Man's brain turned to glass by hot Vesuvius ash cloud | Hacker News Reader

Man's brain turned to glass by hot Vesuvius ash cloud

bbc.com

240 pointstartoran103 comments

> ... scientists have discovered that his brain was preserved when it turned to glass in an extremely hot cloud of ash.

> The pea-sized chunks of black glass were found inside the skull of the victim, aged about 20, who died when the volcano erupted in 79 AD near modern-day Naples.

> ... a cloud of ash as hot as 510C enveloped the brain ...

I don't think I'd use the word "preserved" to describe there being a few glassy cinders left over, after someone's brain was incinerated.

It was preserved in the sense a fossil is preserved. The object was destroyed but some of it's structure remains. (Which is ludicrously cool. The world is so strange and wonderful. Though it's also a world where a volcano can erupt in your vicinity.)

Also, it's a great lesson in randomness - that day when that mountain erupted, the wind was blowing away from Naples, essentially sparing the disaster for its residents. Imagine if it was the other way around?

Would wind direction have made the difference for pyroclastic flows? I thought flow direction was largely driven by local topology.

Not a volcanologist, and vastly simplified, but...

After the collapse of the eruption column, when it's roaring down the volcano's flanks and incinerating/burying everything in its path, it is driven by the local topography.

But before that collapse - especially toward the end, when it's getting seriously unstable - the wind can have huge effects on the direction in which it "fall down".

And the wind (at altitude) pretty much controls which way the ash fall goes. That's technically not pyroclastic flows - but often just as deadly for a major eruption.

EDIT: Add Wikipedia's account of the eruption - https://en.wikipedia.org/wiki/Eruption_of_Mount_Vesuvius_in_...

- And note that descriptive terms for volcanic eruptions (Plinean, Pelean, Vesuvian, etc.) are subject to "it's all constantly-changing shades of gray" caveats.

Thanks for the explanation.

The wind _positions_ the ash cloud and then the terrain channels it once it falls. I suppose I'd been thinking of a Mount St Helens style flow where the mountain gave way.

Lots of glass brains?

LPT: A volcano can only erupt in your vicinity if you're in the vicinity of a volcano.

I don't think that phrasing implying there was a whole brain instead of a part fares any better by tying it to a definition of fossil. It's not essential to the definition of fossil that what is recovered is a smaller subset of the whole thing. A fossil can be a fossil of an entire organism.

I think the that commenter's point is that the article makes numerous unqualified references to "the brain" which on a most natural reading would imply the whole brain when it's some (albeit very interesting) individual shards.

I would have said the same if I discovered those scrolls charred to a crisp too, but here we are reading them like it wasn't sci-fi.

So, can we start an X Prize to read the contents of this brain?

Decoding any part of his memories would be the greatest archeological coup in history, by a massive margin.

By some estimates, the cerebral cortex can store hundreds of terabytes of information. Recovering even 0.1% of that would amount to possibly hundreds of GB of information about life in a major Roman urban center.

Or hundreds of GB of “how to stand up without falling” or “what fish taste like” or “don’t grab thorny bushes”.

Or “how to contract muscle here to make lungs intake air”

I thought a lot of this stuff was basically below conscious thought. You can interrupt it but otherwise the urge to breathe and the corresponding breathing proceeds apace. I've been told that in ventilator-dependent patients if the ventilator has too low a breath rate they still have the urge to breathe and feel it quite painfully until the ventilator takes a breath for them.

If you pass out under water because you run out of oxygen you will immediately resume breathing and drown.

A key difference is that we did know how to read non-burnt scrolls before that.

In 1000 years perhaps we will understand how the mind works and scientists will take out this glass brain and read it.

What an incredible journey it would be if it were possible for a man’s glass brain to yield enough data to reconstruct his consciousness and awaken him in the far future! There must be some sci-fi story like this.

Almost every week there's a discussion on hacker news that leads to a great idea for a sci-fi story and this is one of them.

Any compilations of such ideas?

I have jotted some of them down over the years. But I bet searching the full catalog of hacker news comments would yield a bunch of entries.

Yes, can’t remember the name, but there’s an old short story about alien explorers who find a dead Earth with futuristic buildings.

They revive a knight who attacks them and gets killed, a hippie who they kill off after questioning, and finally a future human with psychic powers who steals the revival device.

A.E. Van Vogt wrote it. Brilliant author. I am sure one of the tales in "The Voyage of the Space Beagle" was the inspiration for "Alien".

1000 years? We already have brain to machine interfaces that seem able to interpret words with 50%ish accuracy and sometimes even beyond.

By no means perfect but it's surprisingly close. Granted I think that interpretation is from looking at stuff like electrical activation and not just brain structure itself and the fossil is the latter.

If only "read words" was the same as "read brain" eh?

In certain contexts it is in fact the same thing because one thing brains can do is think of words, and this brain activity can be interpreted by machine-to-brain interfaces.

https://www.vox.com/future-perfect/2023/5/4/23708162/neurote...

Well, you could probably ask an LLM today what knowledge his brain contained and get plausible sounding answers. ;)

Right, and we have to be careful not to misinterpret such declarations by LLMs as if they're the real thing. I will never forget the first wave of Chat GPT reporting from the likes of Kevin Roose, who took its conversations completely at face value, not realizing it was reacting to his cues, and frankly playing down to the quality of his input.

But bring to machine interfaces are an entirely different thing altogether and they test it with controls and training unknown examples and then proceeding to prediction based off of training.

It may be a bit anticlimactic when we discover that he was just thinking about food and sex

Roman Empire

I developed a theory a few years ago (probably during the pandemic when I would go on a lot long walks alone) that there are many human information artifacts that are preserved that we haven't yet discovered.

Take sound waves. When they come into contact with matter, most of the waves energy turns into heat, yes? But might a mechanical imprint be made onto a surface? For example tree sap that hadn't yet completely solidified. Could we resurrect the roar of a dinosaur?

That's a fun one! And the reality is that there almost certainly *is* information out there, it's just a function of us figuring out how to pull the signal out of the noise. (No pun intended :P)

Perfect pun!

This research project therefore entails spreading some gooey substance on the trees in a forest, playing some known voice recording or songs, letting the gooey substance harden, and then training the AI model to look for that signal.

You know someone's gonna do this someday.

It's not clear that enough information really survives:

https://en.wikipedia.org/wiki/Archaeoacoustics

Wow! I had no idea that the concept had been explored for over a hundred years.

>Take sound waves. When they come into contact with matter, most of the waves energy turns into heat, yes? But might a mechanical imprint be made onto a surface? For example tree sap that hadn't yet completely solidified. Could we resurrect the roar of a dinosaur?

I think it's a fascinating idea and something like that instinct has a "there" there. I suspect that, as another commenter is noting, you probably lose too much information in most cases, but it's intriguing enough to merit consideration.

I think I'm sympathetic because I have a version of that too - I'm curious about the idea that the earth's light, sent out into space, may in some instances be recoverable due to spacetime curvature sending old light back to us. Imagine seeing a stream of light livestreaming what the earth looked like millions of years ago! But my understanding is that light spreads out more and more the further it goes so it's not likely to hold together as a decipherable image, but I still wonder about a middle ground of what might be partially recoverable.

This reminds me of the short story Time Shards:

https://www.lightspeedmagazine.com/fiction/time-shards/

What a good read, thank you.

Many parts of the charred scrolls are readable...but those are in vastly better physical shape (comparing "before" & "after") than the few pea-sized cinders remaining of this brain.

Plus - information storage in the brain is both distributed, and micro-scale.

At some point, you are trying to recover compressed data from a HD where only one or two bits can be still read from each 512-byte sector - basic information theory says that the only good-enough tech to do the job would be a time machine.

To be fair, I'd wager the contents of said brain were roughly 'aaargh! I don't want to die! Jupiter save me!', but in Latin.

"Aaargh! Nolo mori! Iuppiter, serva me!"

4 years of Latin in school finally came in useful!

> The scientists believe the skull gave some protection to the brain.

I don't think we're on the same page about the word protection either.

It did provide some protection, as it wasn't burned like the rest of the body. They're not saying that it provided perfect protection, only some protection.

It would lend to some very FAA-style euphemisms. Civilization-Scale Cold Storage Archival Backup == nuclear winter

There some life forms on earth that will never become preserved. Squids for example are a high ammonium based life form. The Ammonium prevents fossilization. If humans didn't exist while they do, we would have never known about them. [0] There would have been no preserved record.

[0] https://www.barnesandnoble.com/w/squid-empire-danna-staaf/11...

Being entirely naive to squid prehistory, the wikipedia page for squid shows two apparent-to-me squid fossils [0]. Is there some nuance here that I'm not aware of?

[0] https://en.wikipedia.org/wiki/Squid#Evolution

Not sure about squids in particular, but the easiest way isn't necessarily to find remains, but to see negative prints on sediments.

Tons of fossils are "negative prints on sediments". GP's assertion that "If humans didn't exist while they do, we would have never known about them" is clearly wrong.

Is there a marine equivalent of amber that things can be trapped in and preserved?

Apart from ice I guess.

https://www.amazon.com/Anatomy-Lab-Preserved-Specimen-Dissec...

From the study itself:

> Moreover, exceptionally well-preserved complex networks of neurons, axons, and other neural structures have been revealed by Scanning Electron Microscope (SEM) investigation of the brain remains and those of the spinal cord

https://www.nature.com/articles/s41598-025-88894-5

Supposedly this vitrefied brain is mostly carbon and oxygen.

Interestingly, there is a glass-like form of carbon (just carbon):

https://en.wikipedia.org/wiki/Glassy_carbon

Thanks, this was the sentence that was missing from the article and made me confused knowing that humans are basically made of carbon, but glass is not.

The article should have said “a glass”, not “glass”. The former is a term from physics that is different from the lay-mans’s “glass”. https://en.wikipedia.org/wiki/Glass_transition:

“The glass–liquid transition, or glass transition, is the gradual and reversible transition in amorphous materials (or in amorphous regions within semicrystalline materials) from a hard and relatively brittle "glassy" state into a viscous or rubbery state as the temperature is increased. An amorphous solid that exhibits a glass transition is called a glass.”

The Nature article is clearer. https://www.nature.com/articles/s41598-025-88894-5: “Glass forms when a liquid is fast cooled preventing crystallization, across a reversible process known as the glass transition.

[…]

Here we demonstrate that material with glassy appearance found within the skull of a seemingly male human body entombed within the hot pyroclastic flow deposits of the 79 CE Vesuvius eruption formed by a unique process of vitrification of his brain at very high temperature”

The layman’s term includes such things as safety glass, which may have polymer layers.

So, confusingly, not all glass is “a glass”, and not all glasses are glass.

> humans are basically made of carbon, but glass is not

A glass is something that underwent a glass transition (that looks like a liquid at the atomic scale but behaves like a solid microscopically, resulting from cooling a liquid too fast to let it crystallise). It can be made of a huge diversity of things: pure elements (like carbon or sulphur), some metallic alloys, oxides, sulphides, fluorides, polymers, etc.

But the article says "glass", not "a glass".

Using “glass” as a strict equivalent for “silicate glass” is incorrect. Metallic glass is glass. Not the glass you are used to, but still glass.

I'm not convinced that that is common usage. Maybe in technical circles.

However, I've not been able to find much on carbon-oxgen based glass. It's possible to make glass out of CO2 gas, under high pressure. However, at standard pressure, the glass boils off into CO2.

There are definitely some unconnected dots in the story. I have a sense that what is needed is to reproduce this allegedly vitrefied organic material in the lab.

Could this actually be more like a plastic? Some thermoplastics share characteristics with the category of glass, like having amorphous structure and a gradual softening resembling glass transition temperature. Conversely, we could say that glass, such as a common silica glass, is a kind of thermoplastic.

In:

https://en.wikipedia.org/wiki/Thermoplastic

we have language like "[a]bove its glass transition temperature and below its melting point, the physical properties of a thermoplastic change drastically without an associated phase change."

Also:

https://en.wikipedia.org/wiki/Glass#Polymers

"Many polymer thermoplastics familiar to everyday use are glasses."

https://en.wikipedia.org/wiki/Glass#Non-silicate_glasses

"Besides common silica-based glasses many other inorganic and organic materials may also form glasses, including [...] nitrates, carbonates, plastics, acrylic, and many other substances."

i wonder if this guy ever imagined he would be on hacker news

Well if he didn't back then - he did now.

He'll know after his avatar is a resurrected from his preserved brain.

The true Roko's Basilisk: being subjected to social media scrutiny thousands of years after one's death.

That would be Ea-Nasir

there's nothing to wonder. He never would have imagined that, it is entirely impossible.

you guys really didn't get the joke did you

Are you disparaging early Roman hackers?

Maybe not, I think hacker news didn't exist at the time.

There was a Tiktok trend where (typically) women were shocked to discover the (typically) men in their lives thought about the Roman Empire as much as they do, often on a daily basis. The question is why.

To me, it comes down to Rome not being the oldest or even the necessarily the largest or longest-surviving empire. It's that it's the most well-documented ancient civilization. Sites like Pompeii and Heculaneum provide a time capsule into ordinary existence that is often missing from ancient accounts that typically talk about kings, emperors, wars and so forth. In addition, we have a ton of texts from that time, including the direct writings of the likes of Julius Caesar.

Rome continued to influence European history beyond the fall of Constantinopole up until the 19th century through the Holy Roman Empire.

But the impact is still felt today. Classics such as Marcus Aurelius have arguably been co-opted into the alt-right pipeline.

There's also interesting psychology at play here. People like to imagine themselves in such a world. Where in the real world they might be just an average working Joe, people rarely imagine themselves as being peasants or slaves or a grunt in the army despite those being the majority of people.

I find that last point needs highlighting because there is an effort to reshape our current society, driven by real yet misplaced legitimate anger. Human ego being what it is, nobody acknolwedges the statistical likelihood that if you're suffering or oppressed in the current organization of society, you're probably going to be oppressed or otherwise suffer in a new society, particularly one built around an autocrat.

But when the central organizing principle becomes cruelty, perhaps aspiring to being a Brownshirt is the goal.

When the only perceived means of winning is making others lose, most people are going to lose.

The US should never have used plurality voting. It functions as the inputs to the Nash Equilibria decision matrix, our individual votes being against a perceived evil rather than for a value which supports civilization.

If instead of {+1, 0, 0, 0...} we used {+1, +0.5, -0.5, 0, 0, 0...} with each non-zero value used at most once and without duplication of candidate, we would be able to vote for the outputs of the decision matrix--our combined decision--and avoid the tragedy of the commons. I believe the coordination problem is the Great Filter, and going interplanetary won't solve the underlying math of shooting first being incentivized by winner-take-all, and the risk of mutually assured destruction.

The Partial Vote system as I call it would still be one voter one vote, it would just be easier to express it in separate components rather than listing all permutations.

Edit: Also, try applying ranked choice to a nash equilibrium matrix. There are some pathological cases to using rankings for a single-seat (result) selection process, where a voter might have had a better result for them if they hadn't voted. That can't happen with the partial votes described above.

Too complicated. Americans don't even understand the much simpler and (IMHO) sufficient ranked choice voting [1]. Alaska, a deep red state, sent a Democrat to Congress because the voters split their votes between 2 Republican candidates because they didn't rank both candidates.

While I think RCV would be better, I still don't think it solves the problem. There are a bunch of ways in which our system is designed to create a two party system, such as what constituionally happens if no candidate gets a majority of votes in the electoral college [2].

That aside, look at other countries. Has more than two parties really helped in practice? Germany, the UK, Israel and France all have 3+ parties in their house of representatives equivalent and all have swung to the right.

Practically speaking, we could solve a bunch of our problems by simply repealing the Permanent Apportionment Act of 1929 [3], which set the number of House members to 435 and a district size of 700k+. This would take a simple majority in the House and Senate and would revert district sizes back to 30,000. This would kill gerrymandering, practically speaking.

[1]: https://thehill.com/blogs/blog-briefing-room/news/3624553-re...

[2]: https://www.usa.gov/electoral-college

[3]:https://en.wikipedia.org/wiki/Reapportionment_Act_of_1929

Ranked choice for single-seat elections can create situations where your ballot backfires, which is why it has been tried and rolled back. It works for proportional representation, but then you've got people divided ideologically rather than by region and their local communities.

The divide by ideology (proportional), or into "safe" one-party states and "battleground" states (plurality in the US) is the biggest issue, the two parts of the human experience losing touch with why the contrasting values exist in the first place.

That said, good point on the issue of the size limitation on the House.

Part of the reason is that the Rennaisance scholars had high regard for the Romans, and that passed along to the modern age. Much that was created since then, for example some Shakespeare plays and English rhetoric, was influenced by the Romans.

> Classics such as Marcus Aurelius have arguably been co-opted into the alt-right pipeline.

What else from Rome (or Greece) is used this way? Is this related to why some SV leaders embrace the Stoics?

Are the neurons preserved in a recoverable way, like 5D optical glass-based data-storage? Also Chinese (and probably other languages) has a common derogatory expression "Glass heart" by which they mean "I think you are too sensitive to what I see as valid criticism of you" - this glassy brain preempted that slur at an intellectual level.

It's "preserved" in the same way an atom bomb "preserves" your shadow if you're stood at ground zero.

Think more "something remained" than "some thing remained"

Hehehe - Nice distinction!

Once fell down, from volcanic gas.

Soon turned out, I had a brain of glass.

dammit, now I have that tune stuck in my head.

thanks

Reminds me of the brain of a man that went down with the Swedish warship Vasa that was found preserved (turned into soap.)

Someone should have shown that to the "Rings of Power" screenwriters, because in that show Galadriel comes out of the volcanic ash cloud with just her hair messed up :-D

I learned a while ago that more than half the population of Pompeii and Herculaneum fled and survived the destruction of their homes. And the ensuing refugee problem lasted a while in nearby settlements, including Napoli.

So it's likely that the ones who remained and died were either stubborn or disadvantaged somehow, as well as those just unlucky.

Nat Friedman circa 2048: we were successfully able to extract the man’s last thoughts (“oh shit, my boss is going to be so mad when he hea”)

Oh yeah, I'm going to end up on hacker n..

Nature article: https://news.ycombinator.com/item?id=43213563

And Ars Technica: https://arstechnica.com/science/2025/02/study-hot-vesuvian-a...

Crystallising his memory and mental processes

there must be millions of glass dinosaur brains deeper than their already mustered bone fragments

Why would they be deeper, if you mean 'deeper under the surface of the earth'?

because none have been found at the depths of bones..

Sounds awesome, where do I sign up?

Read this article while eating breakfast, hoping to be able to continue enjoying breakfast. Failed.

On the bright side, you didn't have a volcano erupt on you.

Bright side APPRECIATED! TY.

Now, this is a fossil I would like to have!

... Oh, in the Pompeii eruption, not, like, breaking news. Given the general surreality of 2025 news thus far, would not be surprised.

To be fair, it is a very active volcano.

Now he's definitely smoothbrained.

Is it why in Stargate advanced alien computer/memory is from a crystal?

Nope, that's because glowy crystalline or glassy things look super fancy and futuristically high tech.

There was also a certain amount of hype about 3d optical data storage at the time. In practice, this is at this point more or less a dead issue, but it was very much The Future in the 90s. Also shows up in Star Trek TNG and on.

Why is it dead

Flash and streaming.

Flash because it's cheap enough, high density, arbritarily scalable for whatever storage is needed.

Streaming because the difficulty making any optical rewritable (especially with respect to write speed) meant the niche overlapped with bluray and DVD, but why bother with yet another optical disk for films and games when bandwidth becomes more important than single-disk capacity.

There's also a bunch of technical difficulties, but when there's no market for the result, R&D efforts are often less well funded: https://en.wikipedia.org/wiki/3D_optical_data_storage

Some combo of flash and spinning rust storage really makes the market for very high capacity optical storage a bit niche. It might still have a place as archival media (and multi-layer optical disks, which are 3d optical storage of a sort, though not to the level of sophistication imagined in the 90s, are used in that role to some extent), but it's just really hard to see it rivalling flash for mainstream applications at this point. In particular, flash is really, really, _really_ fast.

And make for nice props that show up well on a 90s CRT, and they fit the Stargate aesthetic that lifted many 60s hippy occultist ideas including both "pyramids are too advanced, space aliens must have built them"[0] and "crystal skull has psychic powers"[1].

[0] https://en.wikipedia.org/wiki/Chariots_of_the_Gods%3F (hmm, even the book name sounds like two SG-1 episode titles, pilot and when thrmey meet Thor)

[1] https://en.wikipedia.org/wiki/Crystal_skull

Reply on news.ycombinator.com