Latest / The Last Observer: Science and Beyond / New Galaxy Survey Reveals the Universe Isn’t Running Out of Fuel After All

Transcript

New Galaxy Survey Reveals the Universe Isn’t Running Out of Fuel After All · The Last Observer: Science and Beyond

• 0:00
  Speaker 1: Welcome to The Last Observer, a journey through science, history,
• 0:04
  Speaker 1: and the universe's greatest mysteries. From ancient enigmas to the
• 0:08
  Speaker 1: deepest questions at the frontiers of reality, we explore the discoveries, ideas,
• 0:13
  Speaker 1: and mysteries that challenge what we think we know. Stay curious,
• 0:18
  Speaker 1: keep questioning, and let's search for the truth beyond the known.
• 0:27
  Speaker 2: Welcome to The Last Observer Podcast. Today we will talk
• 0:31
  Speaker 2: about a quiet crisis in the cosmos. Star formation across
• 0:36
  Speaker 2: the universe has been collapsing for billions of years, and
• 0:39
  Speaker 2: a new study says the usual explanation is wrong.
• 0:42
  Speaker 3: The usual explanation being that the universe is simply running
• 0:45
  Speaker 3: out of fuel.
• 0:46
  Speaker 2: Exactly. Hydrogen, the raw material of stars. You would think
• 0:51
  Speaker 2: that if star birth dropped by more than half, the
• 0:54
  Speaker 2: hydrogen must have vanished.
• 0:55
  Speaker 3: And it hasn't.
• 0:56
  Speaker 1: It hasn't.
• 0:57
  Speaker 2: That is the punch of a paper published on September
• 1:01
  Speaker 2: 1st in Nature Astronomy. China's Fast Radio Telescope teamed up
• 1:06
  Speaker 2: with the Dark Energy Spectroscopic Instrument, DESI, and they stacked
• 1:11
  Speaker 2: signals from about 2.5 million galaxies.
• 1:14
  Speaker 3: 2.5 million. That is not a small sample.
• 1:18
  Speaker 2: No, they covered something like 12,000 square degrees of sky,
• 1:22
  Speaker 2: looking back over the last 4.5 billion years.
• 1:26
  Speaker 3: So from around the time our sun was born, roughly,
• 1:29
  Speaker 3: to now.
• 1:30
  Speaker 2: Yes, and in that stretch, the cosmic star formation rate
• 1:33
  Speaker 2: fell by a factor of about 2.46. Stars were being
• 1:37
  Speaker 2: born two and a half times faster then than they
• 1:40
  Speaker 2: are today.
• 1:41
  Speaker 3: Meanwhile, the reservoir of neutral atomic hydrogen, HI, only dropped
• 1:46
  Speaker 3: by a factor of 1.35 in the raw measurement.
• 1:50
  Speaker 2: And after they applied conservative corrections for systematics... That decline
• 1:55
  Speaker 2: shrinks to about 1.12. Almost flat.
• 1:59
  Speaker 3: Almost flat. That is the part that feels wrong in
• 2:02
  Speaker 3: the gut.
• 2:03
  Speaker 2: Because textbooks and popular astronomy have this tidy story. Galaxies
• 2:08
  Speaker 2: use up their cold gas. The fuel tank empties. The
• 2:11
  Speaker 2: lights go dim.
• 2:12
  Speaker 3: End of story. This paper says that story cannot be
• 2:16
  Speaker 3: the main driver. Not for the late-time decline.
• 2:19
  Speaker 2: They put it more sharply. Rapid depletion of galaxy HI
• 2:23
  Speaker 2: is ruled out as the primary cause.
• 2:26
  Speaker 3: Okay, walk me through how they even measured this. Neutral
• 2:30
  Speaker 3: hydrogen from distant galaxies is faint. You do not just
• 2:33
  Speaker 3: look through a telescope and count atoms.
• 2:36
  Speaker 2: Right. Individual HI 21 centimeter lines from most of these
• 2:40
  Speaker 2: galaxies are buried in noise. FAST is the most sensitive
• 2:44
  Speaker 2: radio dish on Earth for this kind of work. But
• 2:47
  Speaker 2: even FAST cannot pick out every galaxy one by one
• 2:51
  Speaker 2: at these distances.
• 2:51
  Speaker 3: So they stack.
• 2:52
  Speaker 2: They stack. DESI gives precise redshifts for millions of galaxies.
• 2:58
  Speaker 2: You take those coordinates, you shift every radio spectrum so
• 3:02
  Speaker 2: the expected hydrogen line lines up, and you add them
• 3:05
  Speaker 2: all together.
• 3:06
  Speaker 3: The noise averages down. The average hydrogen signal rises out
• 3:10
  Speaker 3: of the floor.
• 3:12
  Speaker 2: That is the trick. Optical radio synergy. DESI knows where
• 3:16
  Speaker 2: and when. Fast listens for the gas.
• 3:19
  Speaker 3: I like that image. A forest of invisible whispers that
• 3:22
  Speaker 3: only become a voice when you line them up.
• 3:24
  Speaker 2: And once you hear that voice, it is stubborn. The
• 3:28
  Speaker 2: average HI gas fraction at fixed stellar mass barely moves.
• 3:33
  Speaker 2: Less than 0.2 decks over that whole stretch of time.
• 3:37
  Speaker 3: So it is not just a few monster galaxies hoarding gas.
• 3:41
  Speaker 3: The weakness of the evolution is across the population.
• 3:44
  Speaker 2: That is what they report. The relationship between a galaxy's
• 3:48
  Speaker 2: stellar mass and its atomic gas content stays surprisingly stable.
• 3:52
  Speaker 3: Then the mystery relocates. If the atomic hydrogen is still there,
• 3:57
  Speaker 3: why are fewer stars being born?
• 3:59
  Speaker 2: Because stars do not form from atomic hydrogen directly. They
• 4:04
  Speaker 2: form in much colder, denser clouds of molecular hydrogen.
• 4:07
  Speaker 3: H-I is the middleman.
• 4:09
  Speaker 1: Exactly.
• 4:10
  Speaker 2: Neutral atomic hydrogen sits between the big cosmic supply, the
• 4:14
  Speaker 2: stuff flowing in from the cosmic web, and the molecular
• 4:17
  Speaker 2: gas that can actually collapse into stars.
• 4:20
  Speaker 3: So the bottleneck is conversion, atomic to molecular.
• 4:24
  Speaker 2: That is the leading idea the authors push. As the
• 4:27
  Speaker 2: universe expands, the supply of fresh gas from the cosmic
• 4:31
  Speaker 2: web weakens. Average densities drop. It becomes harder to pack
• 4:35
  Speaker 2: HI into the dense molecular phase.
• 4:38
  Speaker 3: The tank looks full. The pump is clogged.
• 4:41
  Speaker 2: A little crude, but yes.
• 4:43
  Speaker 3: I want to sit with how counterintuitive that is. We
• 4:47
  Speaker 3: grew up with the idea of a universe slowly starving.
• 4:50
  Speaker 2: The starving universe is still true in one sense. Star
• 4:54
  Speaker 2: formation peaked about 10 billion years ago and has been
• 4:57
  Speaker 2: falling ever since. That Maddow-Dickinson curve is famous for a reason.
• 5:02
  Speaker 3: The peak of cosmic noon and then the long afternoon.
• 5:06
  Speaker 2: What this work does is change the diagnosis of the afternoon.
• 5:09
  Speaker 2: Not empty cupboard, inefficient kitchen.
• 5:13
  Speaker 3: Molecular gas is known to track star formation more closely.
• 5:17
  Speaker 3: That part already existed in the literature.
• 5:19
  Speaker 1: Yes.
• 5:20
  Speaker 2: The new constraint is on the atomic reservoir at late times,
• 5:24
  Speaker 2: with a sample size and statistical punch that earlier HI
• 5:27
  Speaker 2: surveys could not match.
• 5:29
  Speaker 3: FAST plus DESI is doing something earlier radio surveys could not.
• 5:34
  Speaker 3: Width and depth at once.
• 5:37
  Speaker 2: And they are honest about systematics. They built a forward model.
• 5:41
  Speaker 2: They quote both the raw factor of 1.35 and the
• 5:45
  Speaker 2: corrected 1.12. I appreciate that.
• 5:48
  Speaker 3: When a result is this clean, you want to know
• 5:51
  Speaker 3: they tried to break it.
• 5:52
  Speaker 2: They also stop short of claiming they have solved the
• 5:55
  Speaker 2: whole baryon cycle. They say the numbers are a stringent
• 5:59
  Speaker 2: benchmark for models of accretion, phase conversion, and regulation.
• 6:03
  Speaker 3: Regulation. Galaxies are not passive bags of gas.
• 6:07
  Speaker 2: Supernovae heat and stir the interstellar medium. Black holes launch winds.
• 6:12
  Speaker 2: Radiation ionizes. Gas can be ejected or prevented from cooling.
• 6:17
  Speaker 3: So even if HI is sitting in a galaxy, feedback
• 6:21
  Speaker 3: can keep it from becoming molecular clouds.
• 6:24
  Speaker 2: And on larger scales, cold streams from the cosmic web
• 6:28
  Speaker 2: become unstable in massive halos. The delivery system itself changes
• 6:33
  Speaker 2: as structure grows.
• 6:35
  Speaker 3: I keep thinking about our own Milky Way. We still
• 6:38
  Speaker 3: have hydrogen. We still form some stars, just not like
• 6:41
  Speaker 3: we used to.
• 6:42
  Speaker 2: The Milky Way is a relatively quiet adult galaxy now.
• 6:46
  Speaker 2: Occasional bursts, a future collision with Andromeda that will stir
• 6:50
  Speaker 2: things up again. But the cosmic average is this long fade.
• 6:55
  Speaker 3: Does this change how we think about the far future?
• 6:58
  Speaker 3: The heat death people talk about, or at least the
• 7:01
  Speaker 3: era when star formation finally sputters out.
• 7:04
  Speaker 2: In a way, it delays the fuel crisis version of
• 7:06
  Speaker 2: that story. There is still a lot of hydrogen left.
• 7:10
  Speaker 2: The question becomes how long galaxies can keep converting any
• 7:13
  Speaker 2: of it.
• 7:14
  Speaker 3: Hundreds of billions of years of trickle, maybe, rather than
• 7:17
  Speaker 3: a sudden dry well.
• 7:19
  Speaker 2: Simulations have said for a while that most hydrogen atoms
• 7:22
  Speaker 2: will never become stars. They will stay diffuse. This observation
• 7:27
  Speaker 2: puts a late-time empirical stake in the ground for that idea.
• 7:31
  Speaker 3: I want to talk about FAST itself for a minute,
• 7:34
  Speaker 3: the 500-meter aperture spherical telescope. That dish in a karst
• 7:39
  Speaker 3: depression in Guizhou.
• 7:40
  Speaker 2: It is a monster. A fixed spherical reflector with a
• 7:44
  Speaker 2: moving feed cabin. Built for sensitivity on the 21-centimeter line
• 7:48
  Speaker 2: and pulsars and fast radio bursts.
• 7:51
  Speaker 3: And now it is writing cosmic history by listening to
• 7:54
  Speaker 3: the average breath of millions of galaxies.
• 7:58
  Speaker 2: DESI is the other half. Installed on the Mayol telescope
• 8:02
  Speaker 2: at Kitt Peak. 5,000 robotically positioned fibers. Redshifts at industrial scale.
• 8:09
  Speaker 3: Two instruments that were not designed as a pair suddenly
• 8:13
  Speaker 3: becoming one experiment.
• 8:15
  Speaker 2: That is modern astronomy. You stitch surveys. You steal precision
• 8:19
  Speaker 2: from one wavelength and sensitivity from another.
• 8:22
  Speaker 3: Is there a risk they are missing a hidden population?
• 8:26
  Speaker 3: Galaxies so gas-poor they fall out of the stack? Or
• 8:29
  Speaker 3: so gas-rich they bias it?
• 8:31
  Speaker 2: That is exactly the kind of systematic they try to model. Selection,
• 8:36
  Speaker 2: beam effects, confusion, the way stacking weights different masses, they
• 8:40
  Speaker 2: call the corrections conservative.
• 8:42
  Speaker 3: Still, I would want independent confirmation from other HI surveys
• 8:47
  Speaker 3: as they come online.
• 8:48
  Speaker 2: So would they, I suspect. A result this important should
• 8:51
  Speaker 2: be poked from every angle.
• 8:53
  Speaker 3: Let me try to say the scientific importance out loud.
• 8:57
  Speaker 3: This is not acute anomaly. This is a constraint on
• 9:00
  Speaker 3: how baryons cycle through galaxies in the last third of
• 9:03
  Speaker 3: cosmic time.
• 9:05
  Speaker 2: Yes, galaxy formation theory has to produce a steep drop
• 9:08
  Speaker 2: in stars without a matching drop in atomic gas. That
• 9:12
  Speaker 2: is a tight box to live in.
• 9:14
  Speaker 3: If your simulation drains HI too fast, you are wrong.
• 9:18
  Speaker 3: If it keeps forming stars at a high rate because
• 9:20
  Speaker 3: the HI is still there, you are also wrong.
• 9:23
  Speaker 2: You need the conversion physics to be the throttle.
• 9:26
  Speaker 3: What would that physics look like in a single galaxy?
• 9:30
  Speaker 3: A spiral like ours, say.
• 9:32
  Speaker 2: You have a disk of atomic hydrogen extending well beyond
• 9:35
  Speaker 2: the bright stars. Inside, molecular clouds live in spiral arms, dust-shielded, cold.
• 9:42
  Speaker 2: Stars ignite there.
• 9:44
  Speaker 3: If the incoming flow from the halo slows and the
• 9:47
  Speaker 3: disk becomes less dense, those molecular clouds become rarer even
• 9:51
  Speaker 3: while the outer HI disk still looks healthy.
• 9:54
  Speaker 2: That is the picture. A galaxy can look gas-rich in
• 9:57
  Speaker 2: 21-centimeter maps and still be anemic in new stars.
• 10:02
  Speaker 3: There is something almost melancholy about that. Plenty of raw material.
• 10:07
  Speaker 3: No workshop.
• 10:08
  Speaker 2: That is a sentence. It is, and it is accurate
• 10:19
  Speaker 2: to the measurement.
• 10:21
  Speaker 3: Should we be skeptical that Nature Astronomy would publish something
• 10:24
  Speaker 3: this clean? Clean results make me twitch.
• 10:27
  Speaker 2: The tension is real and previously suspected, but the precision
• 10:32
  Speaker 2: is new. I do not read this as a miracle paper.
• 10:35
  Speaker 2: I read it as a survey paper that finally had
• 10:38
  Speaker 2: the numbers.
• 10:39
  Speaker 1: Fair.
• 10:40
  Speaker 3: Strength of evidence is high for the observational claim. The interpretation,
• 10:44
  Speaker 3: the conversion bottleneck, is well-motivated but not uniquely proven.
• 10:49
  Speaker 2: That distinction matters. They measured omega-HI. They compared it to
• 10:54
  Speaker 2: the star formation history. They pointed at the missing step.
• 10:58
  Speaker 3: To close that step, you need molecular gas surveys over
• 11:01
  Speaker 3: the same volume and time, with the same honesty about systematics.
• 11:05
  Speaker 2: ALMA and future facilities will keep chewing on that. This
• 11:09
  Speaker 2: paper just moved the atomic side of the ledger.
• 11:12
  Speaker 3: I also keep circling the human scale of Four and
• 11:16
  Speaker 3: a half billion years. Earth has been around for a
• 11:19
  Speaker 3: bit more than that. Complex life is a late guest
• 11:22
  Speaker 3: in a universe that was already turning the lights down.
• 11:25
  Speaker 2: We arrived in the quiet era. Most of the stars
• 11:28
  Speaker 2: that will ever exist already exist.
• 11:31
  Speaker 3: That fact used to ride along with the fuel crisis story.
• 11:35
  Speaker 3: Now it rides with an efficiency crisis story. Same dusk,
• 11:39
  Speaker 3: different reason.
• 11:40
  Speaker 2: Does that change how it feels to you?
• 11:42
  Speaker 3: A little. Starvation is simple. Inefficiency is more like a
• 11:46
  Speaker 3: civilization that still has warehouses full of grain and cannot
• 11:50
  Speaker 3: bake bread. Stranger. More specific.
• 11:54
  Speaker 2: And more interesting, because it points to physics we can
• 11:57
  Speaker 2: still test. Turbulence. Metallicity. Dust shielding. Magnetic fields and clouds.
• 12:04
  Speaker 2: Accretion rates onto halos.
• 12:06
  Speaker 3: A to-do list instead of a tombstone.
• 12:09
  Speaker 2: Yes.
• 12:10
  Speaker 3: There is a small historical echo here. For a long time,
• 12:13
  Speaker 3: people thought the sun would run out of hydrogen in
• 12:16
  Speaker 3: a simple way. Then we learned about cores and shells
• 12:20
  Speaker 3: and mixing.
• 12:21
  Speaker 2: The universe as a whole is getting a similar upgrade.
• 12:24
  Speaker 2: The inventory is not the whole story. The pathways are.
• 12:28
  Speaker 3: I want listeners to leave with the numbers, because they
• 12:31
  Speaker 3: are the spine. Star formation down by about 2.5, neutral
• 12:36
  Speaker 3: hydrogen down by maybe 10-30% depending on how hard you correct.
• 12:40
  Speaker 2: Sample of roughly 2.5 million galaxies. Redshift range 0 to 0.41.
• 12:48
  Speaker 2: Fast HI spectra plus DESI optical redshifts. Published September 1, 2026.
• 12:54
  Speaker 2: Nature Astronomy.
• 12:56
  Speaker 3: Title worth repeating. Weak evolution of cosmic atomic hydrogen over
• 13:01
  Speaker 3: the past 4.5 billion years.
• 13:04
  Speaker 2: Weak evolution. Those two words are doing a lot of work.
• 13:08
  Speaker 3: If I were writing a model of the late universe tonight,
• 13:11
  Speaker 3: I would tape that phrase above the monitor.
• 13:14
  Speaker 2: Same. And I would remember that molecular gas is the
• 13:17
  Speaker 2: other column, the one that still has to fall with
• 13:20
  Speaker 2: the stars.
• 13:21
  Speaker 3: There is also a quiet international story, a Chinese radio
• 13:25
  Speaker 3: observatory and a multi-institution dark energy instrument sharing the sky.
• 13:30
  Speaker 2: Science at this scale is rarely a single flag. The
• 13:33
  Speaker 2: paper is a collaboration. The measurement belongs to anyone who
• 13:36
  Speaker 2: can reproduce it.
• 13:38
  Speaker 3: Which is how it should be, especially when the claim
• 13:40
  Speaker 3: is this large.
• 13:42
  Speaker 2: Let me voice one last skepticism. Could star formation histories
• 13:47
  Speaker 2: themselves be biased? If the CSFRD decline is overestimated, the
• 13:52
  Speaker 2: contrast shrinks.
• 13:54
  Speaker 3: Possible in principle, but the decline of cosmic star formation
• 13:58
  Speaker 3: is one of the more robust results in extragalactic astronomy.
• 14:02
  Speaker 3: Different tracers, different teams, same downward slope.
• 14:06
  Speaker 2: So the contrast is likely real, even if both sides
• 14:09
  Speaker 2: move a little in future papers.
• 14:12
  Speaker 3: That is where I land. Real tension. Best current explanation
• 14:16
  Speaker 3: is a failing conversion from atomic to molecular gas as
• 14:20
  Speaker 3: the cosmos thins out.
• 14:21
  Speaker 2: And a universe that is not empty, just less willing
• 14:25
  Speaker 2: to light new fires.
• 14:27
  Speaker 3: When you look up tonight, most of those stars are
• 14:29
  Speaker 3: old work, relics of a more generous age.
• 14:32
  Speaker 2: The hydrogen is still out there, dark, spread through disks
• 14:36
  Speaker 2: and filaments, waiting for conditions that are getting rarer.
• 14:40
  Speaker 3: That is a good place to rest the thought, not
• 14:43
  Speaker 3: a finale, a revised question.
• 14:45
  Speaker 2: Why is it increasingly difficult to form stars when the
• 14:49
  Speaker 2: sea of hydrogen has not gone away?
• 14:52
  Speaker 3: That question is going to follow astronomy for a while.
• 14:55
  Speaker 2: Good. Questions that last are the point of a measurement
• 14:59
  Speaker 2: like this.
• 15:00
  Speaker 3: Thanks for sitting with a quiet universe and not rushing
• 15:03
  Speaker 3: to make it louder than the data.
• 15:05
  Speaker 2: Thanks for distrusting the tidy fuel tank story until the
• 15:08
  Speaker 2: numbers arrived.
• 15:10
  Speaker 3: This has been the Last Observer podcast.
• 15:13
  Speaker 2: Keep looking. The night is still full of unused hydrogen,
• 15:17
  Speaker 2: and that is stranger than emptiness.
• 15:19
  Speaker 3: There is another layer I cannot drop. What this means
• 15:22
  Speaker 3: for so-called quenching.
• 15:24
  Speaker 2: The word astronomers use when a galaxy stops forming stars
• 15:28
  Speaker 2: in any serious way.
• 15:30
  Speaker 3: People talk about mass quenching, environment quenching. morphological quenching, a
• 15:36
  Speaker 3: whole taxonomy of ways a galaxy goes quiet.
• 15:39
  Speaker 2: This result does not pick a single villain. It says
• 15:42
  Speaker 2: whatever mix of those processes is operating, the net effect
• 15:46
  Speaker 2: is not to empty the atomic tank on a cosmic average.
• 15:50
  Speaker 3: You can shut down the factory without selling off the warehouse.
• 15:54
  Speaker 2: In cluster galaxies, ram pressure can strip the loosely bound HI.
• 15:59
  Speaker 2: That does remove fuel. But the cosmic average is dominated
• 16:03
  Speaker 2: by field galaxies, the ordinary majority.
• 16:06
  Speaker 3: So the global number is telling you about typical galaxies,
• 16:10
  Speaker 3: not just the ones getting mugged in dense environments.
• 16:13
  Speaker 2: That is why the stability at fixed stellar mass matters.
• 16:17
  Speaker 2: It is not an average hiding a massacre in one
• 16:20
  Speaker 2: corner of the sample.
• 16:22
  Speaker 3: I keep imagining a graduate student in 2030 using this
• 16:26
  Speaker 3: as a prior. You do not get to invent a
• 16:28
  Speaker 3: model where HI crashes after redshift 0.4.
• 16:33
  Speaker 2: The box is smaller now. That is progress, even when
• 16:36
  Speaker 2: the mystery gets more interesting.
• 16:39
  Speaker 3: Talk about dust and metals for a second. Molecular hydrogen
• 16:42
  Speaker 3: likes to form on dust-grained surfaces.
• 16:45
  Speaker 1: Yes.
• 16:46
  Speaker 2: Neutral atoms bounce around on a grain. Meat become H2.
• 16:51
  Speaker 2: Without dust, that chemistry is slower.
• 16:54
  Speaker 3: Late-time galaxies are more metal-rich on average than galaxies at
• 16:58
  Speaker 3: cosmic noon, so you might expect easier molecule formation, not harder.
• 17:03
  Speaker 2: That is a real counter-argument. Metallicity should help. If conversion
• 17:08
  Speaker 2: is still struggling, density and dynamics have to be winning
• 17:11
  Speaker 2: the fight.
• 17:12
  Speaker 3: Lower midplane pressure, weaker spiral shocks, less frequent cloud collisions.
• 17:17
  Speaker 2: A calmer interstellar medium that cannot compress itself often enough.
• 17:23
  Speaker 3: I like that more than a missing ingredient story. The
• 17:25
  Speaker 3: ingredients are present. The weather inside the galaxy has changed.
• 17:30
  Speaker 2: And the weather outside, too. Cosmic web filaments are less
• 17:34
  Speaker 2: dense as expansion wins. Accretion becomes hot and slow instead
• 17:38
  Speaker 2: of cold and fast.
• 17:40
  Speaker 3: Cold mode versus hot mode. A classic divide in galaxy formation.
• 17:45
  Speaker 2: At early times, gas streams in cold and can reach
• 17:49
  Speaker 2: the disk. Later, its shock heats in massive halos and
• 17:53
  Speaker 2: has to cool again, which is inefficient.
• 17:56
  Speaker 3: HI can still exist in an extended disk even when
• 17:59
  Speaker 3: the fresh supply is a drizzle.
• 18:02
  Speaker 2: A reservoir filled in the past, sipping from a thinning stream.
• 18:06
  Speaker 3: Does anyone still argue that we have simply mismeasured the
• 18:10
  Speaker 3: local HI density?
• 18:11
  Speaker 2: Local 21-centimeter surveys are pretty mature. The nearby universe is
• 18:16
  Speaker 2: the easy part. The hard part was always the look-back.
• 18:20
  Speaker 3: Which is why stacking at DESI redshifts is the move.
• 18:24
  Speaker 2: Yes, you accept that you lose galaxy-by-galaxy detail in exchange
• 18:29
  Speaker 2: for a cosmic mean you can trust.
• 18:32
  Speaker 3: There is a philosophical cost. Stacking is a chorus. You
• 18:36
  Speaker 3: cannot hear the soloists.
• 18:38
  Speaker 2: True, some galaxies may be drowning in HI, and some
• 18:43
  Speaker 2: may be dry. The paker argues the average relation with
• 18:46
  Speaker 2: stellar mass does not swing wildly. which limits how crazy
• 18:50
  Speaker 2: the hidden diversity can be.
• 18:52
  Speaker 3: I will take that. Not perfect knowledge, better than folklore.
• 18:56
  Speaker 2: Audience-wise, I think this story works because it violates a
• 19:00
  Speaker 2: sentence everyone thinks they already know.
• 19:03
  Speaker 3: The universe is running out of gas, except the gas
• 19:06
  Speaker 3: is still there.
• 19:07
  Speaker 2: Surprise without needing a monster particle or a collapsing theory
• 19:11
  Speaker 2: of gravity, just baryons being more interesting than we allowed.
• 19:16
  Speaker 3: That is my favorite kind of result. No fireworks, a
• 19:20
  Speaker 3: door quietly reopening.
• 19:22
  Speaker 2: Potential impact is on simulations, on how we interpret upcoming
• 19:26
  Speaker 2: HI surveys, and on the story we tell about cosmic dusk.
• 19:32
  Speaker 3: SKA, the Square Kilometer Array, is going to drown us
• 19:36
  Speaker 3: in 21 centimeter data. This paper is a preview of
• 19:39
  Speaker 3: the questions that flood will have to answer.
• 19:42
  Speaker 2: If SKA finds the same weak evolution, the bottleneck hardens
• 19:46
  Speaker 2: into fact. If it finds something else, this becomes a
• 19:50
  Speaker 2: stepping stone that got us asking correctly.
• 19:53
  Speaker 3: Either outcome is healthy.
• 19:55
  Speaker 2: I also keep one eye on our own future observations
• 19:59
  Speaker 2: of molecular gas at these redshifts. That is the missing
• 20:02
  Speaker 2: twin measurement.
• 20:04
  Speaker 3: If molecular gas fell by a factor of 2.5 in
• 20:08
  Speaker 3: lockstep with stars, The story is almost complete.
• 20:12
  Speaker 2: If molecular gas stayed high too, then the bottleneck moves again,
• 20:16
  Speaker 2: into the collapse of clouds or the survival of young stars.
• 20:21
  Speaker 3: A nested mystery. Very on-brand for astronomy.
• 20:25
  Speaker 2: One more human note. Four and a half billion years ago,
• 20:28
  Speaker 2: Earth already existed. The moon was there. Life was prokaryotic
• 20:33
  Speaker 2: and stubborn.
• 20:34
  Speaker 3: And overhead, the universe was more than twice as busy
• 20:37
  Speaker 3: making stars as it is in our century.
• 20:41
  Speaker 2: We are historians of a quieter sky, using dishes in
• 20:45
  Speaker 2: karst valleys and robots on a mountain in Arizona to
• 20:48
  Speaker 2: reconstruct the dimming.
• 20:50
  Speaker 3: That image is enough. I do not need it dressed up.
• 20:53
  Speaker 2: Neither do I. The measurement is the drama.
• 20:57
  Speaker 3: A factor of two and a half against a factor
• 20:59
  Speaker 3: of barely more than one. Remember those two numbers.
• 21:03
  Speaker 2: And remember that hydrogen is still the most common atom
• 21:07
  Speaker 2: in the cosmos. The shortage was never elemental. It was logistical.
• 21:12
  Speaker 3: The universe has plenty of hydrogen. Star formation is collapsing anyway.
• 21:17
  Speaker 3: That is the story.
• 21:19
  Speaker 2: And the work of the next decade is to watch
• 21:21
  Speaker 2: the middleman, the conversion, until it confesses.
• 21:25
  Speaker 3: I can live with that assignment.
• 21:27
  Speaker 2: So can the field, I think. This is the kind
• 21:30
  Speaker 2: of paper that does not end a conversation. It gives
• 21:33
  Speaker 2: the conversation better manners.
• 21:35
  Speaker 3: Less hand-waving about empty tanks. More attention to how gas
• 21:39
  Speaker 3: actually changes phase.
• 21:41
  Speaker 2: If you take one feeling from this episode, let it
• 21:44
  Speaker 2: be that slight tilt of the world. The night did
• 21:47
  Speaker 2: not get poorer. It got pickier.
• 21:50
  Speaker 3: A picky universe, still holding its hydrogen, lighting fewer new suns.
• 21:56
  Speaker 2: We will be here when the next survey checks the ledger.
• 22:00
  Speaker 3: Until then, look up and remember the unused C.