1.
ASMR is that full-body tingling some people get when listening to sounds like gentle tapping or whispering.
It is detectable in the brain using EEG spectral analysis: ASMR amplifies low frequency and reduces high frequency oscillations (2022).
(Technically: self-reported ASMR correlates with a shared EEG biomarker, but it doesn’t demonstrate the reverse – i.e. this does not yet show how to build an ASMR detector.)
Hey, Apple has a patent on EEG sensors in AirPods (as previously discussed).
Free product concept:
AirPods that use their sensors to detect the ASMR response, then dial in to relentlessly trigger it 24/7. A new alternative to noise cancelling mode.
Tim Apple you can put the cheque in the post.
SEE ALSO:
The ASMR version of Bee Movie, and other neuro-divergent media (2021).
2.
A paper presented at the 2018 Annual Meeting of the Society for the History of Technology (SHOT), St. Louis, MO:
The Early History of Artificial Intelligence in China (1950s-1980s).
Always interesting how words and politics amplify/inhibit technologies and ways of thinking.
e.g. in this paper, a quote from the 1970s:
“Can human build intelligence? No. The term ‘artificial intelligence’ is easy for idealists to exploit. If humans can build ‘intelligence,’ then, in the future, something more intelligent than humans will emerge. … Some Soviet revisionist academicians are vigorously propagating AI. This behavior fully exposes their ugly face as traitors.”
HOWEVER,
I am most particularly tickled by this anecdote:
In the early 1980s, the misunderstanding about AI lead many people to an idea that associates AI with the so-called “exceptional human body functions”, which refers to some unscientific, unproven, imaginary, supernatural functions of human body such as extrasensory perception (ESP), telepathy, and moving things with mind. This fervor stemmed from a news report in March 1979 on Sichuan Daily, a local newspaper in Sichuan Province. The report was about a boy named Tang Yu who were claimed to be able to read books using his ears. The news caused a sensation.
Huge if true.
3.
Every so often I remember this story about Jacques Derrida related on X nee Twitter and can’t stop laughing. Copy-and-pasting it here for posterity.
A professor of mine went to go hear Derrida speak once. The entire talk was about cows; everyone was flummoxed but listened carefully, and took notes about…cows. There was a short break, and when Derrida came back, he was like, “I’m told it is pronounced ‘chaos.’“
4.
The argument in From Newspeak to Cyberspeak (Slava Gerovitch, 2004) about why Soviet Cybernetics didn’t invent the personal computer, as previously mentioned:
the political insistence on materialism stripped cybernetics (and therefore computing research) of metaphorical yet inspirational ideas like “memory” and “learning”, constraining the vision of computing to simple calculation.
The human elements were translated away. So the metaphorical leap to “hey maybe computers could be bicycles for the mind” was just too far.
You know memory was controversial in English too, once upon a time? Like, computer memory.
May I suggest that a protest is overdue against the anthropomorphic character of the terminology that is being applied to electronic digital computers? To the layman the use of such a terminology seems to ascribe human or near-human properties to these machines.
That’s a letter to Nature.
I can’t read the full letter right now (please do send me the PDF if you can) but Kapp objected to the terms: memory, instruction, information, machine language and decision.
The argument settled quickly: memory stayed as the adopted term.
And then a backwash effect: people now often imagine human memory as working like computer memory, a box that you put things into and take things out of, which isn’t the case at all. The metaphor became the exemplar.
So I think about that, while we’re talking about “training” and “hallucinating” and “thinking” with our LLMs. It’s not true, anthropomorphising an AI as a lil guy is so misleading: you type something into a window, a rack of GPUs in West Virginia is lit for a nanosecond with the energy density of a small star, a wavefront of computation sweeps across electron surfaces that holographically encode the entire record of human culture at the speed of light, focusing into a single beam that spits a stream of tokens down a fibre onto your screen. And then it does the same for someone else.
Let’s use those human terms anyhow. It’s not like we have anything better. And they might help with with the next metaphorical leap to what these AIs really are.
(And then I wonder what the backwash is, how it will change how we narrativise ourselves.)
1.
ASMR is that full-body tingling some people get when listening to sounds like gentle tapping or whispering.
It is detectable in the brain using EEG spectral analysis: ASMR amplifies low frequency and reduces high frequency oscillations (2022).
(Technically: self-reported ASMR correlates with a shared EEG biomarker, but it doesn’t demonstrate the reverse – i.e. this does not yet show how to build an ASMR detector.)
Hey, Apple has a patent on EEG sensors in AirPods (as previously discussed).
Free product concept:
AirPods that use their sensors to detect the ASMR response, then dial in to relentlessly trigger it 24/7. A new alternative to noise cancelling mode.
Tim Apple you can put the cheque in the post.
SEE ALSO:
The ASMR version of Bee Movie, and other neuro-divergent media (2021).
2.
A paper presented at the 2018 Annual Meeting of the Society for the History of Technology (SHOT), St. Louis, MO:
The Early History of Artificial Intelligence in China (1950s-1980s).
Always interesting how words and politics amplify/inhibit technologies and ways of thinking.
e.g. in this paper, a quote from the 1970s:
HOWEVER,
I am most particularly tickled by this anecdote:
Huge if true.
3.
Every so often I remember this story about Jacques Derrida related on X nee Twitter and can’t stop laughing. Copy-and-pasting it here for posterity.
4.
The argument in From Newspeak to Cyberspeak (Slava Gerovitch, 2004) about why Soviet Cybernetics didn’t invent the personal computer, as previously mentioned:
The human elements were translated away. So the metaphorical leap to “hey maybe computers could be bicycles for the mind” was just too far.
You know memory was controversial in English too, once upon a time? Like, computer memory.
That’s a letter to Nature.
I can’t read the full letter right now (please do send me the PDF if you can) but Kapp objected to the terms: memory, instruction, information, machine language and decision.
The argument settled quickly: memory stayed as the adopted term.
And then a backwash effect: people now often imagine human memory as working like computer memory, a box that you put things into and take things out of, which isn’t the case at all. The metaphor became the exemplar.
So I think about that, while we’re talking about “training” and “hallucinating” and “thinking” with our LLMs. It’s not true, anthropomorphising an AI as a lil guy is so misleading: you type something into a window, a rack of GPUs in West Virginia is lit for a nanosecond with the energy density of a small star, a wavefront of computation sweeps across electron surfaces that holographically encode the entire record of human culture at the speed of light, focusing into a single beam that spits a stream of tokens down a fibre onto your screen. And then it does the same for someone else.
Let’s use those human terms anyhow. It’s not like we have anything better. And they might help with with the next metaphorical leap to what these AIs really are.
(And then I wonder what the backwash is, how it will change how we narrativise ourselves.)