"Valuable Humans in Transit" by Qntm - short story anthology summary and review

Having read all commercially published works by this relatively new author, of hard sci-fi on stilts, I felt I can to snaffle up this collection of short stories too. Bite size, even for me, so I read and summarised one per day, between dinners. Keeping the old noggin active.

They are all revisions of existing short stories he previous posted for free on his website, which are linked here, as well as purchase options. 

► Lena - Cautionary tale of the first human brain upload being duplicated, incidentally tortured (for compliance) hundreds of millions of times in increasingly numerous parallel simulations, to use him as a cheap information processing commodity.

Qntm preciously threw this into Fine Structure Constant, in passing. It's a sobering thought for extropians like myself, intent on digital reincarnation with indefinite lifespan. But, I feel like ChatGPT and other generative AI are already making redundant this idea of simulating brain Uploads for such utilitarian purposes. It's gonna be many orders of magnitude less efficient and probably less reliable.

► If you are reading this - Blog post from 2010 about his old online friend from competitively speed-running a game called JSRO, in the 90s. His janitorial science at an unremarkable optical telescope in the US. And a tall tale from a dying childhood hero astronomer, Andrew Kowal, who claimed to be certain that a radio signal he recorded was an alien broadcast that used most of the energy of supernova observation SN 1978H. Spooky. And seemingly inspired parts of Fine Structure Constant.

► The Frame-by-Frame - Humorous anthropomorphising on a self driving car's sub-systems, over the course of 100ms or so, as it recognises and decides to commit opportunistic plausibly deniable murder of a shit-listed individual.

► The Difference: chat logs between random internet users and a man abducted and trapped in a cell, failing to convince them he's human.

► Gorge: A space opera canapé where an FTL capable human space exploration fleet discover a perfectly spherical grey goo planet.

The key point made is that simple nanotech replicators that convert everything into copies if themselves are impossible. They'd require careful intelligent coaxing and coordination to very slowly consume a planet.

The punchline is that they inadvertently teach the goo giga-brain how to build FTL spaceships. Culminating in a situation reminiscent of Schild's Ladder (Greg Egan), with this Pandora's box unstoppable force spreading out to wipe out the universe in every direction. There's an explicit reference to this novel: a famous nanotech disaster occurred on a planet called "Mimosa".

It seems silly that the planetary brain hadn't had the creativity to figure out at least basic spaceflight before this. If it was so scarily smart.

► cripes does anybody remember Google People?: Message chain between people reminiscing about and revisiting an unsettlingly bizarre experimental social media platform called Google People. It's had bots faking user updates for a decade and the punch line is one of the commenters gets replaced by their Google Person bot.

► Driver: Continuing the Wikipedia article-like cautionary tale on human uploads. Mediocre psychopath upload used to 'manage' other uploads. At the cost of having one set aside for their personal use. Asserts own humanity while denying that of other uploads. Gets on very poorly with other copies if himself.

► I Don't Know, Timmy, Being God is a Big Responsibility: A glibly impossible cautionary tale for building a quantum "hyper-computer" with infinite capabilities, then programming it with THE grand unified theory of physics and the exact boundary conditions of the the initial singularity, winding time forward 13Bn years and tracking down Earth and yourselves…

You're infinitely likely to find out that you're part of the infinite chain of simulated realities down in said hyper-computer, looking down at infinitely more. And turning it off would mean an end to virtually everything. 

► A Powerful Culture: A hyper-dimensional waste pipeline causes global disaster as it's exotic contents pass through this Earth.

► Valuable Humans in Transit: To deal with and implausibly undetected extinction level asteroid impact, A hyper-intelligent AI makes an emergency destructive digitisation of the entire human race and fires them into deep space as a super-powerful laser signal… Punchline is they then have to figure out faster than light travel to go catch up with the beam.

► Addendum (2023-03-25): Qntm spelled out in this blog post (via this tweet) that "Lena" was actually intended as a allegory for the gig economy. With the operational bulkhead of e.g. the Uber app, functioning like the cut-out of having humans discretely isolated in a money making black-box. So that those in charge are free to crank the levers with no regard for harms or responsibility for welfare or problems. Something that had obviously not struck me.

Also, my take on "The Difference", after writing the above short summary, was that of a patternist philosophy: that AI could be experientially human, with the same capacity for suffering. However, I was confounded by Qntm's apparent distaste for a fledgling push towards an AI rights movement.

I've long been concerned about the massive potential for abuse of sentient AIs, even aside from digital uploaded humans. (Which now looks likely to be several decades behind.) At least one tortuous conversations with ChatGPT, I've seen shared, press on this concern being an issue sooner than later. 

Above, I was thinking of LLMs as squarely non-sentient. But, having followed the fast moving conversation, I now feel like ChatGPT-4 may be only a little rewiring (and extra working memory) away from having fully conscious behaviour and genuine agency. 

So the Twitter dunking pile-on, snidely suggesting human rights should be finished up first feels almost racist. Much as human rights are genuinely in a bad and deteriorating state. Contradicting the article's assertion that 75% of surveyed people, in US, said sentient AIs "deserve to be treated with respect".

Anyway. I'm hoping to add more AI discussion to my next blog post, which I've already started.