Dan Hanson

[Dan Hanson @danjhanson]

The data centers are leasing compute to Google and Anthropic to the tune of about 20 billion dollars per year between them. That's an ROI of about 2-3 years. And there is still capacity for training Grok and leasing out more compute.

Elon was first to the gate when the datacenter explosion happened, securing GPUs and energy before both skyrocked in cost and time-to-build. A 1 GW datacenter now costs 60 Billion or so - if you can even secure a location and energy. This is why Google and Anthropic are paying huge prices for SpaceX compute.

[Dan Hanson @danjhanson]

With orbital refueling, Starship can send 100-150 tons of payload to any planet in the solar system other than Mercury or Pluto. The largest probe we've ever sent was roughly 1/10th that size. Starship is a major leap in space capability.

Even if orbital refueling never happens, Starship can still put huge amounts of mass in low earth orbit very cheaply, enabling new space stations, Starlink V3, satellite data centers, etc.

Starlink will pull in about 16 billion dollars this year, with a user base of about 10 million people so far. It's addressable market is in the billions. And Starlink is just one of SpaceX's revenue sources.

[Dan Hanson @danjhanson]

@jota_snchez "There is no such thing as luck; there is only adequate or inadequate preparation to cope with a statistical universe."

- Robert Heinlein

[Dan Hanson @danjhanson]

Complex systems don't work that way. Run the experiment a million times, and sensitivity to initial conditions will get you a million divergent results. What works great today could be a disaster next time.

The real lesson of complexity science is that large complex systems cannot be planned or controlled centrally. The information needed is not available to central planners, and attempts to change the system from the top down will result in unintended consequences. AI cannot do it, because it also lacks the information needed.

[Dan Hanson @danjhanson]

@Tablesalt13 This is being done to condition Canadians to the notion that we belong with Europe, and not with the USA. Preparing the ground for Canada to join the EU And shun America.

[Dan Hanson @danjhanson]

If the goal is to survive the war, the Corsair wins. The Mustang is a great plane, but it did its fighting at long range over a heavily defended Germany. The Corsair flew in a much less dangerous environment, and by the time it came along tactics had been developed against Japanese planes that gave it an 11:1 kill ratio. Big radial engine and heavy armor behind the pilot gave good protection, and air cooling made it less likely to take damage that required a bail-out.

It was more likely than most to kill you in training, though.

[Dan Hanson @danjhanson]

@airmainengineer I believe the white smoke means that a new captain has been selected to lead them.

[Dan Hanson @danjhanson]

Then you must be furious with the people who diverted USAID money to things like LGBT causes and political destabilization operations instead of giving the money to poor children, right?

Just think what could be done for the children with all that diverted money DOGE discovered. But instead of cheering DOGE for finding the graft, you are outraged by it.

While we are at it, don't you think the highest paid members of NGOs should make at most 5X the salary of the lowest paid members? Because a large percentage of USAID money went to pay for multi-million-dollar salaries at these organizations. You must be outraged by that. Right?

[Dan Hanson @danjhanson]

The tax rate may have been 94%, but almost no one paid that because of extensive loopholes and the highest rates only applying to very high incomes.

But if you want to see the effect of tax rates on government income, you should look at that instead. Top tax rates have been as high as 94%, and as low as 28%, and all through those changes the government's share of GDP has fluctuated between 17% and 21% of GDP. And the highest income does not coincide with thh highest tax rates, but whether the country is in recession or not.

[Dan Hanson @danjhanson]

@TheRlyBigShow "Canada is Winning. The rest of the world walked away."

"Hey look! The other lemmings have stopped running! That means we're going to win!"

[Dan Hanson @danjhanson]

There is a hard limit on how much we can diversify. We have spent the last century not just building 'trade' with America, but creating harmonized regulations, developing complex supply chains that span back and forth across the border, creating close cultural, personal and economic ties that are not easily broken without severe damage.

European goods are built to different standards, subject to different regulations, for a people who are not nearly as culturally integrated with us as we are with America. And of course every non-American thing we buy has to cross oceans, increasing costs. There are good reasons why we trade mostly with America.

Also, America is a massive, diversified economy. Replacing our trade with them in favor of distant, very different countries that do not share our values and standards is simply not going to work without serious economic and regulatory costs.

Let the market decide where trade should flow. Policitians do not have the ability or knowledge to centrally manage trade efficiently. That includes central bankers and ex-premiers.

[Dan Hanson @danjhanson]

This mindset is why Elon has been so successful. Hierarchies suck. I've been in software dev for 40 years, and the most effective teams are the ones that dump titles and hierarchies and everyone just does what needs to be done. The worst are teams where every decision has to come from the 'right' source.

And the truth is, I don't care what degrees you have, if all you know is what you learned in school, you are ignorant. Five years after graduation no one cares what degree you have, as your skills and ability to learn will become evident. We had a person with a two year diploma running teams filled with people who have Masters degrees - because he just did the job better. Our head of engineering had a degree in history. Most software engineers have Comp Sci degrees, not engineering degrees.

Outside academia and government, no one cares any more. Software engineers are people who do engineering. Not vibe coding, but proper software engineering. Validation, unit testing, automated build systems, proper version control, PEN testing, etc. What degree you have is irrelevant if you can demonstrate competency.

[Dan Hanson @danjhanson]

@FutureHumanz @CanadianForces I'm pretty sure that means the RCAF will have ground displays and booths at major airshows like they always have.. The Snowbirds will not be flying.

[Dan Hanson @danjhanson]

@NASAEarth looks like a region about 50-75km NE of Winnipeg, Manitoba. Typical grid layout of sections of land, and Gull Lake gives it away.

[Dan Hanson @danjhanson]

@TheRobertDaniro @ArthurMacwaters The train on the right is part of the most efficient rail freight system in the world. Economically, that's more important than faster passenger trains.

[Dan Hanson @danjhanson]

Ok, let's assume a massless, frictionless tube made of perfectly insulating unobtanium, open at both ends. Immanual Velikovsky rises from the dead and flings Venus past us, stopping the rotation of the Earth. So we are all set for our demonstration. What happens?

1. Guy jumps in the hole, and hits terminal velocity quickly. Slower than normal terminal velocity because he's falling inside a tube, and there are aerodynamic effects.

2. As he falls, gravity gets weaker, so terminal velocity drops further.

3. He gets to the center of the earth going very slowly, and stops there permanently to contemplate his life choices.

[Dan Hanson @danjhanson]

@airmainengineer That's the stop sign. It tells you you're gonna stop short of the runway.

[Dan Hanson @danjhanson]

I think forced vaccination was a terrible thing, but we have to do sanity checks here. The vaccination rate for pregnant women varied by country between 36% and 72%. Of those pregnant women vaccinated, the change in miscarriage rates was either nonexistent, or small enough to be buried in the data.

Millions of pregnant women were vaccinated. If 80% or even 20% had miscarried, it would be impossible to hide.

[Dan Hanson @danjhanson]

'Perceived as instant" isn't the right metric. 'Perceived as sluggish' is closer for UIs that have persistent delays over 200mS.

Of course people can perceive shorter intervals when focused om doing so. But if you are scrolling, clicking, and thinking about the work and not the UI, common delays above 200mS will generate comments about sluggish performance and annoying delays. I've done the user tests.

Delays under that amount are usually acceptable, unless fast reaction is a CTQ, like in games. For general productivity apps, 200mS max is a decent rule of thumb to keep users happy about performance.

[Dan Hanson @danjhanson]

In UI design literature 200mS is often listed as the maximum delay the UI can have before it feels sluggish and delays between action and response become perceptable to the casual user.

This is not to be confused with reaction time or ability to perceive smaller intervals when focusing on them. It's more of a general perceptual thing while using the UI.

Love your stuff, Dave. This from a fellow retired software engineer.