Can we get periodic cost updates?


To Rosa Labs:

In Soylent’s history, costs have changed as is the nature of technical progress. Rhinehart said he was living on roughly $4.25 per 2,000kcal in 2013, which discounts the labor of assembly; when business is added, the costs increase.

In the fullness of time, business will always drive technical progress to lower costs. In 2014, Soylent 1.5 cost $10 per 2,000kcal; while today’s Soylent 1.7 powder costs $7.71 per 2,000kcal, a reduction of 23%.

Of particular interest to me, it is technically possible to get 2,000kcal and adequate micronutrition for under $25/month anywhere in the United States. This is extremely unstable: the amount of meal planning required to keep from overshooting by 100% or more is enormous, and not an inherent or trivial skill for a person to develop. I can do it because I have a polymath background including functional experience in accounting, finance, and project management; the effort I put into beancounting is horrendous, and most people would rather saw their leg off at the knee than do that every day.

More-reasonably, it is possible to put an end to all hunger and homelessness in the United States without an increase in taxes by some careful instrumentation of a Universal Social Security (a form of Universal Basic Income). Politics and philosophy aside, the 2015 numbers for this work out to around $135/month for food, out of a $183/month combined food, clothing, and personal care budget.

Let’s repeat that: We can ensure everyone in the United States has adequate food for no less than $135/month.

My earlier assertion that a $25 food budget is unstable (not to mention unpleasant) warrants some quantification. That quantification is that you can get a half a day’s food at McDonalds for $5, or roughly 1/5 of your $25 food budget. Clothing and soap are actually low-cost expenses (overbudgeted here) and highly-flexible, and can provide some spill-over risk; but you need that 400% headroom incorporated in the portion dedicated to food.

Soylent is stable.

If 50% of your food budget is spent on Soylent, then a full half of the risk of fluctuation is gone. That means $100 of food with a possible 50% overspend requires a $150 budget to remain stable; whereas if $50 of Soylent provides 50% of the calories, then you have $50 of food with a 50% overspend–or $100 total with a 25% overspend, stable at $125. In that scenario, one bag lasts one person two days.

To me, that means Soylent becomes highly interesting at a target price of $4.50/bag (2,000kcal) in 2015. Now, year-over-year, the GDP-per-capita and the total income increase, and so the $135/month a Universal Social Security provides increases both in absolute dollars and in buying power (i.e. faster than inflation). That means even if Soylent gets no cheaper, it will eventually cross that price threshold.

It also means that a $6/bag Soylent or a $5.50/bag Soylent in the next decade will graze the price target. It doesn’t actually need to hit a $4.50/bag price target.

That’s all interesting to me, and highly-verbose and annoying to a lot of other people; and it sets the stage for what should be relevant to Soylent customers regardless:

I would like to hear of Soylent’s updates to the production process in terms of cost savings.

Improved logistics, better packing materials, larger bulk orders and other price negotiating, an improved ingredient source, or even a reorganization of the assembly process to reduce time wasted running back and forth. For customers, business students, and economists, it would be great to have a periodic update of what Soylent has accomplished in the control of costs.

In part I’m just a giant nerd and I like that kind of information. In some other part, I’ve projected a method for remediating the 50,000,000 Americans living without secure access to food, and I want to see progress toward better risk controls on that solution. Either way, it would make reading your blog more fun.


Universal Basic Nutrition is something I’d implement if I were President Of The World.


I would also like to see this kind of information, though the company might see it as proprietary.


+1 <!../>


It’s … difficult. It’s also difficult to discuss without getting into politics because people are like that; although Soylent donates to the World Food Fund so I’ll dip in a little.

Essentially, all things are produced by labor. Money represents labor. If you work for $20/hr, you can induce someone who earns a $10/hr wage to work for 2 hours. Likewise, if 10 workers can produce 100 of a thing per hour, that thing’s cost is 1/10 of all those workers.

Note: the cost thing above is highly simplified. Businesses employ labor, plus buy materials from other businesses (who employ labor to produce those materials), plus pay for power, legal, administration, etc. The sheer number of people whose time is fractionally involved in any product is enormous; many people’s contributions represent less than a second per unit of product produced, while the direct line workers represent more. Ground shipping (trucks, trains) is huge, while intercontinental bulk is tiny (it costs 6 cents per pair of pants to import pants from china, but nearly half the cost of a pair of pants is the truck drivers getting them from the docks to WalMart).

So technology is essentially the reduction of labor to make a thing. In economics, it’s called “technical progress.”

Around 1920, the wooden shipping pallet came into use. One test with a docking crew managed to load and unload canned food into a shipping container in three 16-hour days work days. The same crew palletized, loaded, and unloaded the same canned food in four hours. That means it takes 1/12 as much labor time to load and unload that same product–and thus 1/12 as much wage is paid for that task, even though the dock workers make the same hourly wage.

That reduction in wages per product means a person working as a dockworker would have the same yearly income, yet can afford to buy more things because he doesn’t have to pay for as many labor hours. More product is bought, requiring more shipping; and fewer dockworkers are needed per product bought.

In the end, you necessarily end up with fewer dockworkers, because the total of all income can still only pay for the same number of working-hours at the same wages. If you’re buying more products, it means that someone must be making those products instead of loading them on the ships, trucks, and trains. Half your dockworkers become factory workers, retail cashiers, or mechanics keeping the trucks and trains working–you know, the other people you need.

So what’s this all mean?

It means food can either be half your income or damned-near none of your income.

In 1900, the median-income American household spent 44% of its income on food. In 1950, the median-income American household spent 33% of its income on food. Even then, if 5% of all American households need food supplied, then everyone in the 1950s must cough up 5% of 33% of their income–1.65%. Not much; at the $54k median income, you’d pay $891/year or $75/month. If you’re in the top 10% making $154k, you’d pay $212/month–enough to feed someone all by yourself!

Our level of technology bolted between 1950 and 1990. Take a look. The green is the cost, the blue is the amount of food made, and the red line is the amount of food made per cost. Today, the expense share for a household’s food is about 12%.

12%. That means feeding 5% of all households would cost 0.6%. at the $54k median, that’s $324/year or $27/month. Even the top 10% guy pays only $77/month.

In other words: Feeding everyone can be impossible, or it can be cheap.

There’s another part to this.

We already have welfare. We take money from you, give you nothing, and use that money to feed someone. If we implement another system to replace this, then its total cost isn’t the issue; its cost relative to welfare determines the impact. That $77/month could be versus $83/month, making the system cheaper.

In a UBI system like the Universal Social Security I mentioned, you take a portion of the money and give it back to everyone.

My Universal Social Security makes you an equal shareholder in the economy: everyone receives a percentage of the total productivity. That means you’re still poor or middle-class, and the rich still have like 300 times as much as the average American–and that’s just fine. The poorest American, however, gets a piece of the whole pie–a piece the same size as everyone else gets.

The USS as I’ve proposed it cuts the cost of current welfare out of the general tax (which remains a progressive tax), replaces that portion with a flat 17% tax on all income (business and individual), and then divides that income among everyone. This is handled by the Social Security Administration the same way as Social Security: money is withheld from income each pay period, and it flows through the SSA and into recipients’s hands. Everyone over the age of 18 is a recipient, so a part of what you pay in comes back to you.

As you can imagine, this kind of system relies on wealth–on technical progress to have reached a certain level such that the cost of basic needs are a small fraction of our income.

If the United States implemented my proposed system in 1950, we’d have had a Soviet-Marxist wet dream, and a Soviet-style collapse way before the USSR. Think like the poor get $40k, the middle-class get $50k, and the rich get $60k. It might satisfy the emptiness in some angry idealists’s 501s, but it’ll also basically overturn the entire socioeconomic structure–which doesn’t work even if the end result is better than current. As well, having no class of rich people means you can’t profit from anything expensive, so a lot of next-generation technology can’t be made without government control and an economic dictatorship that basically takes away the people’s wealth to pursue what a small oligarchy decides is worth pursuing (which, granted, isn’t much different than current, if we count high-luxury business executives as an oligarchy–although they do pursue what the rich will buy).

As it stands, it can be done without raising taxes; and incorporating everyone in a share of the economy as such means most people end up with more money. That’s due to an elimination of waste (the welfare system fails to do its job, consumes a lot of resources in failing to do its job, and then leaves us with a damaging level of poverty that consumes more resources; correcting that reclaims those resources, meaning that money can be spent on other things).

The rich end up paying the same and getting back a tiny fraction of that; the middle-class end up paying more while receiving the same absolute number of dollars back, but it’s more dollars than the extra they pay in and so they walk away with more money in total (which is equivalent to just paying less in taxes from the get-go); and the poor… a fraction of a poor man’s income is nothing, and he receives a fraction of a middle-class income in return.

Frame that against the philosophical ideal of getting everyone enough food.

It’s really hard to implement universal food, or universal basic needs in general. It’s not just a complex policy problem; it’s that the cost has to be smaller than the cost of alternative welfare schemes–and, in an absolute sense, it has to be small enough to not exacerbate the problem of poverty in its own right.

Our level of technical progress has reached that point just barely in some wealthier nations. If we can bring technology to developing nations and establish a productive economy, then we can profit from their labor–they make stuff not being made now–and they can trade the products they produce for the things they need to continue making the things we want to buy. That includes food and shelter for the workers, which means suddenly the amount of productivity in Africa is enough to cover everyone’s basic needs as a standard welfare service.

The robots won’t take all our jobs; they’ll just spread us out among more jobs, with fewer people involved in each thing. That means everything’s cheap, we’re all rich, and we can do things like end world hunger. It’s unfortunately not something we can magic into existence (and if technical progress occurs too fast, it generates unemployment faster than the market corrects, and you get an economic collapse).

Annoying, right? We’re literally a mathematical model.


It’s definitely proprietary. I would expect that RL would use their discretion on what level of detail to offer us. “The packing process has been improved, reducing widget correlation time to allow 30% more meals produced in the same effort” etc etc.


It may be something we can tackle when we get more staff. I’ll pass along the request.


Seems to me the corporate priority has been moving away from feeding the huddled masses. Every subsequent product has had a higher price per kcal as well as a more narrow market focus.


I’d be inclined to agree if they’d abandoned Soylent Powder to focus on Drink and Coffiest.

I don’t like their high-price product line myself, and I stick to powder for its economy. It’s perfectly fine for a brand to diversify itself with various options, including premium and convenience options. If their customers don’t care for it, they won’t buy it; and retaining the lower-cost option shows there is a market for both.

At the same time, I dread the day they release a Coffiest powder. I despise caffeine; the drug makes me feel terrible, hence why I drink tea instead of coffee (I’m under 100mg/day). I personally dislike that Coffiest is marketed because caffeine is garbage, but that’s me; Rosa is making a valid business decision, and branching into a powdered Coffiest may be another in the future. The only harm that could do me is marginalizing Powder so completely that it goes away, leaving only Coffiest Powder, which would make me perpetually-ill and so not be an option.

I doubt that would happen anyway. Anyone who’s on Amphetamine or Modafinil would have potential problems adding a larger dose of caffeine into the mix. Nobody wants a bunch of ADHD kids crying that their favorite food substitute is now inaccessible because consuming it causes a horrible drug interaction (although how anyone can stand Amphetamine all the time is beyond me, too, so YMMV).


We are increasing our donations to orgs that feed the hungry. Unfortunately we are literally to small to just send Soylent.


Didn’t mean to imply you weren’t doing anything at the corporate level. Merely that all of your products since powder have been moving away from low cost healthy sustenance (in favor of convenience and variety). I’m sure it’s a smart business move but the early vision of ever more affordable human fuel seems rather distant at the moment.


To play devil’s advocate, even assuming a perfectly benevolent corporation, the big picture stuff probably can’t happen unless they bring in more cash. They’ve also pretty clearly moved to a more opaque development model; there may be exciting developments we aren’t privy to yet.


All of their products since powder have been… other products. They’re options we can buy. I don’t buy them because of economy–liquid products are dumb, and who the hell wants to pay all that to ship water?

As long as they still sell the product I want, they’re not taking anything away.