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.

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".

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.

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?

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.

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.

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!