Hacker News Reader

I am always surprised to hear how much programmers get paid in the US (third world disclaimer: It's a lot!).

> Average pay for manufacturing engineers is about $78k

This is more on par with pay for (top) programmers or engineers that I know of. In South Africa, actuaries get paid much more than programmers. From what I have seen on HN, US based programmers (in top-ish companies) get paid actuarial salaries. As a mathematician, I have a lot to say about actuaries, but alas, that is for another day.

Maybe someone more knowledgeable can clarify this. Median pay at tech companies is sometimes > $100k annually. In my view as a not-a-US-citizen, that is a lot of money.

Bear in mind you might not have the same way to report salaries. In France, for instance, when employees report salary, they don't include the tax the company pays for them. AFAIK, in the US, salaries are reported including them. I still have a very hard time making apple to apple comparisons. If anyone has a good methodology, please ping

But this shouldn't make a difference relatively. In most of Europe (every country I know salary data of), manufacturing engineers tend to get paid at least the same if not more than programmers.

Yeah and there is also the ridiculous cost of living in a lot of the tech areas. A household earning $117k is on the poverty line in SF.

In SA salary before tax is used. So you lose from there between 20%–45%.

In the US, salaries are not usually described including tax, are they?

When someone in the US says they make $100,000 per year, that is usually exclusive of any tax paid by the company. It is also exclusive of other benefits paid, with health insurance commonly being $10-20k paid by the company, if provided. The employee is responsible for their own income tax.

There’s confusion because in the gig/entrepreneur/consultant world, people will say they “make” $200,000 in business income. But then they bear taxes and benefits out of that amount. So you may “make” more, but do worse than someone who is an employee of a bigger company. That is why it takes a substantial increase in income in the US to justify leaving a stable employer.

South African here. We tend to use products, integrate systems and do custom dev. There's really no reason to pay crazy salaries because "pretty good" is good enough.

When you're competing on the output of developers and the difference between first and second is Facebook and MySpace, get the best you can afford and pay them. You're spending VC money and need the best output as fast as you can get it. Good enough will kill you.

> We tend to use products, integrate systems and do custom dev.

As a young South African, this reality is something of a let down. In the US, it's like you have a country with (if you normalise to mean) 30 cities the size of JHB economically. The access to cutting edge tech jobs is just incomparable. The upside is that if you do break through, your relative success is greater. In the US, someone else is already trying your idea.

What is important is not how much you are paid. What is important is how much buying power you have with the salary you are paid (minus taxes) and how much your society gives you for free (education/health/...)

Until you go on holiday or buy online...

I am curious what you have to say about actuaries.

I think they are one trick ponies (evaluting risk in insurance, pensions and medical funds) that can't appreciate why 0^0 = 1, and if they do, would rather make money for the time it takes to explain.

But, to not sound like a completely grumpy cynic, I think they are intelligent and efficient; moreover, these days people have multiple jobs or facets and the critism is more towards the idea of an actuary than the individuals who are actuaries. I guess we are all in competition and actuaries and CAs are maybe just more honest about competing for one's income.

I am actually interested in actuarial research, FWIW.

There are a lot of engineers that work in manufacturing. Typically they'd be referred under controls and automation or maybe industrial engineering.

Controls Eng for plant level PLC/HMI, Automation Eng for interconnect with the higher layers (SCADA/MES/ERP/Cloud) is the general wording I see. But, yeah lots of mix and match with those words.

In the middle of architecting the systems for a new pack line myself.

Here in Australia "Process Engineer" is used a lot for this type of role.

Economies of scale often get mentioned as if it is some sort of magical property that makes things cheaper. Reality is often far more complicated. Amazon, for example, hasn't really had an improvement in fulfillment costs over time. They've gotten more expensive as they've pushed ideas that grow the company: faster shipping, more products, etc. Also, resource constraints tend to have overwhelming impacts: the cost of extracting oil from the earth got a lot cheaper on economies of scale right up to the point where they had to go after more expensive sources to get enough to fill demand. Economists have a name for this effect: diseconomies of scale.

It is pretty safe to say that economies of scale has gotten us to where we are now with batteries. But that does not necessarily mean the trajectory will continue.

There's also Jevons Paradox[1]. This is that the more efficiently a resource is used the more demand there is for it. So if you increase the efficiency of electric vehicles, there will be more demand for battery packs because the battery pack's use value for transportation increases. For example, if there were technological improvements to Tesla that made them get 50% more range off the same weight of battery packs, that would greatly increase the demand for those battery packs because the value of those battery packs would be greatly increased.

[1]https://en.wikipedia.org/wiki/Jevons_paradox

Ah, I've been looking for a name for this.

So this applies to software systems as well. We had a system where, every time we optimized the underlying database, it became slower . This puzzled us for a while. Turns out that people were used to the slowness. When they found out that the system became faster, they actually started entering more transactions. As more people noticed, the system got overwhelmed again. Until it reached a point where it could finally handle the load.

The variable we failed to see is that people were leaving work earlier now. Even though the system's metrics appeared to degrade, that was a metric that markedly improved. They used to work longer hours to get all their work done.

> We had a system where, every time we optimized the underlying database, it became slower.

One thing I've noticed is that the amount of time I have to wait when performing operations has remained more or constant through the past few decades, even as processor instruction retirement rates skyrocketed and memory/storage/network latencies plummeted. I often find myself trying to pull up a site to do something like track a project or buy a ticket, only to be waiting for 5 to 10 seconds before I am able to read the information and/or perform the operation I was there for.

I always ask myself, "Why on earth am I having to wait at all for this?" I assume it's the same concept of "induced demand" for roadways. Under that theory, in densely-populated areas, traffic will always be as bad as it is, regardless of the number of lanes. Add more lanes, and commute times drop. Then people who didn't tolerate the longer commute times before jump in and start commuting, increasing the load on the roads. This quickly raises the wait time to just below the previous wait time, with all the people in the "margin of toleration" who weren't commuting before now commuting.

In the case of the systems at my company for project tracking, I just assume that people build systems on dependencies in a way that trades latency with the convenience of using a particular API/service with an SLO. The system with the SLO costs money to maintain, and so they have a quota system in place based on priority. The project tracking software comes along and says, "Well, we could write the data into a Prostgres instance with a 10 millisecond response time, but then we've got to worry about backup/restore, availability, and all that stuff. Or we can be the lowest priority traffic for this nifty distributed storage backend that has a support team and SLO with a 7 second response time." The seven seconds of each person's life every time they use the tool is an externality, and it's something few people are ever going to complain about. So, 7 seconds it is.

I'm resigned to waiting an average of 5 or so seconds for anything I do on a computer for the rest of my life, no matter how cheap cycles, storage, or bandwidth get.

This equation has very interesting implications.

E. g. more lanes translates to higher automaker profits at expense of less overall population happiness.

Faster computers translate to higher overall rate of software use at expense of more overall time spent waiting

This is precisely what happens with Mobile Networks. The faster the mobile network gets, the more people uses it which causes things to slow down to zero return of performance.

Ultimately though both Economy of Scale and Jevons Paradox reaches Law of Diminishing returns. Adding more capacity to network or faster DB queries will no longer yield any meaningful output.

For a more direct example with future actors:

Cell phone battery life has stayed mostly stable, even as battery capacities have gone up and processor energy/instruction has gone down. This is because mobile device performance is mostly power-constrained: manufacturers have a target of about a day of battery life as the minimum consumers will put up with, and whenever more energy is available they'll turn up performance or add features until battery life goes back down to that minimum.

But the postal system already had immense economy of scale. If Amazon had to bootstrap its own fulfillment from scratch, the per package cost absolutely would have been slashed as volume increased.

To put some numbers on it, imagine you start off delivering one package by hand a week, literally by hiring someone to drive across the country. This might cost $2000 a week in fuel, salary, depreciation etc.

Now say we're delivering a billion packages a week. The rule quoted would put it at $4 per package.

There's wiggle room to argue each end, but considering it's a rule of thumb operating over nine orders of magnitude, that it's not outright nonsensical is actually pretty good.

I agree it's not some iron law, and it's important to know when and where it can be applied.

I think you have to choose a limited perspective to make a diseconomy of scale for batteries.

CPUs, RAM, Energy and Energy storage have each reached the economy of scale point where innovation is heavily rewarded.

Single core, static dram, Oil, and lithium batteries should each fail at some point soon.

Your examples are services, but the argument is about economies of scale for manufacturing.

As for resource constraints, don't forget substitution.

Yep, there are very often "learning curves" that can be used to accurately predict price trends as manufacturing volume increase. The same thing has played out with solar cells and led to big price drops.

https://en.wikipedia.org/wiki/Swanson%27s_law

The PLC/DCS/SCADA world hasn't really caught up to modern open-source software in a lot of cultural ways. Lots of proprietary components and "pay-to-learn" practices. I expect there's software-engineer types who would venture into the hard-real-time control systems world if it were as open-source and tinker-friendly as something like Rails. Nowadays someone like Microsoft has to go open source to have any hope of capturing developer mind share for something like .NET.

The PLC/DCS/SCADA world hasn't really caught up to modern open-source software in a lot of cultural ways.

That may be a feature. Otherwise you get a Raspberry Pi, with enough attack surface to drive a semi through, to control a pressure regulator.

I wish I went into manufacturing engineering. I wanted to work on robot AI and am there now but studying software is the easy part, you can easily get access to all the software and documentation online for free. Learning automation and manufacturing is more costly in terms of time and money.

More expensive to learn and unfortunately pays much less.

Not lost on me that jobs working on systems to spy on and gas light people pay about three times what manufacturing engineering jobs pay.

One of my mechanical engineering suitemates in college cooped/worked at a Frito Lay chip plant for three semesters and got a job there in manufacturing after graduating, I remember because when he got back he always had boxes of product they gave him after each work period.

I wish every government that invested in renewables would invest at least as much in batteries R&D.

Batteries are getting literally hundreds of millions of dollars a year in R&D investments and have been for decades. We could increase that by a few Billion per year, but it’s not obvious what would make a good investment over just buying more batteries and letting the market develop.

Why don't govs give tax breaks for everything that's battery related? Extraction, shipping, manufacture, sales...? I feel the price would halve instantly.

A carbon tax would help a great deal here as well.

I would be interested in a study on how many years it takes the average "battery breakthrough" that never goes anywhere because the lab technique is completely unscalable to be revisited with more advanced technology and declared impractical a second or third time.

I don't think government funding battery research makes as much sense now as it used to. Absolutely everything runs on batteries nowadays, to the point where they are the main source of expense and weight for a large swathe of consumer goods. The commercial incentive to build a better battery is just huge. I remember reading some years ago that the prospectus of a large oil company, ExxonMobil maybe, explicitly singled out a sudden leap in battery energy density as something that could torpedo their $250b/yr business. It's hard to see how the government finds a way to build something that the private sector can't, when there's a trillion dollars at stake.

Interesting that you mention ExxonMobil. ExxonMobil does, in fact, cite "technological advances in energy storage that make wind and solar more competitive for power generation or increased consumer demand for alternative fueled or electric vehicles" [1] as a risk factor on their annual form 10-K.

In a bit of a twist, worth noting that Exxon Research & Engineering was involved in the invention of the rechargeable lithium ion battery [2]. Quoting from [3], "Thus Stanley Whittingham invented the first rechargeable LIB, patented in 1977 and assigned to Exxon."

[1] https://ir.exxonmobil.com/static-files/37b7e1b7-620b-4248-a3...

[2] https://patents.justia.com/patent/4009052

[3] https://www.currentscience.ac.in/Volumes/117/09/1416.pdf

Disclosure: I worked for ExxonMobil from 2012-2017 but don't have any remaining ties.

Economics doesn’t matter. So long as CO2 emissions continue to rise, subsidies are not only desirable they are absolutely mandatory in any rational sense of the word. The cost doesn’t matter.

I don't know how much funding does in batteries that are designed to be stationery. I.e. where weight does not matter, where they can be bulky and potentially run hot.

There are technologies that are irrelevant for cars or phone (which I do think attract most of the battery R&D) that would be a good fit for a battery the size of a building, designed to store the electricity for a whole city during the night/low wind period.

Government can fund the private R&D easily by, for instance, mandate that renewable sources installed on the grid must provide a baseline 24/24.

But public research is nothing to sneeze at. It brought us the lithium-ion battery (Goodenough was working at the U of Texas when inventing it). Private R&D is good at incremental improvements and solving deployment problems. Public research at funding radically new tech. They work well together.

On manufacturing engineering...

Tesla is likely just an example of a general trend. Whether it's big leaps towards "fully automated" manufacturing or gradual, cheap labour is not as important. It's increasingly a "talent" paced industry.

Manufacturing is becoming less la

Doesn't halve the price of raw materials though which is a big part of Lithium Ion battery pricing.

I'd love to see a cost summary of a typical lithium ion battery if you have it. How much is labor, energy, materials?

Depends entirely on the chemistry, but it's mainly Cobalt that is killing the price. Everyone tries to cut it out as much as possible. Tesla is trying to cut costs with less Cobalt but in general it means worse performance and a lot of safety issues with overheating or exploding. Huge amount of savings from the last years comes from cutting Cobalt by making better chemistries, trying to keep the safety record at the same time.