Department Of Simple Pleasures
# 4582 In A Never-Ending Series
Dateline: Wednesday; 2pm-ish; picking up the week’s share of produce from The Farm ® (our CSA, La Finquita del Bujo [1] ). This is only the third pickup of the season, and the first one in which there has been bok choy. As I pick out my bok choy stalks I find moiself talking to them, an interaction which did not occur to me when I picked up the bundles of kale, lettuce, mustard greens, radishes, cilantro, shallots, and mushrooms:
Bok, bok bok bok bok bok bok bok…choy!
I don’t know how not to do this.
* * *
Department Of The Cautious Indoor Kitty Steps Out
We’re trying out a new harness/leash arrangement with our elderly cat, Nova – who officially is 18 years old, as of yesterday (!) – as part of our ongoing effort to make car transport easier for her…and, okay, also (read: mostly) for us. Nova goes to and from the coast with us on a regular basis; the aim is to have Nova in an open carrier (which is belted into the seat belt) in the back seat of the car, with her harness/leash device clipped to that carrier so that she’ll be able to roam from the carrier to her travel litterbox and back to the carrier – but with the least short enough so that she cannot get into the front seats with us and get underfoot/carjack us/otherwise get into trouble, in an effort to reduce “accidents” [2] she’s been having in her other carrier.
* * *
Department Of What Is Wrong With Me, That Even For A Moment,
I Wanted This
Stop wringing your wimple, sister, I didn’t even post what “this” is.
Ok; you can shame me now.
* * *
Department Of A New Thing To For My Brain To Wake Me Up
In The Middle Of The Night Wondering: “What About This?”
Moiself looks forward to the day when I read a headline about how, with the help of AI technologies that could understand and synthesize every available piece of research and information, medical researchers have found the causes of – and therefore the cures for – Parkinson’s, Alzheimer’s, ALS and other devastating, progressive afflictions.
Regular readers of my blog may know that I tend to be very critical of AI across the board, both the emerging technology itself and its applications. Moiself is holding hope (despite what I know of human tendencies and capacities), that for all of the downsides of AI – read: the counterfeiting of almost everything – there are scientists and engineers working on identity/detecting technologies, as in, solutions/warning/watermarks two warnings of the “bad” or counterfeit usage of AI.
A recent episode of the science podcast Curiosity Weekly had me wondering about another AI downside, mere seconds after the episode had me marveling over the same AI application. I left skidmarks going from, This is incredible!!, to Yikes – how will we recognize the counterfeit or be able to identify mistakes?!?
The last part of the episode Indigenous Climate Solutions, Virus Thaw, AI Brain Implant deals with a new AI-powered brain implant. This device aims to translate thought into speech for a person who survived a stroke which left her a quadriplegic who had also lost the power to speak.
“A new AI powered brain implant was able to translate thoughts to speech in an instant for patients with paralysis like the studies participant, and this is the closest to naturalistic speech, a communication device has come yet, a brain computer interface or BCI is a system that lets the brain talk directly to a computer. It works by detecting the electrical signals your brain naturally produces when you think, move, or even just imagine speaking, and then translates those signals into commands that a computer can understand….
‘Anne’ had a stroke in her brain stem, which left her quadriplegic and unable to speak even though she knows what she wants to say. The hope is that BCI can decode what she intends to say and turn those thoughts into real-time audible speech….
…currently, most patients with severe speech impairments rely on devices ranging from simple picture boards to computer programs that synthesize speech from text…. Scientists are still developing these brain computer interfaces, but a new system from researchers at UCSF and Berkeley…was able to simultaneously detect words and thought and turn them into speech within just three seconds….
They trained the deep learning model on data to interpret the patterns of neural activity. Because the system was designed to capture activity in tiny windows of time every 80 milliseconds. It didn’t have to wait for Anne to finish a whole sentence. It could start interpreting what she was trying to say while she was still saying it…
So far Anne has been the only patient to test the device, so there’s a long way to go before a device like this is available to the masses. But as a proof of concept, [3] it’s such a great start. And for an added touch per Anne’s request, they train the system on recordings from her wedding video. So the computer speech sounded like her voice from before her paralysis….”
(excerpts, transcript for Curiosity Weekly 4-23-25 )
YEAH WOW…BUT….
What if the AI device is seemingly working fine – as in, accurately translating Anne’s thoughts into speech, for days, weeks, months – then suddenly mistranslates something? What if the AI/BCI somehow gets it wrong, about what Anne was thinking, what she meant to imply, what she wanted the technology to say for her? What if, from that point on, her AI voice is wrong, every time…or maybe one out of three times, or…?
Whether it messes up one minor thought…
* Anne’s I’d like a lemonade is mistranslated as, Please bring me orange juice,
or mistranslates her attempt to convey a symptom…
* Anne’s I taste blood in my throat translates as, My scalp itches.
Instantaneous or verification of precision or inaccuracy of translation is vital.
Anne can hear what her BCI is saying, as “herself” (Anne). But because she is quadriplegic with no power of speech, she can’t say, immediately or otherwise, Whoa there, that’s not what I meant to say. How can she indicate to friends, family, coworkers, her medical attendants around her, that, No, the BCI/AI voice, which is uncannily mimicking (“speaking in“) her voice, got it wrong:
* She was not asking for an enema; she was asking about her Aunt Emma;
* She was not asking you to mark her ballot for the Republican candidate; she was thinking as strongly as she could that it will be a cold day in Death Valley before she will ever vote for a Republican again…
* Her BCI/AI voice says to Please help me redo my will, when she was thinking ,Will you please redo my rickety wheelchair ramp?
When – not, if – Anne’s BCI-AI translator misinterprets her brain waves, she will (hopefully) know that it has done so, [4] because she hears what other people think she is saying or requesting, via her BCI-AI. How would Anne then communicate this error in communication?
* * *
Department of Employee Of The Month
It’s that time, to bestow that prestigious award upon moiself. Again. The need for which I wrote about here. [5]
* * *
Freethinkers’ Thought Of The Week [6]
“Religion…asks us to deliberately deceive ourselves – to replace reason with its opposite, faith, and when men operate on faith, they can no longer be reasoned with, which makes them more dangerous than any sane man, good or evil.”
( James L. Sutter, American author & game designer )
* * *
May you make a sizeable donation to the ASPCA should you ever
purchase a classic car stroller for your pet;
May you never had the need for a brain implant to speak for you;
May you enjoy making chicken noises to your bok choy or other produce items
(try clucking to kohlrabi for a real change of pace!);
…and may the hijinks ensue.
Thanks for stopping by. Au Vendredi!
* * *
[1] “The Little Farm of the Owl”
[2] Read: barfing and/or peeing and/or pooping. Sometimes she goes for the trifecta.
[3] A proof of concept (POC) is a demonstration of a product in which work is focused on determining whether an idea can be turned into a reality. A POC’s goal is not to seek market demand for the concept or choose the best way to produce it. Rather than focusing on building or developing the idea, it tests whether the idea is feasible and viable. In addition, it enables those involved in the proof-of-concept exercise to explore its financial potential. From techtarget, definition, POC
[4] As long as her hearing remains intact…or her eyesight, if the technology allowed for printout of every thing the AI voice says for her.
[5] Several years ago, MH received a particularly glowing performance review from his workplace. As happy as I was for him when he shared the news, it left me with a certain melancholy I couldn’t quite peg. Until I did.
One of the many “things” about being a writer (or any occupation working freelance at/from home) is that although you avoid the petty bureaucratic policies, bungling bosses, mean girls’ and boys’ cliques, office politics and other irritations inherent in going to a workplace, you also lack the camaraderie and other social perks that come with being surrounded by your fellow homo sapiens. No one praises me for fixing the paper jam in the copy machine, or thanks me for staying late and helping the new guy with a special project, or otherwise says, Good on you, sister. Once I realized the source of the left-out feelings, I came up with a small way to lighten them.
[6] “free-think-er n. A person who forms opinions about religion on the basis of reason, independently of tradition, authority, or established belief. Freethinkers include atheists, agnostics and rationalists. No one can be a freethinker who demands conformity to a bible, creed, or messiah. To the freethinker, revelation and faith are invalid, and orthodoxy is no guarantee of truth.” Definition courtesy of the Freedom From Religion Foundation, ffrf.org
