The biggest feature I miss from Google is the constellation of citc and the distributed build system that revolves around it. Most is described or implied by "Why Google stores billions of lines of code in a single repository" (2016) [1].
The key idea is that any human-written change only ever touches ~1-100 files, and there's never any need to store or maintain the entire source tree in your local persistence store. By only ever working on deltas, you can have highly efficient distributed and cached builds. This architecture imposes significant constraints: you must express your build graph in a modular way [2], you collocate your distributed FUSE backend and build system for speed, and all development is over the network. But it comes with many benefits at scale: near-perfect caching, instant "branch switching", fast builds even with deep and wide dep graphs, and excellent affected target determination so flakes/outages in one part of the codebase don't affect the rest of it.
I've worked in JS for around 12 years now and welcome this change of keeping JS as a scripting language but incrementally retiring it for performance-intensive general-purpose applications.
To my eye, JavaScript has three applications:
* A GUI-building language: A language and ecosystem we use to build UIs, where code runs on the end user's machine, served over the Web or as desktop applications using a compatible API (e.g. Electron).
* A scripting language: An approachable high-level C-like language.
* A general purpose programming language: A server-side or developer machine-side programming language for computational tasks like linking dependencies, transpiling code, type checking, orchestrating tasks.
At the first it's unbeatable, and the ergonomics and ecosystem are phenomenal. This is why it "won" as a language to write even desktop apps like VSCode and Discord.
At the second, performance doesn't matter anyway, so making JS capable of this just makes it approachable for folks who learned the language for the GUI-building purpose.
It's at the third application that the fundamental limitations of JavaScript---the programming language---really show. These are things like the lack of true multithreading (technically possible with significant limitations around shared memory and messaging and high overhead with Web Workers), lack of low-level primitives to fine-tune performance, and the ease with which authors can accidentally write extremely slow code. So much of this (TypeScript compilation, NPM package resolution/linking, Webpack bundling) in the modern Web ecosystem is regrettably still done in mostly JS.
I hope the JS community rallies around one general-purpose language to offload these tasks to as well as a uniform shim layer like Ruby's C extensions so we don't end up with a chaotic mishmash of different technologies. If nothing else, it would be nice to have a good excuse to write not-JS every once in a while!
> At the first it's unbeatable, and the ergonomics and ecosystem are phenomenal. This is why it "won" as a language to write even desktop apps like VSCode and Discord.
I don't really think this is attributed to JS in any way besides it being the only browser language. Desktop apps use JS because they are browsers, not necessarily because JS was the best language option. An ecosystem has developed around it not because it's good, but because there is no other option for working in a browser.
I worked on the vendor (not auditor) side of accessibility audits, which have similar structural issues as security reports. The amount of latitude individual vendors have in shaping the scope of testing, setting cadence/private review phases, and picking auditors in the first place is stunning, and even though there are public standards (VPAT/WCAG/Section 508/EN 301 549), these degrees of freedom made reports completely incomparable across even vendors with the same commodity software products. Downstream from us, 90% of customers didn't closely scrutinize the contents or quality of the report and just needed it to exist.
The problem is the ultimate consumer[1] of these reports, legal and procurement agents at buying companies, themselves don't care about the actual quality of the report except insofar as it satisfies their own transitive legal/sales requirements, and it's turtles all the way down. This harms users/customers at the end of the day because they don't have time to scrutinize the details of each individual report or have any real power. If we care about the end goal (secure, accessible software), we need for the auditing firms to collaborate we the government, judiciary, and ancillary vendors to tighten standards to include random[2], uniform checks (same auditor, same methodology, multiple vendors at once).
In the US, OSHA designates NRTLs like UL to perform safety testing, which are required everywhere from workplace standards to insurance requirements. In comparison, at least for accessibility, merely having your vendor have any assessment report is likely enough CYA to withstand a legal challenge. I acknowledge the power of recent website lawsuits to use the broader ADA to raise the bar here, but ADA's "enforcement through private lawsuits" enforcement mechanism is spotty and I think won't result in enough structural improvement.
[1] - These reports also serve as PR/marketing, which is probably moreso the case with Mozilla VPN, but in most enterprise software where these assessments are taking place, the marketing side is very much a secondary goal compared to the individual legal/sales relationship that hinges on the report.
[2] - I think removing the opportunity for vendors to fine-tune scope or prepare or respond to concerns (at least until the next review cycle) is a big step in the right direction, but unfortunately, the legal climate is very much all-or-nothing and not good at nuance. Section 508 (I'm not personally familiar with PCI/SOX/etc. but suspect those are similar) is formally speaking "all or nothing" check all the boxes things, and in that climate, good random audits will basically be always-failing, and if you make too hard a standard that even reasonable vendors can't meet with an earnest effort, you'll end up constricting the market into a meta-game of who can hack the auditing process. See federal government procurement.
Even if content license holders don't require Netflix to go above and beyond, building sophisticated DRM is a moat against other streaming services. If Netflix can build a sophisticated VPN-detection service, copyright holders will be much less willing to license their content to upstarts who can't match its capabilities.
In the same vein, it can be rational behavior for a market leader who deals in private information (most online advertising companies) to advocate for consumer privacy protections. It "hurts" them, but if the resulting regulations are so onerous that only incumbent(s) can comply, it can restrict the competitive landscape and paradoxically be advantageous to the existing leaders.
There is a different incentive at play here. Netflix wants to track individuals. I wouldn't be surprised if some other agreement were behind this move, perhaps like other secret agreements:
> There is a different incentive at play here. Netflix wants to track individuals.
Maybe this is naive, but doesn't Netflix already have enough information from the account already? Credit card info, name, address… what does having an IP address get them in addition to all that?
It gives you a lot more information. Where you travel, who you travel with, what types of content you associate with different regions, how much you share with friends, how your preferences change as you hang around different areas and individuals.
Information is deceptive because it's exponential. Detectives often make or break cases on a single lucky clue. Data mining on the Internet is no different. Information goes massively further than most people realize.
No, the second idea is silly. They already have buckets of information. In fact, they have a whole team dedicated to making sure they don't hold onto certain types of PII (Personally identifiable Info)
They have their flaws, but they don't waste time doing useless shit (once they have figured that out).
I mean, obviously they notice which accounts are breaking whatever rules are currently in force.
>If Netflix can build a sophisticated VPN-detection service, copyright holders will be much less willing to license their content to upstarts who can't match its capabilities.
Have you missed the events of the last few years? Netflix's future competition isn't going to be scrappy startups, it's going to be content providers with their own streaming service.
This is a super interesting take and I wonder how closely this reflects Netflix's actual strategic thinking. If so it's an interesting game theory question, because the whole reason rightsholders would want this is to guarantee the value of splitting up distribution to multiple local distributors, most of whom would be well outside the realm of matching Netflix capabilities. But if those players can't match Netflix's DRM and thus rightsholders won't license to them, then all they are doing is giving Netflix is a discount for restricted rights without any additional profit.
I worked on the infra team for a TypeScript-based product downstream of the author. His and his team's approach and writings influenced at a lot of my thinking about core development, and I'll always be grateful for having had the opportunity to learn from them by osmosis.
The success or failure with which any companies hires and retains folks like these is a mystery to me. As the evidence shows, Google (and I suspect almost any >100-person company) is far too much an amorphous glob/slime mold to even have a single coherent approach to this problem. But what are the successful strategies here? How important is it even to retain this kind of talent? How do we know how a company is even doing in this regard?
In my meager-in-comparison 6.5 years at Google, while I certainly saw many amazing SWEs come and many go, I'm far from knowing what the real trajectory was. But I can imagine that for many folks, the perceived trajectory would be a major motivating or demotivating factor to stay or to leave themselves.
Retention of very senior people really comes down to those people and their personal reasons. I don't think it's a group of people who all want the same thing, so it's impossible to make a general retention plan.
And realistically, the reason you want them to stay is because they are someone who has figured out how to operate and impact the organization without needing too much guidance
or direction.
But I think it's why there are all the generic lifestyle perks and competition for titles like "Best Place To Work".
Any large corporation needs to eliminate the bus factor of individual contributors (those who don't pay the management and politics game). You can't have an universally respected engineering legend contradict the board of directors; and the CEO it installs and influences.
Thus, you could argue that faceless, kafkaesque, "out of anyone's control" systems for talent management are a feature with trade-offs, not a bug.
While I have not worked at Google, simply being a user of its products for the past two decades has taught me that so much of its company culture and values have changed; and rarely for the better.
Of course, you don't want a bus factor with executives either although it generally becomes more complicated as you get higher up at a company. You're not going to casually replace a CEO though you're not going to just shut down the company if they get hit with a bus either.
> His and his team's approach and writings influenced at a lot of my thinking about core development, and I'll always be grateful for having had the opportunity to learn from them by osmosis.
Would you like to share with the rest of us a few wisdom here?
There's no secret sauce. The typical aphorisms of software development are generally correct and apply here:
* internalize the cost of core changes and don't push them downstream unless you have a really good reason
* staff a core team to take care of high ROI horizontal efforts that have high fixed cost to do efficiently, e.g. TypeScript migration of a massive codebase
* be careful of what you take on
* complexity and code are costs and not end goals
Reading them is one thing but seeing the principles play out in detail amidst the mess of reality was illuminating for me. The author's blog[1] touches on these themes a bit more and I think are a good glimpse into their stewardship of the TS/JS codebase.
- pay. These companies may be paying more than average and amazing if you compare to non-technical jobs, but as an engineer committing your best years to become great at what you do, you find yourself still not being able to afford even a small flat in the area where you work. Meanwhile you hear your company boasting about another record year and billions flowing in. Unfortunately they won't share those with you. That's demotivating and you feel being taken advantage of.
- lack of flexibility. For example managers pushing for pairing at all cost and with disregard to neurodivergent people. At the same messaging about equality, support for various social issues and so on. That creates dishonest image of the company and makes you think this is all bs and cheap PR rather than genuine care.
- Exploitation. Some companies expect you to do unpaid overtime or that you'll answer your phone outside of work hours. If you want me to work, you pay - I am not a charity.
- Rules designed to appease insecure managers, like having to be in person in the office so the manager can "watch you" and be "hands-on" with his or hers team. Again, company is boasting how great they are for the environment and at the same time they drag hundreds of people each day to the office for no reason. We have great tools to do work remotely, there is no longer a need to be in the office. Sell the office and give employees a bonus.
- open plan office. I just can't focus in those. There is only so long I can stand wearing noise cancelling headphones and I don't want to hear people behind me chatting loudly what they had for lunch, with occasional bursts of laughter that is piercing my ears. I found that often I actually done the work in the evening at home, whereas the whole time in the office I had to pretend I am busy. One time when I was offered a promotion I asked for private office. The CTO said it's only for him and the board. Well, I quit next month.
Winners: The investors who bought Airbnb shares at $68 whose shares are now worth $140.
Losers: Existing Airbnb shareholders who have shares in a company that could have raised $7B in cash but instead raised $3.5B. In theory, the mispricing "cost" Airbnb around 4% of its market cap, so existing common shares are worth 4% less than they should be.
But there's a caveat[1].
Companies like Airbnb (no stable profit) are almost entirely valued on sentiment and expectation. Assuming perfect information and rational actors, in an alternate universe, a direct listing would have resulted in around a $146[2] share price, on top of which the company can raise a secondary listing for $3.5B without affecting the share price. They'd have the same assets and liabilities as they have today, except pocketing the difference (by limiting dilution) instead of handing it to investors.
However, there's no certainty that the valuation would be $146 today had they gone this route. Humans, especially hype-powered retail investors who like to jump on flashy IPOs, are very susceptible to things like IPO bumps when making their buy or sell decisions. An entire field[3] of investing with significant buying power essentially relies on predicting sentiment.
If I'm a common shareholder, sure I'm sad that my shares are worth maybe 4% less than they "should" and that the financial world pocketed the difference, but I'm probably pretty happy my shares are worth what they are and aren't too interested in, for example, rehashing the deal or picking a offering strategy that could have swayed investor sentiment in a way that made my shares worth much more than 4% less.
[1] - The impact of "more perfectly" priced IPOs on market sentiment and thus public valuation is the most significant factor, but a secondary one is the practical aspect that the IPO process consists of a lot of risk, both on the part of the underwriting investment bank who gives the company the money raised in the IPO and the small number of powerful investors get access to (often) preferential pricing from the investment bank in exchange for an advanced commitment to buy the shares and hold for some agreed upon time. This illiquid market is dominated by relatively few actors. An IPOing company who doesn't "play ball" may find itself unable to find someone willing to underwrite their funding round or buy into it at all.
[2] - This theoretical loss per share is computable two ways. The easy way is that they "left $3.5B on the table" which divided into Airbnb's current market cap of $84B is ~4%. The other way is the counterfactual way. Airbnb currently has 600M shares outstanding, having issued 50M shares in the IPO in exchange for $3.5B. So if not for the issuance, Airbnb is worth $80.5B with 550M shares outstanding, or $146 per share.
[3] - https://en.wikipedia.org/wiki/Technical_analysis
Disclosure: I work at Google. You can consider looking at Glassdoor[1], bearing in mind this includes data from several years ago before the wage-fixing bust and excluding natural growth. Looking at offers that I have friends have gotten at both bigtech and not, a good rule of thumb for new offers is to take the upper end of what Glassdoor says and add ~10%. If you want you can also look at H-1B data[2] which usually quotes only guaranteed base salary (and excludes bonus and RSUs). As far as I know there is no public data on this.
H-1B applications often exclude expected value of bonus and RSUs (mine does) and can understate comp at bigtech by ~60k (entry level) to ~150k (mid level) to ~300k+ (senior+).
Yeah the salary listed on the Labor Condition Application is a minimum the company must meet or exceed. When I was on H1-B it was lower than my actual salary.
To reiterate my comment another time this was posted in a comment thread:
I feel this graphic, while informative and delightful, is insidious in its choice of scale and its lack of comparisons.
On scale, it colors +3% to +7% real returns as "neutral". This makes it seem like the stock market is sometimes good sometimes bad but overall it may as well be just okay. I feel that 0% nominal returns, or even 0% real returns, is more honest as a neutral anchor, and even with the latter it would need some comparisons against other asset classes to paint an accurate picture.
On comparisons, it does a disservice to its readers by not adding a tab showing bond yields and a tab showing cash/treasury yields (which would be dark red across the board except light red around 1930).
These slights in the graphic unfairly make the stock market look unfavorable and makes the suboptimal strategy of keeping your money out of the market seem much more favorable than it is.
Adding a specific example, the "worst 20 years" is 1961 to 1981 with a BIG RED -2.0% a year, as if someone loses money and in retrospect would have been better stowing cash away under one's mattress. According to the BLS[1], US inflation was 5.7% over that period, so the mattress strategy yields at best -5.7% in real returns, not at all better than investing in the S&P.
True. Bond yields would be super useful. In 1961, 10yr US treasury yields were 3.84%, and in 1971 they were 6.24%, so you'd have made some 5% nominally by rolling (I don't have 20yr yields at hand), or -0.7% in real terms, beating equity handily.
So, for that period, treasuries > shares > cash.
HOWEVER, I think the chart is not so useful (or intended) as a decision aid for asset allocation (equity/debt/...), but quite good to disabuse people of that notion that "over the long term, you can't lose with equities", or the idea that you can pretty much rely on a 6% real return on your savings.
These returns are before taxes. There is no inflation discount on taxes so your real after tax returns can be negative with a 1% nominal ROI.
Further people don't really invest all their money in year X, and then take it all out in year Y. Cost dollar averaging helps returns and needing to take money out to live off of in down years hurts returns.
Finally the baseline is not the mattress strategy, it's spending all your money now and investing nothing.
I'm pretty sure I replied to you last time ;) One important aspect of looking at returns is comparing different strategies, but another one is in realistic planning. There's a lot of literature and marketing out there touting, in my view, grossly unrealistic numbers like 7-8% annualized compound returns as a reasonable expectation for sticking your money in an index fun on the S&P. Considering the huge differences in the effect of small changes to the annualized returns, it's important people have a realistic idea of the volatility in that expected number when they allocate the amount of money they save for the kind of retirement they want.
This graphic is awesome primarily because it shows that it is not correct to assume that volatility in the equity markets is averaged out completely during a timespan that is comparable to the average savings portion of a career.
Thank you for the reply again :). I saw your comment previously and I think you make a great point. As much as I criticize the chart for being unfairly pessimistic about equity returns, there is far too much literature suggesting you can get 7-8% real returns by parking your money in X, especially in the <20 year time frame for stocks. In comparison this chart is a good factual dose.
My concern is, that while this comparison is useful for people further along in their research trying to understand the volatility of the stock market, this chart has a number of misleading (IMO) traits that can dangerously/unfairly steer people who are newer to managing their own money away from index funds altogether.
I would hope that people see this chart, my comment, yours, and FabHK's excellent comparison to bond yields. But if you have limited attention and are getting started, I would hate for the original link to be the only thing you see.
Speaking for experience with family and friends, too many good people scared by charts like this bought gold in 2011 or trusted mutual fund managers to buy into funds with 4% front-end loads and 2% AUM fees.
It's funny because I agree with all those statements and add "people will listen to Jeff Siegel and just jam their money into index funds and close their eyes until its time to retire". So they lose coming and going (but lose less relying on index funds than buying gold funds).
Personally I think actively managing your money is the better solution, but the active desire not to manage money from so many people (even otherwise active and engaged people like the HN crowd) has led me to being in favor of a stronger govt-backed pension system rather than tax-deferred accounts that hurt our tax base and are a windfall for trustees.
> Personally I think actively managing your money is the better solution...
To what degree do you believe people should actively manage their money? Are you advocating that people should be more active in choosing their distribution of assets across risk:reward categories, or are you advocating for more active trading?
I think the short answer to your question is "both". You need a portfolio with diversified product risk and diversified strategies. You don't need to be a quant to make a basic stab at this with the typical retail portfolio size, there are tons of tools for free on the internet to do this kind of thing. Most people who know enough to not be in managed funds still have no idea how to have anything but basically a 100% long equity market portfolio (I'm intentionally grouping together mostly meaningless 'diversification' between highly correlated segments like midcap/largecap/nasdaq/dow) except to make it long bonds. So I think there's basic product and strategy knowhow to organizing and maintaining a portfolio.
The reason I said both is because of the 'maintaining' part. Without some level of activity, its effectively impossible to be engaged with the market enough to take advantage of opportunities and manage your portfolio to keep enough diversification and reduce the internal correlations in your holdings/strategies.
It might sound complicated but it can be learned and it isn't rocket science, and there is a lot of great technology to assist anyone, not just software devs. Managing your life savings is a better investment of time than many other pursuits, in my view.
Sounds good in principle but how well does advice like this scale? Similar problem to Waze - side streets are great when you're the only one taking them, but once everyone does your advantage is gone. It's hard to expect a large population of amateur investors with no edge to outperform the market. In the general case, what's the marginal return on time and effort spent actively managing your money vs dumping it in a vanguard 50 and learning a different hobby?
That's possible, but it isn't currently the case. At a minimum, having a long vanguard 500 position has a roughly 50% + positive drift - fees chance of success. Part of the long vanguard 500 price bakes in the unlimited theoretical upside that comes along with it. Selling option contracts against that long position to give up that upside beyond a certain price reduces your cost basis and pushes your position's success rate over 50%. Repeated over many events creates a net positive expected value.
Even if there was no edge in the market, as in your premise, it's still the case that the upside of a long S&P 500 equity position is unlimited, and the upside of a long S&P 500 equity position with an option sold against it is limited, therefore would be priced to have a superior chance of success relatively speaking. More market participants would improve the price accuracy of risk, it wouldn't reduce the price of risk to zero.
As for whether its worth it, I think the aggregate effect is significant and, of course, is subject to the benefits of compounded returns, so it doesn't take much to severely outperform your other prospects in the long term. It's up to each of us to decide if its worth learning.
I don't know about every buy-write index investments, but BXM specifically is done with essentially ATM (technically the very first strike OTM I believe) calls against the long position, then held to expiration and cash settled. In a long bull market like the present day, this approach will always underperform the market while having reduced volatility. It should overperform the market in down or sideways markets. Also, volatility induces drag so in a compounded return, all else being equal, lower volatility will yield higher returns.
Diversifying amongst uncorrelated products is an important missing feature to this strategy for the purposes of reducing volatility. If only considering writing covered calls/puts, I'd personally prefer to reduce volatility through diversification, and sell further OTM options to reduce basis so I keep more of the directional risk in each individual position and have lower transactional costs.
Also, BMX holds contracts to expiration rather than benefit from cyclicality in price and implied volatility by closing/rolling options early when they move in your favor or scaling into positions during volatility expansions.
it sounds like you're basically advocating something like an s&p 500 buy-write index investment instead of just an s&p index investment, right? basically selling covered calls on your index investment.
I agree with your premise, and make a pleasant bit of side-income selling covered calls on individual stocks that I own, but it seems like the buy-write indices don't actually fare well, or at least $BXM doesn't. Way worse than I would have expected actually...any idea why?
that's a good point -- i suspect the s&p 500 receives so much attention in comparison to any particular individual stock that there is very probably little change to fish out of its couch cushions
The average person likely doesn't have the time, energy or interest to go beyond the absolute basics of investing which is why simplified advice such as that on Bogleheads is popular.
For those wanting to go a step beyond but still lacking time to go deep, where would you recommend starting education wise? Any links or a syllabus with links would be super useful.
Sure, the guys at tastytrade.com put a ton of effort into educating people and providing a platform (dough.com) that provides this information. I think it has some organizational issues, but my way to start would be to start with the 'Where Do I Start' series they have.
In short, that site largely revolves around the fundamental premises of a random-walk view of prices, and using the time-decay of selling options to reduce the cost basis of holdings over time. There's a lot of treatment and research on correlation of different assets (equities, different kinds of commodities, currencies). My advice if you follow this is to start small and stay actively engaged without getting over-confident at early success. There is a lot of getting used to the mechanics and learning the products so that you can make it a manageable part of your life, time-wise. Also you need to make sure you properly understand the relevant notional values you're dealing with so you can do proper sizing.
BTW quick answer to your first question, here's a sample basket of lesser-correlated assets that the typical index funds that all boil down to being long the market. One of the cores of having a random-walk view of things is that the choice of direction (long/short) is less important than the strategy & cost basis reduction (all of these have liquid option markets):
Long S&P (/ES or SPY)
Long Gold (/GC or GLD)
Short Bonds (/ZB or TLT)
Short WTI Oil (/CL or USO)
Long Euro/USD (/6E or FXE)
Thanks! I'm not sure how shorting bonds/oil is like being long the market? I've always been under the impression that going long had an upward bias, and buying options had a downward bias (but maybe selling them gives some upward bias), so that playing with options was a bit like playing at a casino, where the odds are biased in the favor of the house.
I'm not saying shorting bonds/oil is like being long the market. I just picked a set of underlying assets that are much less correlated than e.g. S&P & Nasdaq. The directional choice (long/short) is really a choice of the investor. Correlations between products are not stable over time, so picking one vs the other is similar to a price bet (i.e. normally distributed). The main point is that the underlying assets are not highly correlated, and that they have liquid derivatives markets that can be used to reduce cost basis.
You may or may not make money on your directional choices, but the core strategy is to be short option premium to make your expected value positive, and to have low internal correlation amongst your assets to reduce volatility in your portfolio.
I'm not 100% sure I understand what you're trying to say re: upward bias/downward bias. However, buying options do have a negative expected value so I agree about that. Selling options is the strategy, and conceptually is similar to selling insurance. Limited profitability, positive expected value. Just like an insurance company, you keep your risk diversified to reduce volatility and keep positions small enough to prevent busting out during drawdowns.
I'm not suggesting buying options (except as part of a spread)
You're missing that there are other investment options.
My college expenses were largely paid for with US Savings Bonds which paid between 4-9%. I cashed in the last one, which ceased paying 9% interest in 2011. My ING Direct 5%, 10-year CD matured last year.
The problem is is with this weird never-land world where we're printing money with no inflation, it's difficult to make money anywhere. My son's 529 plan is in a basket of volatile stock and bond funds. I'll probably make a similar return to my dad's savings bond portfolio, but at substantially higher risk.
> My son's 529 plan is in a basket of volatile stock and bond funds.
Depending on the age of your son, you may really want to consider investing in more stable assets. 20 years is a typically considered a minimum time-horizon for being heavily-weighted in stocks due to their volatility. If your son is older than 5, preservation of the earnings you've already made should start to become a significantly higher priority than high returns.
That's true iff the 529 plan money is "required" for college. If you take instead the view that the 529 plan is just one type of asset that your child has, that you are optimizing for lifetime returns, and that attending vs not attending college is not changed by the 529 returns, then you could take the view that the 529 funds are just as subject to optimal asset allocation as any IRA or brokerage investments the child might have.
Though my kids are 5 and 7, I plan to invest on their behalf in 529, UTMA, and other accounts with their anticipated lifespan as the controlling factor, not which moment I expect to need the funds (because it is overwhelmingly likely that there will be other funds available if needed).
Also, it does not seem to be an accident that they have chosen the range 3% to 7%. The historic average excess return of stocks vs. government bonds is 6.5%. If their range was 2% to 6%, the graphic might look very different.
The key idea is that any human-written change only ever touches ~1-100 files, and there's never any need to store or maintain the entire source tree in your local persistence store. By only ever working on deltas, you can have highly efficient distributed and cached builds. This architecture imposes significant constraints: you must express your build graph in a modular way [2], you collocate your distributed FUSE backend and build system for speed, and all development is over the network. But it comes with many benefits at scale: near-perfect caching, instant "branch switching", fast builds even with deep and wide dep graphs, and excellent affected target determination so flakes/outages in one part of the codebase don't affect the rest of it.
[1] - https://dl.acm.org/doi/pdf/10.1145/2854146 [2] - https://bazel.build/