All of this must require a ton of computing power. Is it local or are you cloud?
We have a data center in Pasadena, but the nice thing is the technology is advancing very rapidly. The biggest computing challenge is in adaptive optics and because of latency issues we need the computer right on the telescope. But you can make computers about the size of this table that have more power than we need.
We use GPU boards that have hundreds of cores per board. We need about a teraflop, so you need a pretty powerful machine. But that's something that's relatively straightforward to put on the telescope and deal with the heat it generates.
What's the expected lifespan of the telescope?
We have a baseline of 50 years, but there's nothing that really wears out. The things that really limit ground-based telescopes is light pollution and climate change.
We need to make sure there's not much human activity around the observatory. When I was a student at the University of California I did my observing on Mount Hamilton outside of San Jose. It's kind of bright right now, that's not really a forefront place any more, although they do interesting work.
We have a location that's remote, dry, and protected by the government, so we should be able to manage the light pollution issue. But climate change is also affecting observatories. Those that are close to the equator are seeing more cloud buildup.
But we think the zone we are in is more like Southern California – it's drying out – and while climate change is bad for everyone it might mean more clear skies in the deserts. It's a small comfort, but we'd be happier seeing carbon concentrations come down.
Is there much in the way of daily maintenance?
There's a huge amount of maintenance, it's like a hospital, it never sleeps – 24 hours a day. The optics slowly accumulates dust and debris and the coatings slowly oxidize so we will clean them regularly in place. But every two years we take them out and completely strip off the coating and put on a new one.
The way we do that is to have one mirror as a spare. We can swap it in, clean the old one, recoat it in a vacuum chamber, and in the meantime swap in the new one. That way we only lose one night of viewing. From a cost/benefit analysis it's a no-brainer to get the extra mirror, even though they are not cheap.
Seven into one, with a spare in the back shed
What's it like to live up on these remote observatories?
It's hard, it's not glamorous, but it's a wonderful place to be; I love it. You're in a place where you can completely focus, there are no distractions. The sky is beautiful and you're sitting up there at one with the sky – it's wonderful.
The only limit is that you don't want to spend too long up there. It's too dry and that's bad for you. So we rotate our crews; you go up for a week and then you go down for a week. When you go to Mauna Kea at 14,000 feet that's a little harder, and you have to limit the number of hours a day, but our site is at 9,000 feet and you can go a week or so without a problem.
The part that's hard is that in the winter you work for 14 or 16 hours, sleep for a few hours, and then get right back to it. That's when the glamour wears off, it's pretty hard. But when you know that this is your window and you have to get great data, and it's all going well, it's gratifying.
How long will you personally be involved in the project?
I'm hoping to see it through to the end. I was there at the beginning and I want to see this thing finished. It would be great to know that you've spent close to 20 years building something that will produce great science for at least 50 years. That would be nice, a real feeling of accomplishment.
What's been the biggest problem that you've had?
Finding enough technical talent. Unemployment among engineers is about 0.5 per cent or less. There are so many exciting things going on – whether it's SpaceX or Tesla or Silicon Valley or biotech – that young people with mathematics, engineering, and analytical skills for big data are in very high demand.
Even though we have an extremely exciting project, it's hard to find really strong engineering talent. It's not just us, the talent is the challenge in any technological field and the immigration system is not making that easier.
As a non-profit we go to the top of the queue for visas, but there's still a problem of staff retention. I know Silicon Valley companies have a much tougher time on this.
Astronomy evokes passion, but sometimes in really specialized engineering areas you have to pay the market price and we've lost people to SpaceX and other tech projects – you don't find these people in unemployment offices, you have to lure them away. There are also personal issues in people's lives and the pressure of families that can take them other places.
You're funded by a consortium of universities and research institutions. Do you get enough support from governments?
In the priority list for funding they look – appropriately enough – at life sciences, climate, and the quality of life. Understanding the universe is on the list, although it could be higher.
Understanding the benefit of curiosity-driven research is often difficult. The payback is enormous, but the timescale and connections are harder to see. It takes a long time, but everything that goes on in the high tech industry and Silicon Valley comes from physics, astronomy, biophysics, chemistry, and biochemistry – it's all rooted in curiosity-driven science but it just takes time.
I'd like to see more federal investment in this kind of science. We are grateful for what we have, but everyone would benefit from investment in pure science. ®