“Death really did not matter to him but life did, and therefore the sensation he felt when they gave their decision was not a feeling of fear but of nostalgia.”
— Gabriel García Márquez, One Hundred Years of Solitude
I quote García Márquez on death not because I live for drama (I do), but because Macondo is described at the end as a city of mirrors and mirages, exiled from memory at the precise moment Aureliano deciphers the parchments. Astro2020 is Melquíades’ parchment for the Astronomers’ city of reflections. It tells the fate of the great mirrors we point to the sky. Of those we lift into the sky. And of perhaps the most important mirrors of all: those we point to ourselves. We near the ending, when some visions become prophecies while others are eaten by ants; mirages, indeed, condemned to another decade of Solitude. It may as well be a hundred years.
I know. I really need to chill.
***
Astro2020’s release is imminent. I take refuge in knowing that fate is written, and await the report with optimism and dread. Regardless of what the Survey prioritizes, we should remain grateful to every wonderful human who served on the Steering / Science / State of the Profession, and Program Committees, as well as all NASEM/SSB/BPA staff who enabled the Survey’s completion. I am thankful to the whole community, which was so invested and participatory in this long process. I will remain thankful even if the projects I love go down in flames.
A Decadal survey is an immense, high-stakes, and somewhat thankless undertaking. Amid the most acute public health crisis in a century, our colleagues on Astro2020 have dedicated untold hours of unpaid time to crafting a consensus vision for the future of discovery. That vision must (a) reconcile with reality (i.e. 💰) and (b) represent a huge community that inevitably fractures into Jeans-length clumps of scientific and parochial interests. Amid constrained budget envelopes, a hyperdimensional discovery space, competing interests, and (mis)aligned incentives, maybe the only way for Astro2020 to achieve maximally proportional representation is to make every community equally unhappy.
It’s thankless. So, thank you.
But the report is also a ghost at the door. A shadow on every wall of my bedroom. My colleagues are stressed. My family and friends are concerned for my health. My every look is a distant stare. I need to write some thoughts down, however incoherent they might be. I’m making them public not because they’re interesting or intelligent, but because our team is one of many cats in Schrödinger’s box, and I have proof of life.
The last five Decadal Survey reports. Their words have given life to five Great Observatories and the world’s most powerful ground-based telescopes. The power of the Decadal is literally hard-coded into the law. They’re a huge deal. You can read all past reports here.
Contents
DISCLAIMERs
🥺
Every single word in this post (as with everything on my dumb blog) is my personal opinion. None of it remotely reflects the opinions of NASA, the Smithsonian, Harvard, or indeed the Lynx team.
Perhaps more importantly, none of this — none of it — is meant to patronize, lobby, or “game” the decadal. I have the utmost respect for everyone on the Committee and panels, and have no doubt that they’ve done right by our community. Besides: It’s way too late to lobby them, anyway 😉.
Lynx is the best thing I've ever worked on
At a meeting in 2016, in front of 50+ friends, a very senior, famous astronomer said “…and then there’s X-ray Surveyor, which we all know will never fly”. The entire crowd laughed. I was pissed, but mostly worried that, if this sentiment were broadly imprinted in the community, it foretold certain doom.
We may well be doomed. I’ve been assuming we’re doomed, not just because the odds are long (they are), but mostly as emotional preparation to lose a thing I’m so passionate about.
But my overwhelming pride in every single aspect of the Lynx study has carried me to a realm beyond fear. Whatever our fate, I feel nothing but gratitude, and lucky to have played a small role in such an inspiring vision for a New Great Observatory. And, no matter the outcome, I think we’ve been immensely competitive at the Survey. Lynx will fly one day. Maybe not from this Decadal. Maybe not the next. But it’ll fly. It needs to. Absent a sub-arcsecond angular resolution, extremely high effective area X-ray Observatory with a microcalorimeter, our eyes will remain closed to most of the Cosmos without it.
The more than 1,000 pages that comprise the Lynx Final Report and supplementary documents, including a 200+ page Cost Book and an entire journal’s worth of refereed academic papers. You can read them all here (except for the stuff rated SBU - Sensitive But Unclassified - by NASA 👀).
The report, which I hope you at least skim through, speaks for itself. So instead of ranting about the inspiring science, here’s a minor but illustrative anecdote:
In those frantic last few months before the Lynx report needed to be finalized, I wrote a little bit of bullshit into the draft and pushed it to Github. Just a tiny bit of bullshit. It was nothing egregious. It wasn’t explicitly wrong, but it wasn’t rigorous. It was a hand-wave. Fine.
Within minutes, my dear friend and mentor Alexey Vikhlinin, co-chair of the Lynx Study, silently walked into my office and taped this “CTFM” sign to my wall. “Check the F!$cking Math”. Our Study would accept no bullshit. Zero. Every simulated Lynx datacube, every calculation, every DRM trade, cost estimate, figure, and statement went through tons of iterations to ensure that it was absolutely rigorous.
Lynx was the ultimate optimization problem. The team found the absolute maximum on the curve in capability-cost space, creating a DRM that was both the “dream” configuration while being achievable. We invert the risk structure. We put revolutionary instrumentation at the focus of the most exquisite X-ray optic ever conceived by human minds, and fly that payload on a proven and (relatively) simple spacecraft, ready for more than two decades of discovery without servicing.
Of all the simulated Lynx datacubes, the Legacy Field is probably my favorite. Read more about it here.
The X-ray community has (both fairly and unfairly) been accused of being insular, of being unable to compellingly sell why X-rays are so essential. I think Chandra’s legacy unambiguously proves this wrong 🙄, but whatever — many think that’s true, and maybe they’re right to some degree.
But Lynx has become something different. I don’t hear jokes about it anymore. I think it is the perfect distillation of what it means to be a Great Observatory. A discovery engine for the age to come. An observatory for all. A facility powerful and flexible enough to pursue questions not yet known. A perfect avatar for the austerity and majesty of X-ray astronomy.
To all 200+ people on the Lynx Team. To Jessica, Feryal, Karen, Harvey, the whole STDT, the whole Study Office, and most especially to Alexey. To Paul Hertz, who has been a visionary leader at APD. To the great Dr. Z. To NASA. To our “competitors”, who are truly our friends, on the Origins, HabEx, and LUVOIR teams. To Astro2020, even if you bury us deep on your list. To the community. Thank you.
The secret fate of all life
I hope Astro2020 is absurdly ambitious. I hope that it quite literally recommends pursuit of a new constellation of Great Observatories. A capital-P Program of Great Observatories. Not just one ambitious Flagship, but three or four. I mean this even if Astro2020 doesn’t want to start with Lynx! I mean this even though such a recommendation would implicitly require the 2030 (and even 2040) Decadals to re-endorse the program, which is wildly out of scope for the Survey’s charter. I mean it with the utmost conviction, even though it’s completely absurd. 🧐
The difference between the absurd and the real, in this case, is exactly twenty years.
I think we should do these absurd things in the face of a cosmic horror perspective on discovery: that it’s the singular thing we leave behind, the only real memory we imprint in the cosmos. The rest is just one big pluck on the guitar string of time, a fading note in the eldritch song. We’ll probably never leave Earth in meaningfully large numbers, no matter how many lunar or martian colonies lay before us in the coming millennia. We will surely not be extending the light of consciousness to the stars, because even the second-nearest star to Earth is a ~fifty thousand year trip on a Generation Ship we should call the Paleolithic.
My favorite image ever. Taken by Al Worden on Apollo 15’s return home. Life highlight: meeting Al, and hearing about the moment he took it.
Assume instead that we are doomed forever to spend our entire cosmic moment on this spectacular paradise of a shining planet. The Strong Anthropic Principle is almost beside the point. We already know it’s the most beautiful ball of mud that hurdles through the void. Earth is the Giving Tree; the epochs of life upon it are a series of concentric rings. Fossils of great reptiles and the earliest protozoa lie just a few rings below us. In time, we’ll sink too.
The Anthropocene now builds algorithms to better sell the apples. We build paperclip machines unto the heat death. We accelerate the expansion of entropy with engines of destruction.
And, lest you think me a cynical nihilist, I think we do great things, too. We can be good stewards of our cosmic moment. We do great things billions of times a day, as life on Earth pursues a trillion unique definitions of meaning. For me: seeing the joy and mystery and wonder on our daughter’s faces at their first encounter with literally everything on the planet. For a bird: finding great worms (?). For worms: … avoiding birds (??🤷♂️). We do our greatest things when we care for one another, especially those we’ll never meet. When we care for the Earth in all its splendor, and for the manifold manifestations of cosmic structure on Earthbound scales, from tiny bugs** to ancient mountains. To be clear, I think these are probably the most important things we can do as short flickers of life in the void. Yes, way more important than building telescopes.
But. BUT.
For the distant horizons of time, those we will certainly never see, I think the product of discovery is our most important lasting message. Let the last Murmur of Earth be to tell the cosmos all we’ve learned. That life here has been some arcane machinery invented by nature as a way to better understand itself. That our tiny planet in its brief existence matters, because complexity increases with decreasing mixing times. That a blade of grass is vastly more complex than a megaparsec-scale intracluster medium, and here’s all we know about both.
We’ve long agonized over the intrinsic value of fields like astronomy. How does a galaxy luminosity function grow the economy? As accessible and “instantly cool” as astronomy is, our contribution to things like the Solow Residual is far less obvious than, say, “gee whiz” research of the 1920s giving rise to the Internet, or investment in mRNA technologies in the early 1990s now saving so many lives from a terrible pandemic. I don’t know how CMB polarization helps humans live better lives. It probably doesn’t. I don’t have satisfying answers to those questions, other than it being one of the manifold ways in which we grasp meaning from the relentless march of entropy.
I don’t know. I’m inarticulate about it because it’s hard to articulate. I think the intrinsic value of fields like astronomy should be self evident, and if you don’t see that, then I don’t have good answers for you. Arrogant and condescending and dismissive, no? Things we probably shouldn’t say to stakeholders. Maybe we need to keep crafting reasonable economic / jingoistic / nationalistic arguments (CCDs! National Competitiveness! Soft Power!), and cross our fingers behind our back. To hide the cosmic horror of it all. That discovery is our Foundation amid a thousand collapsing empires. That it’s our one message in the bottle, bound for shores we’ll never see.
I hope Astro2020 goes bonkers big, because from our tiny island in the boundless cosmic ocean, discovery eclipses everything. Astro2020 at least has the chance to built our great ships of discovery, and so I hope it at least tries.
Let’s write a long f!%cking message for the bottle. Let’s stuff it with maps of cosmic structure on all fractal scales, from gigaparsec filaments to mitochondria and quarks. Let’s tell the story of water across the cosmos. Let’s look at all the other planets and decide over and over again that Earth is where we want to be. Let’s take vacation videos of Old Faithful on the surface of Europa. Let’s push outward on the edges of understanding, and then write down what we see. In ApJ, fine. But also in the great everlasting cosmic memory. Yes, I realize this presents a challenge for Archivists. Not my problem.
Discovery is worth it. Our species can write only one sentence in the book of the Universe. A one-sentence summary of that message in the bottle. Let’s write one that survives the edit of time.
** Full disclosure: I chase flies in my house with a vacuum cleaner.
There are rational reasons for Astro2020 to go small. Few are compelling.
tl;dr: Had I been on the steering committee (ha!), I would have argued that the Report provide a maximally ambitious set of prioritizations for Space. “Do them all. Start with this one. Phase (tech dev) roughly like this.”, basically.
We could then reconcile our Great Observatories recommendation with a number of future realities by draping it in a cascading tree of decision rules. A detailed, nested series of if / then statements to reflect possible budgetary futures, buttressing the “Three Worlds” of Colleen Hartman’s model. Two of these worlds would probably include explicit cost-caps or suggested mission descopes.
I think a unified vision for a new constellation of Great Observatories is both scientifically and indeed narratively compelling enough to win exactly the NASA budget increase needed to enable it (a few percent of the total budget, or ~$400M/yr). This is a lot of money. But a ~2% NASA budget increase isn’t that crazy, and indeed has recent historical precedent with our Planetary colleagues. They were successful for a number of reasons (yes, Culberson), but it’s largely because, among stakeholders, Flagships appeal to each of Flyvbjerg's Four Sublimes.
Both the (current) White House as well as Congress are excited and energized by NASA’s visions for the future. Why surrender to an army that isn’t invading? Despite what you might have heard, they’re not that upset about JWST (much more on that below). The Perseverance EDL this past February was a light in a dark winter of grief and loss. It was super popular at the White House, including with the President, and it did impact the PBR, which was in formulation at the time.
NASA is about to send a Time Machine a million miles beyond the Moon. Its arrival at L2 will be our Dare Mighty Things moment.
The first two Great Observatories are now the last two. They fly ~1984 and ~1994 technology, and yet still maintain their hegemony over the discovery space they cracked open. It is a testament to the raw power of glass in orbit. The twilight of the Great Observatories is regarded with a global sun-drenched affection and gossamer nostalgia. They have woven their way into the global consciousness. Kids tape their images onto bedroom walls. They push outward on the edges of human understanding. They are Generation Ships. Our greatest engines of astrophysical discovery.
Let’s do it again. Let’s at least try.
Okay. That’s the gist of what I was trying to say. A longer and less coherent rant is below.
***
Astro2020 recommending a new Program of Great Observatories is absurd. There is barely enough inflation-adjusted budgetary free energy in Paul’s sand chart to allow for one Flagship out of Astro2020, let alone three or four. I mean, my god, the Flagships from the two previous Decadals still sit on their cleanroom floors! JWST has been a f!&king quagmire, and Roman is the gift that keeps on taking. Flagships eat astronomy. They eat other programs, i.e. (SM/MID)EX and ROSES. We’re better off with probes. Faster, better, cheaper. Let’s pause for the decade, do some tech dev, and give Astro2030 the chance to recommend something big.
While acknowledging I’m a naive zealot, I find few of these points all that compelling. Some are demonstrably incorrect (“Flagships eat astronomy” has basically never been true). Some are good and fair. But I think none are compelling with respect to Astro2020’s charge.
Astro2020 recommending small things (i.e. skipping a Flagship for a Probe and some tech dev) virtually guarantees a contracting budget, and therefore a receding horizon of discovery. It would give OMB and/or Congress the easy out of saying “the Decadal didn’t ask for it, so why should we fund it?”. This is because: (1) Congress does not allocate funds that NASA hasn’t asked for; (2) NASA SMD basically cannot ask for big things that the Astro/Planetary/Helio Decadals haven’t prioritized, pursuant to 42 U.S.C. § 18384 (2014); and (3) both OMB and Congress love the Decadal process because it gives them a “community consensus” to reference.
Dr. Colleen Hartman’s “Three Worlds” model. An excerpt from Lessons Learned in Decadal Planning in Space Science.
I hope instead that Astro2020 aims really big, giving us the chance to land somewhere moderately close. If we go small, we’re guaranteed to do something small. Relative to prior Decadals, decision rules will be a huge deal this decade (it’s an explicit element in Astro2020’s Statement of Task, inspired by the NAS workshop Lessons Learned in Decadal Planning in Space Science).
When recommending high-profile missions, survey committees are advised to explicitly state which aspects of the project are essential to retaining the mission’s consensus priority and which can be further considered during design development to enable cost control.
Clear decision rules for high-profile missions and large facilities that include both de-scope and cancellation options can provide some level of protection against unconstrained cost growth and possible collateral damage to other programs.
— Dr. Alan Dressler, summarizing best practices for Decadal Surveys
I hope these Decision Rules allow Astro2020 to go bigger than the (current) budgetary reality might suggest is feasible. I hope they allow Astro2020 to envision a future in which we grow the wedge, not merely “protect” it.
When the 1991 Decadal Report was released, Hubble was an absolute disaster. The Decadal still endorsed the program of Great Observatories.
Look, there are plenty of lessons to be learned from the development of JWST. But it hasn’t taken that much longer than other flagships have (especially considering its unprecedented complexity).
Hubble (LST) was endorsed in 1972 and launched 18 years later (some of this delay was due to the Challenger tragedy, but HST was hugely behind regardless). It didn’t really work until 1993 (post-COSTAR). Chandra (AXAF) was endorsed in 1982 and launched 17 years later. Spitzer (SIRTF, descoped almost to a probe class mission) was 11 years. JWST is (🤞) 20 years. Roman seems on track (🤞🤞) to be ~16 years.
So maybe some of this frustration has less to do with development hell, and more to do with the long march of time. We need Decadals to recommend a new Flagship every decade, but those recommendations take ~two decades to actually build. Some songs need a half-time beat. Programmatically, NASA APD might need that 10 year impulse frequency to keep the Flagship song in rhythm.
Every single Flagship recommendation from every single Astro Decadal has resulted in … a Great Observatory in the sky, driving a cascade of paradigm shifts and transformative science. Yes, this will soon include JWST, which is being shipped to Kourou as we speak. Yes, this will soon include Roman, who has flight hardware in-hand and is approaching CDR this month.
So, I dunno. I’m bullish. I think if I said all this at a Steering committee meeting (which, trust me, I’d never have been invited to), I would have been shot down by better arguments from vastly more experienced, smarter people. But I still have yet to see a truly compelling argument for Astro2020 passing on a Flagship recommendation. Maybe the Report will provide one. But I doubt it.
***
I say this not because of selfish career goals. Not because I love that GO cash and Granty needs another MacBook Pro. Not because Flagship is my favorite flavor of self-licking ice cream cone.
It’s instead because we are forever bound to the abyssal cosmic dark by 1/r² and λ/D. Our best weapon against this veil of dark is to use big f!%king pieces of glass to see the light behind it. The Cosmos is a symphony on mute; it’s all muffled by the geometry of distance. (1+z)⁴. To the fourth! We can only fight back with big f!%king mirrors. It’s fundamental, unbreakable physics. We see past the veil of dark by collecting more light. I’m in a spiraling ramble because at some point, we’re gonna have to convince Congress why you can’t do LUVOIR with a CubeSat. I’m calibrating a message.
This is why I think, when it comes to astrophysical Discovery, that Faster / Better / Cheaper is axiomatically a “pick two” proposition.
A (marginally out of date) illustration of the Programs of Record for (clockwise from top-left) Earth, Astro, Helio, and Planetary (credit: NASA). It’s actually a lush, balanced portfolio. We need Flagships and Probes and Explorers (and their cross-divisional equivalents).
Yes, there are good, semantic arguments over what “better” means. Yes, “the largest obstacle to low-cost innovation is the belief that it cannot be done”. There are indeed decent arguments in favor of FBC as an underlying principle that I really don’t disagree with. And yes, I really, really do love Probes. I think they make great discoveries. Do great science. I love Galex and NuSTAR and WISE. Love love love. Let’s do more. I hope Astro2020 does recommend a probe, or even a probe line. Really!
But as our community faces the false, narrow-minded question of whether we’d rather have five $1B probes or one $5B flagship (which, trust me, is not the A/B proposition before our community 😉), we should pick the Flagship every time. And the simple reason is that the biggest astrophysical questions — Are we alone? How does the Universe work? How did we get here? — just simply need big f!&cking pieces of glass to address. It’s because Flagships have the power (area), flexibility (payload) and longevity (everything from Phase E budgets to FTEs to serviceability, even at L2) to drive the accelerated expansion of their own discovery space. They’re the platforms long-lived and powerful and flexible enough to answer those questions we haven’t yet thought to ask. Unlike for directed physics experiments, the science case for Flagships grows as they age. JWST will be doing its best and broadest work near the end of its life, just as HST and Chandra are doing right now.
My buddy Prof. Emily Levesque, shortly after HST’s launch in ~1991. Three decades later, Emily is an active user of the very telescope that inspired her interest in astronomy in the first place. It still remains one of the greatest and most powerful optical observatories in history. And there is immese power in the ability of flagships to span decades of discovery: when HST first launched, we didn’t know for sure that exoplanets existed. Now, HST is one of our greatest discovery engines for the field. Note: Emily has kindly granted me permission to use her awesome photo. Thanks Emily.
Many people much smarter and vastly more experienced than me have argued this for many years. Our friend Paul Hertz, Director of Astrophysics at NASA, has eloquently said as much for the past decade. “Great visions inspire great budgets”. “If you plan for a diminishing budget, you get a diminishing program”. “Without JWST, the Astrophysics budget would be $600M / yr”. (I’m paraphrasing here, Paul. Forgive me. And I think that last quote may have been John Grunsfeld). So I’ll stop there.
But I truly believe we should move the Overton Window to talking about a new constellation of Great Observatories. We’ve done it before. It can be done again with merely a few percent added to NASA’s top line. A few percent increase to the NASA budget, directed toward APD, isn't just possible, it has recent precedent with our planetary colleagues. And whether it had more to do with certain members of congress believing in alien squid on Europa is basically beside the point. Both the astute cynic and the bleeding heart are correct. I try to argue this in the below conversation with Mat Kaplan of the Planetary Society. If you listen, drink every time I say the word “literally”.
JWST must be great. It will be.
(🤞🥴)
tl;dr: JWST required the (nearly) from-scratch invention of multiple Unicorns, all under the rule of Conway’s Law, Halberstam’s Law, Hirschman’s Hiding Hand, and Augustine’s XVI Law (inflation, basically). Inventions aside, these laws scale self-similarly to complex projects of basically all sizes, from your kitchen renovation to every megaproject ever attempted. Astro2020 “going small” therefore doesn’t solve “the JWST problem”.
***
Astro2020 should’ve been meeting with JWST long at L2, writing a Nature paper every hour. Instead, when the committee was first forming, Astro2000’s Flagship was dropping nuts & bolts on the cleanroom floor, a decade and a half behind schedule, and a factor of nine to fifteen over budget (depending on what you count as the first “real” cost estimate for NGST). It was such a looming shadow, such a squawking albatross around the community’s neck, that NASA commissioned a survey (via the PAGs) to determine whether Astro2020 should be delayed until after Webb’s launch (um, yeah… let’s be glad that didn’t happen 🥴).
"News that the Transbay Terminal is something like $300 million over budget should not come as a shock to anyone. We always knew the initial estimate was way under the real cost. Just like we never had a real cost for the Central Subway or the Bay Bridge or any other massive construction project. So get off it. In the world of civic projects, the first budget is really just a down payment. If people knew the real cost from the start, nothing would ever be approved. The idea is to get going. Start digging a hole and make it so big, there's no alternative to coming up with the money to fill it in."
— Hirshmann’s Hiding Hand, illustrated. An excerpt from former San Fransisco Mayor Willie Brown’s San Franscisco Chronicle column, July 2013. Note that there are good arguments that the Hiding Hand is, um, bad.
There are many lessons to learn from JWST, from (systems) engineering to technology maturation to minimizing Inventions Needed Per Flagship™ to cost estimates to optimizing risk structure to better exploring the trade space to DEEP BREATH program management to advertising to naming and on and on and on. There are so many lessons to learn, that maybe the first lesson is that we need to learn them.
The big dig NASA — an agency I love — is that it escapes needed reckoning for catastrophes of its own creation. The biggest tragedies (Apollo 1, Challenger, Columbia) have inspired global mourning and sympathy, and I think this is right and human and good. But there is an unhealthy ablutionary side to sympathy in the absence of reckoning. Flower-draped gates cannot take the place of an all-hands-on-deck mission to correct the root cause(s) of the problem(s).
Now, to be absolutely clear, NASA worked its ass off to address problems from the above tragedies. Of course nobody at NASA wanted them to occur, and of course nobody wanted them to occur again. But NASA arguably has never faced full reckoning for these catastrophes. The mere act of NASA returning a fundamentally unsafe flight system** to the launch pad — twice! — was regarded — twice! — as a national triumph. It was just another roll of the dice. We got lucky 133 times.
** I’m sorry I went there, but: there are many words to describe a flight system that killed one human being for every ten launches. “Safe” isn’t one of them.
NASA’s programmatic and cultural issues are fractal, and found at basically all levels. Casey Handmer, in discussing STS and SLS, smartly suggests that this is just a manifestation of Conway’s Law, neither unique to nor entirely the fault of NASA itself. It evinces one reason why complex government organizations are so profoundly bad at learning lessons. At the risk of comparing apples to, um, wars, this is clearly true for the largest and worst of all governmental activities — those for which, despite so much human suffering and loss, a full reckoning never comes. But it’s also true for far less tragic, more nuts-and-bolts, and often worthwhile megaprojects run by large organizations.
A slide that Paul Hertz showed to Astro2020 in the early days of the Survey.
We literally invent new words to explain it. “Undercosted” (🙄🙄🙄) was the word we invented for JWST. Yes, some of this is Hirschman’s Hiding Hand — some parts of “undercosting” (🙄) may be strategic, because if stakeholders knew the real costs and challenges ahead, nothing big would ever get done (see Willie Brown’s famous quote above). Some of it is simply the tyranny of inflation — JWST is expensive for the same reasons your kitchen renovation is expensive. Those early NGST cost estimates were so wildly wrong because they were made in 1996, and we built the damned thing between 2003 and 2021. The tounge-in-cheek version of this is Augustine’s XVI Law.
An excerpt from the JWST Independent Review Board’s Report. This is really just Halberstam’s Law.
Some of it is just Halberstam’s law — humans have a preferential bias for good news, and downplay evidence of problems, especially when reporting up the chain. This creates a growing pile of problems that, while ignored for now, must be addressed later (and later is always more expensive). In the context of Flagships, someone (I can’t remember who, sorry!) at a recent NASA Pause ‘n Learn said “the biggest problems are baked into the cake at the beginning, and they aren’t discovered until you’ve eaten half the cake”.
An excerpt from Bent Flyvbjerg’s 2014 review work on megaprojects. JWST suffered from almost all of these.
Fully assembled and fully tested, NASA’s James Webb Space Telescope has completed its primary testing regimen and will soon begin shipment preparations.
Credits: NASA/Chris Gunn
But maybe the biggest problem with JWST is that, when it was prioritized by the 2001 Decadal, it required the invention of multiple new technologies and spacecraft/payload elements with low average TRL. In other words, JWST’s biggest invention was an entirely new class of problems to solve. The sunshade. Microshutters. Deployable segmented optics naked in the void of space. A moment of inertia from hell. Of course it created a volley of rolling snowballs of schedule slip, scope creep, and an accelerated expansion of cost. And maybe the only way to cut costs while preserving capability was to relax tolerance for risk (which sounds like a great idea, until your telescope breaks a million miles beyond the moon).
The point? I think Astro2020 going small for the next decade does not solve “The JWST Problem”. Whiffing the decade to a probe and some tech dev isn’t the reckoning we need, because Ice Cream Cones still come in small, medium, and large. Yes, let’s build the next generation of Flagships upon a solid foundation of lessons from JWST (we’ve made a helluva start, by the way). But let’s not retreat to the dark. That’s not the goddamned lesson.
I know that we’re embarking right now on missions that could make JWST look small by comparison, and such missions should keep the JWST experience in mind. I would say that the next decadal should wrestle with that problem.
— Tom Young, chair of the JWST Independent Review Board
(a committee on which Fiona Harrison served)
I dunno. I’m kinda done with swoons on the fainting couch. I’m ready to launch this infernal machine and watch it be great. JWST is the best sentence from Blood Meridian. It will witness the birth of stars at the edge of time for two billion less than Apple’s revenue on AirPods in 2020. That’s fifteen billion less than the lump of extra cash the Senate added to the Pentagon’s FY22 appropriation last month. That’s four hundred billion less than the F-35 program, which the leaders of the Military probably don't even want (🥴🥴🥴).
I’ve been scolded for these comparisons. Industry and the military live in such different universes, play on such larger scales, that comparing a Great Observatory’s cost to Apple revenue isn’t a fair or meaningful comparison. I disagree, but fine. I concede the point, and will say only that we could fly all four New Great Observatories with a budget increase smaller than annual consumer spending on Candy Crush gems.
A 1996 cost / timeline estimate for JWST (then NGST). Yes, costing is hard. Yes, this was just a “yardstick” mission concept. No, the people who made this estimate were not idiots or ill-intentioned. But still: we absolutely cannot let this happen again. 🥴
But hey, I don’t mean to downplay the problem. Ten billion bucks is a lot of green. Give the number the respect it deserves. Talk about how this could’ve been better spent, whether on science or society or both. The early cost estimates for JWST (NGST) were insane. It’s embarrassing. It made some people at OMB despise Flagships. We can’t do it again.
What’s clear is that JWST must work and be great. The most complex piece of civil space robotics in all of history now must now write its own hagiography, as Hubble did in the days after COSTAR was inserted. Hubble had been an even worse disaster, and then it became one of the greatest missions in all of scientific history. JWST can do the same.
Pete Worden calling for HST to be abandoned and deactivated, in his famous missive on Self-licking Ice Cream Cones. This was one year before Servicing Mission 1. Note that this isn’t meant to be a hindsight-enabled “gotcha” or dig at Pete Worden, who is a smart and accomplished leader. This was not crazy thinking at the time: Hubble was truly a disaster. And then it was great.
Predictions, smartass?
The report will be out very soon. I’m gonna make some predictions, because 🤷♂️. If I were smart, I’d make some that lie at each extrema of all axes of possibilities. The report will inevitably land somewhere in the middle. Then I’ll look prescient. 🧐
Astro2020 will not invent a Flagship over coffee break. This, of course, is the creation myth for WFIRST — mostly unfair and partly true. And to be clear, I think Roman is great. I’m so excited. I think it will do remarkable things. An HST-like PSF, with that field of view? Holy shit. But Astro2010 also (un-intentionally) created a mission without a constituency. They thought they could make both the physicists and the exoplanet people happy, and all they really did was piss both contingents off. The proposed architecture was confusing in the context of Euclid. When the Report landed, the community spent the first month googling what the hell WFIRST is. They spent the next month assembling circular firing squads. When you-know-what was offered to NASA, a famous astronomer on the Astro2010 committee said we should name the observatory GHOST, the Gift Horse Optical Space Telescope. Anyway, some or all of the above attitudes are embedded at many levels of Astro2020. They’ve been presented with four exquisitely detailed Design Reference Missions that hundreds of brilliant people have dedicated hundreds of person-years to optimizing in the trade space. Astro2020 cannot do better in a few zoom meetings, and they know it. If they’re gonna prioritize a flagship, it’s not gonna be one we haven’t heard of.
Along the above lines, the chances of Astro2020 truly relitigating WFIRST (Roman) are, and always have been, zero. They will endorse mission success for Roman and move on. There is no better way to gut the power of the Decadal process than to un-endorse the last Decadal’s top priority. It’s just not gonna happen.
The Aerospace TRACE will come in high (even eye-wateringly high) on basically everything it was asked to asses, including ground-based projects. This is because Aerospace has precisely zero incentive to come in low. They will use conservative fudge-factors more liberally than I use parm on spaghetti. Nobody wants a JWST-like costing situation again. I get it. I think it’ll be potentially more interesting to see the TRACE results on ground-based projects. Ground-based telescopes are very out-of-wheelhouse for Aerospace.
I think alignment of incentives is the secret sauce of meteorology for the winds of change. In other words, if you want to predict an outcome, a good first step is to write out the full matrix of incentives (for people, constituencies, sub-groups, etc.), and calculate the eigenvector. Astro2020 is charged with pulling “consensus” from an immensely complex, multidimensional trade space. Predicting outcomes (“it’s gonna be ngVLA and HabEx!”) is pretty pointless, and certainly non-actionable. We just need to wait to see the Report. So I’m going to stop wasting your time, and predict that Astro2020 is maximally incentivized to throw as many bones to as many communities as possible. To check as many scientific and political boxes as possible, while still being reconcilable with some version of reality (even an optimistic one). Example: the Exoplanet community is gonna get a big bone, so if they don’t get a Flagship, Astro2020 might compensate by giving them (and many other communities, obviously) a thirty meter.
Dumb exercise, isn’t it? Let’s not game the Decadal. Let’s allow our friends on the committee to finish their difficult job, and then thank them. And then let’s unify, and advocate for the report and our community.
To salt the Earth behind you
Look, I know I take this too seriously. It’s a report. Life will go on. Regardless of what the report says, great science will be done, guaranteed. Yes, even if Astro2020 “goes small”.
And to be clear, I know that all of this hand-wringing is an immense expression of privilege, almost to the point of being utterly tone deaf. Our world faces a fractal universe of vastly more urgent problems to address. Those that impact lives, and the ability of so many to pursue happiness and meaning. To be absolutely clear, I think observing and acting within this universe, rather than the capital-U Universe that surrounds the Earth, is so much more urgent and important.
And, for what it’s worth, I think we astronomers are really lucky. We should be utterly grateful for ongoing investment from a global public that so effusively shares in our wonderment of the Cosmos. We are lucky to have Hubble and Chandra. It will be such a privilege to watch JWST embark on its voyage of discovery. We’d be lucky to get another. There are better things for society to spend money on (except we don’t 🥴).
We should also be grateful to every human being in our own community. Above, I’ve anthropomorphized glorified optical benches of metal and glass. Humans build them, and analyze their light. And so maybe the most Discovery-oriented prioritization that Astro2020 can make is that we work toward a more inclusive future for all explorers. Science is a labyrinth of fractal human biases and hegemonies that exclude and marginalize more brilliant scientists than they welcome. That is why, for all my navel-gazing about telescopes, the State of the Profession pillar of the Survey will arguably be the most impactful and important.
Humans are the engine of our discovery. We write the message.
When my father worked on the Hubble Telescope, he said
They operated like surgeons: scrubbed and sheathed
In papery green, the room a clean cold, a bright white.He’d read Larry Niven at home, and drink scotch on the rocks,
His eyes exhausted and pink. These were the Reagan years,
When we lived with our finger on The Button and struggledTo view our enemies as children. My father spent whole seasons
Bowing before the oracle-eye, hungry for what it would find.
His face lit-up whenever anyone asked, and his arms would riseAs if he were weightless, perfectly at ease in the never-ending
Night of space. On the ground, we tied postcards to balloons
For peace. Prince Charles married Lady Di. Rock Hudson died.We learned new words for things. The decade changed.
The first few pictures came back blurred, and I felt ashamed
For all the cheerful engineers, my father and his tribe. The second time,
The optics jibed. We saw to the edge of all there is—So brutal and alive it seemed to comprehend us back.
And so, I’ll end with thanks.
Thank you to all members of Astro2020. Absolutely none of the above is meant to “game” your Survey or lobby you (it’s a little late for that, anyway 😉). You will do right by our community. I am ready to defend and advocate for the Report and all of its recommendations. All of them. Even if none of them are “what I was hoping for”.
Thanks and ❤️ to the New Great Observatories coalition. To Aki, Heidi, Jason, Scott, John, Marc, Margaret, and Matt. When I read the LUVOIR, Origins, and HabEx reports, I didn’t sneer and jeer at “the competition”. I was floored in awe at the profound visions of discovery your teams put forward. In the two years since, our little ragtag group has become genuine friends, recognizing that while we’ve all designed different boats, we’ll all be jumping in the one we ultimately build, together.
Absurdity has a short half-life. The New Great Observatories idea began as a way to re-frame the narrative, to move the Overton Window from absurd to not-that-absurd. It began as a bit of science marketing. It ended with a group of friends. I think our community is pretty unified, after all.
To Lynx, and the entire team. To Feryal, Jessica, and Karen. You inspire me. To Harvey, a force of nature who gave life to not just one Great Observatory, and instead advocated for the whole damned fleet. It worked.
And mostly to Alexey. Спасибо тебе друг. CTFM. Thank you.
***
If you think I’ve been overly dramatic about telescopes and astronomy, you should’ve met be before we had kids 🥴. My family, especially my wife and two daughters, have brought meaning and perspective to my life. I’m so lucky and thankful. They’re my one sentence. The brightest light in my cosmos. I don’t even need a telescope.
Gabo was right. We do not have a second opportunity on Earth.
