Hacker News Reader

Disclaimer: I work in the space industry for one of SpaceX's competitors. None of my work is in competition with SpaceX (totally different systems), but it bears mentioning. I've also never worked for SpaceX, so a lot of what I'm saying here is opinion and hearsay.

SpaceX does excellent work, and has an amazing track record for such a young space company. That being said, it's important to highlight that the 'absurd level of improvement' that we're seeing not just the product of good engineering management practices. Much of it comes as a result of some pretty extreme expectations placed on the employees working there.

I've lost track of the number of people who've told me that they'd love to work at SpaceX, if only the culture wasn't so oppressively work-centric. These comments have come from students looking for full time jobs having previously interned within the aerospace industry as well as NASA civil servants and contractors who have been putting stuff into space for decades.

The aerospace industry is practically built around overworking its engineers, and, for almost everyone in it, working here was a conscious sacrifice to be able to do some Very Cool Shit. SpaceX takes that concept and turns it up to 11, expecting people to eat, breathe, and live their jobs. The last thing I would want is a firm dedicated to propagating these ideals.

There are a lot of places that will work you like a dog for something that feels meaningless. I don't think anybody feels that way at SpaceX.

SpaceX takes that concept and turns it up to 11, expecting people to eat, breathe, and live their jobs. The last thing I would want is a firm dedicated to propagating these ideals.

On the other hand, if that's what you love it starts to not feel like work. In my experience getting that the match right is super difficult but perhaps that's what SpaceX is aiming for - not really looking for employees but members.

> I think Musk should open an engineering management consulting firm, and hire a few smart people to analyze SpaceX and transfer some of its engineering management methodologies into a set of training modules.

It's a nice idea, but it wouldn't make any sense. Big institutions are arranged a certain way because of the incentives. You're not going to change the incentives because the people at the top fundamentally don't understand the way R&D works. Further, big organizations tend to reward a track record of successes no matter how small, rather than rewarding herculean efforts that only succeed halfway (or fail). The culture is generally anathema to taking risk.

Because the organizations are the way they are, nothing will change unless you change management to the highest ranks and empower people at all levels to make things better. Sadly that rarely happens in big companies; people manage to claw their way up and they're not going to risk that job, power or prestige unless it's a sure thing. Which R&D never is.

I completely agree, and if you look at my past posts, I am quite cynical with most modern companies precisely because I understand that human primal impulses, such as the insatiable drive for status and wealth accumulation, tend to completely sabotage any potential for innovation.

Musk's companies are two of the very few exceptions to this universal empirical reality, and as such underscore the need to try and understand what makes them (SpaceX and Tesla) different, right down to the psychology of the executive team and founders. In my opinion, nothing is more important for the success of a company than the psychological foundations of the executive team and investors. They set the incentive structures, and their subordinates do their best to carry them out.

In most companies, the incentive structure is to get paid as much as possible by playing office politics. Hence, we get golden parachutes, "cost-centers", and an obsession with stock price and quarterly profits.

There is one possible answer that is both obvious and unfortunate (for businesses) - it's because they actually care. For most of the companies, profit is the primary goal and the thing they do is an afterthought (I like to call such entities "toilet-paper companies", because they'd happily drop whatever they're doing and start selling toilet paper if that was more profitable). For SpaceX and Tesla, making humanity multiplanetary and electrification of transport seem to be the primary, terminal goals, while profit is only instrumental. I don't know how many employees care about those goals the way Elon does (but I suspect many, at least in SpaceX - from what people are telling, you wouldn't be able to stand working there otherwise), but if there is enough of them, if that forms the company culture, then it's quite easy to see how the operations will be mostly immune "insatiable drive for status and wealth accumulation". I bet many would love to work for them even for free. I know I would.

Why is this answer unfortunate (for other businesses)? Because you can't really fake caring about something. Many of the problems you see in most modern companies stem from the very fact that they exist primarily to make money, and they comprise of people thinking mostly profit - looking at their company through IAbstractMoneyMakingEntity interface, without really caring about what services or goods the company actually provides.

> This is an absurd level of improvement for an already mature, commercially successful rocket.

What are you basing that on? And an absurd level of improvement can just as easily mean that V1 was just really bad, not that V2 is just that good.

I mean, the Merlin-1D has really good thrust to weight ratio, but the actual amount of thrust and the ISP at sea level is lacking compared to, say, the space shuttle's main engines (even with +30% the 1D is going to come short of 1,000 kN thurst, whereas the RS-25 on the shuttle was 1,860 kN).

It also runs on RP-1 which is a supply-constrained fuel choice that also has a max theoretical ISP that's lower than the ISP that the space shuttle's main engine actually achieved at sea level. And the RS-25's vacuum ISP is so far out of the 1D's reach it's not even funny. Not saying the RS-25 is particularly unique here, there are 17 liquid-fueled rocket engines that have an ISP over 400 in a vacuum, 9 of which also have more thrust than the 1D. The 1D by contrast has an ISP of 340 in a vacuum. 9 engines that have more thrust and more efficiency than the 1D. 9. But somehow Space-X is the one that blows you away?

So sure, it has good thrust to weight - at the cost of using a limited fuel, with low efficiency, and low total thrust. Wow that sure is worthy of just being blown away by absurd level of improvement

The one thing Space-X and Elon Musk do deserve credit on is they sure know how to market effectively. He is really, really good at making mundane tech sound amazing.

To bad people down voted you without commenting on why. Here's what I think:

>>I mean, the Merlin-1D has really good thrust to weight ratio, but the actual amount of thrust and the ISP at sea level is lacking compared to, say, the space shuttle's main engines (even with +30% the 1D is going to come short of 1,000 kN thurst, whereas the RS-25 on the shuttle was 1,860 kN).

The Saturn's F1 engine had an even lower Isp, and was also KeroLox. Your focus on Isp is a bit misguided. Sure the SSME's have a stellar Isp, but guess what? The shuttle solid rocket boosters had a terrible Isp (Shuttle basically ran the RS-25 at lift off only to verify they'd start. They weren't throttled up until well into flight). When it comes to hauling millions of pounds off the launch pad, past the tower, Isp is not what you want. You want mass expelled at a tremendous rate. In other words, why not use electric ion engines with >2000 Isp for the first stage? Thrust to weight matters on lift off is why.

Sure, but the Merlin 1D is still on the very low end of the spectrum for total thrust, which you seem to ignore.

You'll note I only focused on Isp for engines with similar or better thrust than the 1D. And you can't just blindly ignore this. The lower the Isp of the engine the more fuel you need and the worse the thrust-to-weight of the vehicle total ends up being.

Having a very light but very low Isp engine will not get you off the ground efficiently. And if your claim to fame is being cost-effective, efficiency is kinda important.

Or put another way, the RS-25 at 7,480lb each represents a tiny, tiny fraction of the shuttle's total launch weight of 4,400,000 lbs. Using a 5,000lb engine instead would change nothing. Similarly your claim they only ran at lift off to verify they'd start is wrong. The RS-25 represented 30% of the launch thrust.

Even on the fairly light and small Falcon 9, the total vehicle weight is 1,115,200 lbs. The engine weight here is important but far, far from being the single most important attribute. And with a higher Isp engine with enough thrust you could then carry less fuel. Pretty sure you could spend 10,000 lbs on the engines to save 240,000 lbs of fuel (Falcon 9 has ~822,000 lbs of propellant, the RS-25's Isp is +30% the 1D at sea level)

Don't want to diminish SpaceX achievements, but for some comparable systems results could be similarly interesting. For example, Soyuz rockets are launched at least twice (at least 12 per year, https://en.wikipedia.org/wiki/List_of_R-7_launches_%282010%E... ) more often than Falcon last year (highest frequency so far), their pumps and chambers, modules of first stage are manufactured in corresponding numbers (6 pumps per rocket, 20 main thrust chambers for 1 and 2 stages alone, 4 modules of first stage), the rocket is largely - though not completely - manufactured in one closely tied cluster of companies (in Samara city on Volga river, where it is also designed and upgraded). Admittedly these results came after half a century of incremental improvements - but it's interesting to compare SpaceX results with some other offerings.

And how many launches can a single R7 make? That's the difference that has all of our jaws droppping, because it's going to reduce the price for access to space by an order of magnitude.