Sarah Kendzior is one of my favorite writers. All of this latest issue of her newsletter is worth your time. And admiration. đ
Sample quotes:
"Like many technologies, AI is designed to kill humans, but in the most soulless way possible, stripping away even our ability to own and express our pain."
"A song written by AI isn't a song any more than dust is skin."
"AI wins not when it becomes more human, but when humans become more like AI: disengaged and devoid of original thought."
https://sarahkendzior.substack.com/p/soul-stripping
Tagged: Akkoma
@kolombiken Bra frĂ„ga. Kanske vissa tester mĂ„ste göras om i en del fall. Jag Ă€r osĂ€ker pĂ„ hur journalinfo ârĂ€ttasâ.
Men ja, för mÄnga mÀnniskor blir det övermÀktigt att ifrÄgasÀtta vÄrden. Allvarliga fel kommer hÀnda.
@ambivalena Ett grundlÀggande problem Àr, som jag ser det, att tekniken alltid kommer utifrÄn. Om det fanns it-verktyg som personalen direkt kunde börja anvÀnda utifrÄn egna förutsÀttningar pÄ ett sÀtt som passar in i deras vardag, alltid pÄ deras egna villkor, sÄ vore det en annan sak. Men det förunderliga jag tar med mig efter 30 Är inom it Àr att ingen litar pÄ att mÀnniskorna som gör jobbet har bÀst koll pÄ hur man gör jobbet.
@ambivalena Ja, att det skulle bli sÄ hÀr illa med digitaliseringen Àr jÀttesorgligt.
Vid mitt senaste vĂ„rdbesök frĂ„gade lĂ€karen om han fick spela in. âJaha, ni har ett program för att transkribera?â frĂ„gade jag. âJa, precis, kĂ€nner du till det?â
SĂ„ förklarade jag att jag jobbar med it och etik, och vet att det ofta blir fel i de dĂ€r transkriberingarna. Han brast ut i skratt⊠âhaha, ja, verkligen!â
Inte superförtroendeingivande.
VÀnner, begÀr ut en kopia av er journal och sammanfatta den sjÀlv manuellt och koncist. Ha med dig en utskrift av denna lÀsbara, tydliga journal vid era vÄrdbesök.
VÄrdpersonalen orkar inte lÀsa allt skrÀp som hamnar i digitala journaler.
https://www.svt.se/nyheter/inrikes/vardpersonal-fastnar-i-it-istallet-for-att-varda-patienter
Most "AI" consultants are in fact autocomplete consultants. It's like when a window washer's business card says "Transparency Enhancement Facilitator".
âI have FOBI. Fear of being included.â
â #Wednesday Addams
Finally decided on a name for my newsletter and set up a landing page for it.
It's still the same newsletter (or newsletters) I've always had, just with a name that resonates with me. :)
My Next Heartbeat
https://heartbeat.email
We really need to stop calling everything AI. It's misleading at best, delusion-inducing at worst.
If you always have all your digital stuff set to English you may not be aware of how often those of us who use multi-lingual interfaces come across bad auto-translation that force us to do a double-take. (worst case they can of course be very dangerous).
This is from an email log in Gmail today.
The label "Delstat" is presumable from "State" in English, meaning what condition/status the email is in.
But delstat in Swedish essentially refers to U.S. State. Which makes this information appear very odd đ
«This can be immensely empowering, and can improve public trust: itâs hard (and unwise) to give yourself over to a technology that wonât tell you how it works, particularly when its predefined settings allow only for meagre approaches to âuser privacyâ. As I ask my students, if you could develop an AI at your own home, and programme it to reflect your values and prioritise your safety, wouldnât you trust it more? Well, the idea isnât so outlandish â it only feels impossible because big tech firms want us to think it is.»
https://www.theguardian.com/commentisfree/2025/aug/02/billionaire-big-tech-frugal-elon-musk-innovation
@julian This is helpful. Thanks.
@julian You were right, it was a bot protection mechanism at Cloudflare. When I turned that off it started working. Oddly, the same system was turned on for the other domain which was working - meaning the LinkedIn bot was behaving "suspiciously" towards only one of my sites(!) Live and learn. Thanks again.
Open question about OpenGraph to any geek who feels like responding đ
BlueSky and LinkedIn fail to fetch the featured image for posts on axbom.com (my English site) but does it with no issues for posts on axbom.se (my Swedish site).
From my perspective they are set up in exactly the same way (Ghost sites). My estimate is that this started happening more than a month ago and I have no idea why.
So, what could be the issue?
LinkedIn, plattformen för professionellt nÀtverkande, uppdaterade för nÄgra dagar sedan sin policy gÀllande hatiskt och krÀnkande innehÄll. Specifikt har man tagit bort formuleringar som inneburit ett tydligt stÀllningstagande mot trakasserier riktade mot transpersoner.
LÀs hela inlÀgget: https://axbom.se/linkedin-raderar-tydligt-stod-for-transpersoner/
This is of course very bad, but I also found it funny. AI evangelists are suddenly privacy-aware and repeating this as a security risk. And it is. But the idea that using ChatGPT itself hasnât always been a security risk is ridiculous to me.
TLDR: If youâve ever used the share function on a ChatGPT chat, that full chat can be found via Google, viewable for anyone in the world.
https://www.fastcompany.com/91376687/google-indexing-chatgpt-conversations
If people are wildly optimistic and non-critical about deploying something that is destructive, that should never be framed as a positive.
And continuing on this, if I am being critical of optimism in this scenario, thatâs not âslowing progressâ or âbeing backwardsâ. Thatâs being cognizant of bad outcomes and resisting them.
This all reminds me of when people are understandably confused by ânegativeâ results in healthcare, when they can be the most welcome of news.
Reading McComasâ article from 1996, and I have to say itâs very refreshing.
TEN MYTHS OF SCIENCE: REEXAMINING WHAT WE THINK WE KNOW...
Some examples:
* Myth 3: A General and Universal Scientific Method Exists
* Myth 5: Science and its Methods Provide Absolute Proof
* Myth 6: Science Is Procedural More Than Creative
* Myth 10: All Work in Science is Reviewed to Keep the Process Honest
https://homepages.bluffton.edu/~bergerd/TenMyths.html
Something to always be aware of: Many wheelchair users can stand and move around for brief periods of time. Not all wheelchair users are paralysed. Reasons for wheelchair use are numerous and varied.
Some wheelchair users choose not to stand in public because chances are they will be chastised and harassed if they do. With more awareness and understanding this risk can hopefully diminish over time.
For example, if a wheelchair user is able to retrieve their own wheelchair from the boot/trunk of their car, this does not mean they are âfakingâ, and accusing them of this is ableist in itself.
âAmbulatory wheelchair userâ is something you can search for to learn more, as many are sharing their experiences online, like Lauren:
https://www.rareyouthrevolution.com/post/ten-things-i-wish-you-knew-from-an-ambulatory-wheelchair-user
#accessibility #wheelchair #ambulatory
LĂ€ser om algblomning och hur farligt det Ă€r för hundar. Att det pĂ„verkar hundar sĂ„ hĂ€r snabbt och Ă€r sĂ„ hĂ€r allvarligt hade jag inte koll pĂ„. SuperlĂ€skigt... đł
https://firstvet.com/sv/artiklar/algfoergiftning-hos-hundar