Hacker News Reader

Comment by eloff | Hacker News Reader

I think the rule with Elon Musk is all his estimates are extremely optimistic and you need to multiply them by two or three to get closer to reality.

But I'm sure it serves a purpose in giving his employees an aggressive target to aim at. Even if they miss it, at least they probably got there sooner.

View on HN

Well, it depends on a lot of initial conditions that outsiders can't see. If Tesla wants to throw up a steel frame building in an already empty field in Texas or Oklahoma, that's not hard to do. It would take a few months. And during the Model 3 production ramp up, Tesla put together an entirely new assembly line using on-hand spare parts in a matter of days. So if they have an entire Model Y assembly line sitting in storage rooms in California, Nevada, and New York, which very likely they do, it might only take a few weeks to truck all of that to Texas or Oklahoma. Of course that completely ignores quarantine complications. Which might actually help Tesla. Right now, red states are so eager to get positive "restart the economy" news and millions of recently unemployed people are so eager to get jobs that Governor Abbott of Texas might cut mountains of red tape to green light whatever Tesla needs. I would say it's totally feasible for Tesla to be producing Model Y's from a new facility in Texas by the end of 2020. Certainly not at high volume. Potentially from a tent in the parking lot of an unfinished factory building. They've done it before.

Musk rarely gives estimates, but targets (used to motivate people, for instance, and to overcompensate any unexpected issue or mistake).

I'm looking for his most ambitious and oldest targets and so far, most of them have been reached/exceeded. The most important one being the 500,000 production vehicles run rate for 2020 and 35 GWh battery production for Gigafactory Nevaga. Multiplying that duration by 2 or 3 would bring us to 2027 or 2034... If you can find an even bigger and older target, please share so we can re-evaluate.

His track record on things like self driving is not even close to accurate. But that's to be expected, manufacturing is more engineering and self driving is more research. Engineering is easier to estimate (note, still not easy!)

His track record on the NUMMI plant wasn't very good either. Over-automation and many other issues.

Yet, they keep completing their overall objectives. Why do subsidiary goals like level of automation (remember the fluff bot failure [0]) matter? How could one learn and progress without trials and errors?

[0] https://qz.com/1268647/elon-musk-says-a-fluff-bot-showed-him...

> But I'm sure it serves a purpose in giving his employees an aggressive target to aim at. Even if they miss it, at least they probably got there sooner.

That only works until the employees quickly realize the estimates are impossible bullshit, and stop taking them seriously.

Also lowers morale. If employees can never hit a target, they will become depressed.

Well that’s where management steps in and doesn’t let you disregard the deadline. It makes for a miserable work experience but you can probably achieve a lot that way.

> Well that’s where management steps in and doesn’t let you disregard the deadline. It makes for a miserable work experience but you can probably achieve a lot that way.

That's the kind of management that would go whip the sea for disobedience.

https://en.wikipedia.org/wiki/Xerxes_I#Invasion_of_the_Greek...

> According to the Greek historian Herodotus, Xerxes's first attempt to bridge the Hellespont ended in failure when a storm destroyed the flax and papyrus cables of the bridges. In retaliation, Xerxes ordered the Hellespont (the strait itself) whipped three hundred times, and had fetters thrown into the water.

Reminds me of the iterated prisoners dilemma style party planning I've experienced with my compatriots, where the host expects everyone to be at least an hour late and sets the party start time an hour earlier than they want, and of course the guests expect this and show up two hours late, and so on. The few people (host or guests) who take start time literally are slowly driven insane.

No. Unrealistic schedules don't help getting anything done.

I hope they have real plans for the real work, while the CEO throws out PR bombs.

"First, I believe that this nation should commit itself to achieving the goal, before this decade is out, of landing a man on the moon and returning him safely to the Earth."

Seems to have worked

Once.

With a cherry picked sample of one, you can prove anything.

Shanghai GF was built in time, for example.

Though, I guess in this case it's to put some pressure on Texan government burocracy to hurry up with permits etc.

Over here at Giga Berlin, the local government issued temporary permits extremely quickly (for German standards) but it takes still longer to get all the paperwork done than it took to actually built Giga Shanghai.

> But I'm sure it serves a purpose in giving his employees an aggressive target to aim at. Even if they miss it, at least they probably got there sooner.

That would require an executive board entirely detached from both reality and what their own experts have to say regarding the feasibility of those plans.

He does occasionally hit the estimates - people have provided examples in this thread.

I think he tries to just publicise the unpadded, best-case-scenario estimates rather than the best-guess estimates to push people to eliminate any slack in the project.

It always stuns me what people are able to explain away when it comes to these Captains of Industry.

If we can realize the goal is ridiculous, surely the engineers who work on this full time will realize it's ridiculous.

And maybe their takeaway is the same as ours: "This is a big priority and I need to get it done ASAP". Isn't that exactly what Musk is trying to get across?

Urgency and scarcity also encourage folks to find solutions that they wouldn’t otherwise search for if they could just wait for the status quo approach.

The flip side to doing things that don’t scale is that once those practices become a standard part of your operations you are forced to figure out how to scale them. This causes you to find the scalability hell or high water whereas if you didn’t do the thing in the first place you never would have invented the scalability.

If you were in a meeting where an executive set a clearly impossible goal, would you be more likely to feel a huge swell of motivation or to grumble at the water cooler about clueless business types ignoring the engineers?

Perhaps they got there sooner, but at what cost to their health, relationships, and sanity?

>I think the rule with Elon Musk is all his estimates are extremely optimistic and you need to multiply them by two or three to get closer to reality.

What about all the promises that have been made that never came to fruition? Do we just exclude those from this formula?

Did he say "I promise"?

Look at the time line for the model 3 factor in China.

Reply on news.ycombinator.com