosen. How can we tell whether it's justified?
When markets move, there is a whole industry of people coming up with explanations of why it might have happened - more guesswork!
A lucky guess can be helpful if it can be verified, but knowledge should be about more than making guesses. Sharing evidence is important.
You could let ordinary people piggy back participate too, and then use the results to filter through internet/media noise but that starts to smell too social score-ish.
Some examples:
- A question about asset prices specifies FTX as resolution source, but then FTX stopped existing.
- There was a question about wether submarine cables in the Red Sea would be destroyed by a hostile act before a certain date. The cables got damaged, but it seemed to be a (suspicious) accident, with very limited independent media coverage.
- There is a question about wether YouTube would be blocked in a certain country before 2025. It got throttled, to the point where it is unusable in practice, but not technically blocked.
- There is a question trying to forecast LLM progress. How to quantify that? It chose "What is the state of the art score on the Penn Treebank at a certain date?", which was a standard benchmark at the time. But as LLMs evolved, new benchmarks got developed, and although current LLMs probably score much better on the Penn Treebank than a few years ago, nobody reports Penn Treebank score any more. The question ended up being about the popularity of the benchmark rather than LLM progress.Is there a point spread on journalistic accuracy? How do I take the under?
To make explicit what I'm assuming is your point, this statement from the article can't be correct because A. POWER is by IBM, not Intel and B. Apple had already launched such a computer back in 1994, before Google existed.
Maybe x86 was meant instead?
however interpreting it as a probability, or an average of agent beliefs, or anything like that, seems tricky. i assume these internal markets are not deep and liquid enough that you can just throw up your hands and say "EMH". it works if you assume risk-neutral traders who will just trade up to their correct price but as I understand it, breaks down with realistic traders who may limited capital and are usually somewhat risk-averse.
i wonder how these prediction markets dealt with that. was there any postprocessing of market prices to get final probabilities? based on interviews with traders or observed trading behavior, did the traders behave in such a way that the market price could be interpreted as "pretty much" just a probability?
Google published [1] one such calibration chart on its current prediction market in late 2021. Also, the 2009 paper in the article [2] on Google's first prediction market published one too.
[1] https://cloud.google.com/blog/topics/solutions-how-tos/desig... [2] https://static.googleusercontent.com/media/services.google.c...
You cannot really do postprocessing to the market price to get the average belief back out, because the bounds aren't very tight: a market price of 50% could correspond to an average belief anywhere between 29% and 71%.
I've said this many times: innovation at google is dead. This is yet another supporting data point. Firms that innovate aren't concerned with what competitors are doing. At google it appears it's their main concern.
As currently formulated, prediction market outputs are just a fancy opinion poll, where participants have some incentive for accuracy. To rise above the simple wisdom of the crowds, you would want to identify the subset of market participants that are constantly beating the market (because they have a more accurate mental model of the world). I think this necessitates both 1) long term tracking of bets and 2) likely withholding individual positions from the market to prevent follower effects.
Similar to the title article, this raises the question of who the ultimate customer is prediction market is. Individuals can be incentivized to bet by winnings, but who else is the customer for aggregated data?
I wonder about the extent to which current prediction markets have internal outputs and derivative statistics, and what they might do with it.
If polymarket or similar companies put Trump vs Harris at 55-45, do they have internal statistics that that put the race at 80-20% among their most accurate betters? Was this data for sale?
To that end, I'm not sure that these factors are mutually exclusive and would like to hear more of your thoughts.
If I understand you correctly, I would think that a market could have both volume and depth.
With respect to volatility-stability, what do you see as the drivers there? Is it that gamblers would need a significant upside to drive betting? Is this solved by the size of the bet? I suppose there is an internal conflict. If the line of a bet is 49-51% on a binary outcome, the risk of ruin is high, and the upside is low. You would need to aggregate outcomes over many distinct events to mitigate.
I suppose this could hang on ability/inclination of professional forecasters to research and take several positions.
Given two prediction markets, one which varies wildly and one which is stable, the former will attract more users. Particularly the most profitable ones. Even if the underlying odds are unchanging, the volatile market is more “fit.”
More people get to feel like winners for longer [1]. And reward uncertainty makes gambling more additive [2].
For purposes of information discovery, volatility is bad. But for purposes of gambling, volatility is good. Running an information-discovery (or financial) platform is less profitable than running a gambling platform. Herego, operators will optimise their prediction markets for gamblers.
It's difficult to see the niche for the academic market.
More profit to the gambling platforms means more money for R&D, customer service, user retention and marketing. That means more liquidity. Gamblers means dumb money, which in turn attracts the smart money: if you're commissioning private polling to place more informed bets [1], you want to place a big bet against dumb money.
[1] https://www.bloomberg.com/opinion/articles/2024-11-07/predic...
> To rise above the simple wisdom of the crowds, you would want to identify the subset of market participants that are constantly beating the market (because they have a more accurate mental model of the world).
Identifying, and then working with, the top traders at Google (including one card-carrying superforecaster) was a great joy.
And yes, they're sitting on some great data, on what the employee crowd tends to get right and wrong, who individually is good at forecasting what. Though one complication is that being a great trader is not the same as being a great forecaster.
To the author: if you have a personal blog, can you post that here?
My personal blog is defunct (for now!). But some of my recent writings can be found on the research page [1] of my startup, FutureSearch. We're building an AI that can forecast accurately.
We've written some pieces on topics like the problems with using crowds to forecast, and contesting recent papers' claims of good forecasts coming from simple LLMs.
There are some interesting jupyter to blog tools like quarto.org. Or, my Svelte based blogging tool: svekyll.com (I use it to blog about AI/ML because Svelte is the best visualization front end tool).
It's a great time for you to start blogging again!
If polymarket had an internal 80-20 model based on super-forecasters, they wouldn't disclose that, but they might have paying customers for it outside the prediction market itself. For example, stock traders might pay for the private model.
If I could pay the NYSE for real time trading info on the buy/sell limits from warren buffet and other whales, I would.
Uninformed bets should wash out as noise, and informed bettors should reverse uninformed moves so long as they are profitable.
The stock market analogy would be the predictions you could make as an individual if you knew the internal limits and assessments of the best trading firms, and not just current market prices.
If I could pay the NYSE for real time trading info on the buy/sell limits from warren buffet and other whales, I would.
I think the analogous situation to what I was proposing would be if a platform had open betting and organic odds, but sold the sharp betting data to 3rd parties.
Is it gambling to do an extra project in hope of getting a "Spot Bonus" (~$250-$1000) in recognition? That was a very standardized process at Google.
Is it gambling to file a patent application, which if approved, would lead to a $1000 bonus and a trophy you'd often see on the desks of Google Brain researchers?
Is it gambling to decline auto-sale of RSUs, and have your compensation determined more by movements in GOOG than in your cash salary?
2. No, Google isn’t taking actual money bets from their employees.
3. What employees do with their vested RSUs isn’t at all the same as hosting an internal gambling platform. Once they are vested, employees own the stock, they can do whatever they want. One could sell them all and use the money to bet on roulette, which I think is obviously gambling, but that’s also obviously not Google hosting a gambling platform internally?
I’m just surprised that Google hosting a platform where employees gambled was… allowed? This isn’t a moral judgement, I’m surprised that Google was operating a gambling platform internally and legal etc. thought that was a good idea.
I assume they were given free credit for the system, and had the chance to turn it into cash bonus if they won.
It is closer to a company giving out a prize for the winner of a free fantasy sports league.
The Google case turns out to not be real money, but it’s weird that the article never said that, and it’s weird that the author responded with three points of whataboutism instead of just saying “it wasn’t real money”.
If someone wrote an article that Google set up a roulette wheel in the microkitchen and employees “bet” on it, yes I’d assume they were running an actual casino and find that weird too!
Then it occurred to me that there are a lot of major lawsuits that depend on litigation finance, someone to front the cost of the lawsuit and return for a share of the winnings. So you could have a situation where gambling addicts are forced to resort to “their old ways” to get it off the ground!
(And leading to a paradox where, if they can bet on this case in a controlled way … maybe they really weren’t addicted the whole time? Like in the paradox about the lawyer who sues over getting a bad legal education.)
> as of August 2024, the team continues to refine its approach to make Gleangen a useful source of information for Google senior management.
And it is apparently now officially a way to get information to Google senior management.
> As Prophit had done, I got approval to pay out valuable prizes to complement the play-money leaderboards.
Some traders won things like iPads, and similar rewards that even highly paid tech employees considered valuable.