Marty Sullivan

“Afoot and light hearted I take to the open road. The long brown path before me leading wherever I choose”. Walt Whitman. Licensed psychotherapist and author

[Marty Sullivan @martysull]

A.I. Companies Are Recruiting Electricians and Carpenters by the Thousands https://t.co/usTtc5IgYr via @NYTimes

[Marty Sullivan @martysull]

@elizay560 enough time wasted this morning on your puzzles!

[Marty Sullivan @martysull]

@elizay560 Ok. This one is a puzzle. Think outside the box. Sorry, I asked Grok first. One assumes both numbers are dates, but the first could indicate a room number for example, not the date she was born.

Good one.

[Marty Sullivan @martysull]

@elizay560 One needs to know the math order in solving it. Otherwise it's easy. Give us harder ones.

[Marty Sullivan @martysull]

@elizay560 Oh come on, Eliza. Make them harder.

[Marty Sullivan @martysull]

@elizay560 holy shit, the more I looked the more I saw ... 8.

[Marty Sullivan @martysull]

@elizay560 948

[Marty Sullivan @martysull]

@elizay560 16

[Marty Sullivan @martysull]

@elizay560 76

[Marty Sullivan @martysull]

Do you think he has a point? Or is he uncaring?

[Marty Sullivan @martysull]

@tymboyszn 12 volts

[Marty Sullivan @martysull]

@hcat3h 27

[Marty Sullivan @martysull]

@Magahalla @HumdardAlfaaz @grok And 3 is acceptable.

[Marty Sullivan @martysull]

@Magahalla @HumdardAlfaaz @grok Yessss.

[Marty Sullivan @martysull]

Actually, it’s not too far off from traditional pensions. If you assume an interest rate of 5% and the company contributes 6000 per year over 45 years, you end up with about $1 million. This doesn’t assume that salaries also increase and therefore the contributions would increase. It is not as horrible as you portray.

[Marty Sullivan @martysull]

@sgeine @MattWalshBlog At 5 percent interest contributing 6000 per year you get almost a million dollars. Not bad.

This doesn’t assume that salaries grow overtime so the contributions would grow overtime. It is not far off from what a traditional Pension would be.

[Marty Sullivan @martysull]

@kwelkernbc

I just watched your interview of the U.S. UN Ambassador and the Iranian foreign minister.  You came across as sympathetic to the Iranian cause and hostile to the U.S.   This was not journalism, it was a thinly disguised effort to amplify criticism of the Trump administration’s policy to neuter Iran’s malevolent activities in the Middle East.

You gave softball questions to the Iranian foreign minister, and were clearly hostile to Michael Waltz often interrupting him.  You allowed Mr. Araghchi to speak at length not interrupting him, and asked easy questions such as “Do you trust President Trump” and President Trump said he would annihilate you, are you prepared?

Please reconsider your priorities.

[Marty Sullivan @martysull]

I just watched your interview of the U.S. UN Ambassador and the Iranian foreign minister.  You came across as sympathetic to the Iranian cause and hostile to the U.S.  

This was not journalism, it was a thinly disguised effort to amplify criticism of the Trump administration’s policy to neuter Iran’s malevolent activities in the Middle East.

You gave softball questions to the Iranian foreign minister, and were clearly hostile to Michael Waltz often interrupting him.  You allowed Mr. Araghchi to speak at length not interrupting him, and asked easy questions such as “Do you trust President Trump” and President Trump said he would annihilate you, are you prepared?

Please reconsider your priorities.

[Marty Sullivan @martysull]

I just watched your interview of the U.S. UN Ambassador and the Iranian foreign minister.  You came across as sympathetic to the Iranian cause and hostile to the U.S.  

This was not journalism, it was a thinly disguised effort to amplify criticism of the Trump administration’s policy to neuter Iran’s malevolent activities in the Middle East.

You gave softball questions to the Iranian foreign minister, and were clearly hostile to Michael Waltz often interrupting him.  You allowed Mr. Araghchi to speak at length not interrupting him, and asked easy questions such as “Do you trust President Trump” and President Trump said he would annihilate you, are you prepared?

Please reconsider your priorities.

[Marty Sullivan @martysull]

I just watched your interview of the U.S. UN Ambassador and the Iranian foreign minister.  You came across as sympathetic to the Iranian cause and hostile to the U.S.  

This was not journalism, it was a thinly disguised effort to amplify criticism of the Trump administration’s policy to neuter Iran’s malevolent activities in the Middle East.

You gave softball questions to the Iranian foreign minister, and were clearly hostile to Michael Waltz often interrupting him.  You allowed Mr. Araghchi to speak at length not interrupting him, and asked easy questions such as “Do you trust President Trump” and President Trump said he would annihilate you, are you prepared?

Please reconsider your priorities.

[Kristen Welker @kwelkernbc]

This morning on @MeetThePress, we'll have the only interview with an Iranian official after President Trump rejected Iran's latest ceasefire proposal and reportedly considers resuming strikes after the midterms.

In addition to Iranian Foreign Minister @araghchi, I'll interview U.N. Ambassador @michaelgwaltz, Rep. @RoKhanna (D-CA) and @BillGates.

[Marty Sullivan @martysull]

@SpaceX @elonmusk ChatGPT response to exact same question

The excerpt you shared actually isn’t about orbital data centers. It’s about the labor demands of constructing a large terrestrial data center in Michigan, using it to illustrate how AI infrastructure is affecting the skilled-trades workforce.

That said, if you’re asking whether moving data centers into orbit would solve the kinds of problems described in the article, the answer is: only to a limited extent.

The article highlights a few key bottlenecks:

Construction labor. Massive data centers require hundreds or thousands of electricians, pipefitters, steelworkers, and other trades.
Power supply. AI data centers consume enormous amounts of electricity.
Local infrastructure. They require substations, transmission lines, roads, water systems, and land.

An orbital data center would address some of these—but create others.

What it could solve

Power generation. Large solar arrays in orbit receive far more consistent sunlight than solar panels on Earth, so the facility could generate much of its own electricity.
Land use. It wouldn’t occupy hundreds of acres that might otherwise be used for homes, farms, or industry.
Local electrical grid strain. It wouldn’t compete with homes and businesses for grid capacity.

What it doesn’t solve

The hardware still has to be built somewhere.

Thousands of servers, switches, processors, cooling systems, solar panels, and structural components would still require factories and skilled workers on Earth. Instead of electricians wiring buildings in Michigan, you’d need technicians assembling spacecraft and rockets.

New problems it creates

These are the reasons most experts remain skeptical.

Launch costs. Even if launch prices continue to fall, sending thousands of tons of computing equipment into orbit would cost billions of dollars.
Heat rejection. Computers convert almost all of their electrical power into heat. On Earth, fans and water cooling remove that heat efficiently. In space, heat can only be removed by thermal radiation, requiring very large radiator panels.
Maintenance. If a server fails on Earth, someone swaps it out in minutes. In orbit, servicing could require robotic systems or expensive crewed missions.
Communication delay. Most users are on Earth. Every request has to travel to orbit and back, adding latency and requiring enormous communication bandwidth.

Where orbital data centers do make sense

There are a few niche applications where they could be practical:

Processing imagery from Earth-observation satellites before transmitting only the useful results.
Analyzing data from space telescopes.
Supporting lunar or Mars missions.
Military or intelligence applications where keeping data in space has strategic value.

Could they replace data centers like the one in Michigan?

Probably not in the foreseeable future.

For services like ChatGPT, Google Search, streaming video, cloud storage, and enterprise computing, terrestrial data centers remain much more economical because they’re close to users, connected by fiber-optic networks, and easy to maintain and upgrade.

So the labor shortages described in the article are unlikely to disappear because of orbital data centers. Even if they become technically feasible, they would probably complement Earth-based facilities rather than replace them. The Michigan project—and others like it—reflects the reality that demand for AI computing on Earth is growing much faster than any realistic deployment of computing infrastructure in space

[Marty Sullivan @martysull]

A.I. Companies Are Recruiting Electricians and Carpenters by the Thousands https://t.co/usTtc5IgYr via @NYTimes

[Marty Sullivan @martysull]

@SpaceX @elonmusk

•The X post by @martysull shares a NYT article on AI companies hiring thousands of electricians and carpenters for massive terrestrial data center construction, then directly asks about the practicality of placing data centers in orbit as an alternative.
•Orbital data centers could leverage constant solar power and space’s natural vacuum for cooling, potentially reducing Earth’s energy and land strain from AI’s growing compute demands.
•Practical challenges include extremely high launch and maintenance costs, radiation damage to hardware, signal latency for Earth communications, and current technological limits making near-term deployment unfeasible.