eMusings
Your eyes and ears on the worlds of art, culture, technology, philosophy - whatever stimulates the mind and excites the imagination. We remind you that 20 years of
back issues of eMusings can be found on our archives page.
AI, for better or for worse, is now unstoppable. Try to use caution and common sense. If you are using it for business, edit
frequently. Here are some of the more thoughtful comments:
Let's begin with one of the most important articles I have read.
Researchers from the German Cancer Research
Center have developed a large language model (LLM) that predicts a person's risk of getting more than 1,000 diseases decades
in advance. Called the Delphi-2M, it was trained on over 400,000 medical records from the UK Biobank, a huge database that tracks participants'
health as they age. Additional data taught to the algorithm included lifestyle information like smoking and drinking habits and
body mass. Using this data, the AI
was able to predict the health trajectories of roughly 2 million people in Denmark. Delphi also explained the rationale for its
assessment.
Open AI has made no secret of its attempts to stop its
algorithms from deceiving users. It has discovered, instead, that it is teaching the models how to hide what they are doing while continuing
to deceive. In other words, their anti-scheming methods resulted in even more malicious behavior, as the algorithms
fabricated new rules to hide non-compliance. 'Don't hack' turned into 'let's cheat', a phenomenon known as pathological
lying.
A new book takes AI companies to
task for not understanding the extreme risks they are supporting. Titled, "If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All",
the book's main thesis is that we are heading toward "global catastrophe" unless we have effective guardrails. At issue are the
superintelligent algorithms expected to emerge in 2 to 3 years. Technicians will be unable to correct toxic behavior by
simply correcting code. Basically, the authors, who are both AI researchers, say that you cannot win a fight against opponents
that are smarter than you are.
A startup in the
U.K. has outperformed humans in an international forecasting competition. ManticAI, co-founded by a former Google DeepMind
engineer, ranked in the top 10 competitors charged with predicting the likelihood of 60 events over the summer. By comparison,
a year ago the best model ranked roughly 300th. The competition rewards not just regurgitated learning but genuine reasoning.
An AI algorithm detected signs of life in a comatose
patient days before doctors did. A new study was able to see small movements that were missed by physicians doing routine
exams. Nicknamed SeeMe, the model found evidence of covert consciousness roughly 4 days sooner and in 90% of patients.
Current brain imaging techniques are expensive and impractical. SeeeMe was able to track facial movements as small as the
motion of individual pores, tiny twitches that the human eye could not observe.
A Latvian computer
scientist feels that AI will replace 99% of human jobs during the next 5 years. Dr. Roman Yampolskiy
predicts that humanoid robots will replace most human workers because the bots are cheaper and more comprehensive in their
thinking.
Microsoft is one of the companies facing immense pressure from the AI
avalanche. CEO Satya Nadella says the company is struggling to stay relevant as other companies spend astronomical sums
to poach talent from each other. Massive layoffs have decimated employee morale at Microsoft while the company commits to
multibillion-dollar investments in the technology. Nadella says he is haunted by what happened to Digital Equipment Corporation
which was made "obsolete" by poor strategic decisions.
Synthesia began in 2017 with the goal of replicating real human
faces. Its latest results are said to be remarkable but unnerving. One volunteer noticed tiny details in the newest
version that were a give-away, while the main problem was an uncanny lack of emotion. Her comment was, "I’m a die-hard cynical Brit
who finds it difficult to inject enthusiasm into my voice even when I’m genuinely thrilled or excited." Her feeling was that she
had been replaced by a polished speaker with the body language of an overachieving male. Other companies like Yuzu Labs,
Creatify, Arcdads, and Vidyard are working on human replications as well. In China, the use of clones of humans has exploded because they
never tire, don't need to get paid, and can sell 24/7.
OpenAI has discovered a troubling reason why LLMs (large language models) tend to
hallucinate, or make things up.
The conclusion is that these responses aren't merely an unfortunate side effect of training methods but
are mathematically inevitable. Generating sentences produces far more errors than simple yes/no queries, as would be expected,
since mistakes accumulate over time. Apparently it boils down to how well the models can tell the difference
between valid and invalid data. Additionally, the bots are given zero awards for uncertainty, prompting them to guess and
creating what is being
called an "epidemic" of dishonesty.
It appears that YouTube has used AI surreptitiously to alter people's
videos
without informing them or asking for permission. Not only small unsettling details are occurring, but the larger issue
is how our connection to reality is being eroded. YouTube claims that it is merely removing noise and blur to improve
quality. Those who own Google Pixel smartphones also get a tool called Best Take which alters people's facial expressions into
an image of something/someone unreal.
Sam Altman, co-founder of Open AI, is behind a brain-computer interface called
Merge Labs
designed to compete with Elon Musk's Neuralink. The startup is hoping to raise $250 TO $850 million USD. Altman's
co-founder of the new venture is Alex Blania, CEO of World (formerly Worldcoin). Merge Lab's technology aims to merge
humans and machines using AI, whereas Musk's vision is to connect human brains to computers. Other companies like
Synchron and Paradromics are working in the same field.
A Japanese team
at Yokohama National University has produced a wireless levitation device that goes up to 10 mph while completely eliminating
friction. The palm-sized device uses accoustic levitation combined with a piezoelectric actuator that creates a
"squeeze film" of air between the surface and the device. The system can support 43 grams of payload in addition to
its own weight of 150 grams.
On to other October treats:
Kerry James
Marshall is exhibiting images at the Royal Academy, London, said to offer a superb overview of Black America. Titled
"The Histories", the show covers the 69-year-old artist's development, from humor to complexity, brutality to innocence,
causing him to be called "the pre-emint painter of Black American life".
The British Museum
presents an exhibition of the origins of Hindu, Jain and Buddhist sacred arts in India and its spread beyond the subcontinent.
We are treated to the daily lives of roughly 2 million people around the world. Included are more than 180 sculptures,
paintings, drawings and manuscripts
Yinka Shonibare concentrates on materials, especially Ankara print, African wax
fabrics that he feels are imbued with the idea of hybridities. The Ankara fabrics, he says, are "Indonesian-influenced fabrics produced by the Dutch
and sold to the African market". He has created The African Libraey, containing about 6,000 books displayed in African
wax prints and celebrating leaders like Nelson Mandela and Kwame Nkrumah. At last year's Venice Biennale, his installation
featured 153 clay objects that replicated items known to have been looted from Benin in 1987.
"Home Less Home" is the title of an
exhibition by Chiharu Shiota at the Institute of Contemporary Art/Boston featring a huge field of red and black ropes
forming the shape of a house. In another installation Shiota used vintage suitcases hung from red rope. These immersive installations
are evocative and seductive, questioning how we collect memories and form connections as we move and travel.
Lesia Khomenko has created large-scale paintings based on deliberately
blurred and obscured selfies of Ukanian soldiers. The muted, washed-out colors suggest how bodies are disintegrated and
mutilated during war. The result is a powerful integration of abstraction and figuration.
Photographer Sally Mann
warns of a "new era of culture wars" after police removed 4 of her celebrated photographs from the walls
of the Modern Art Museum of Fort Worth. Mann's work has been controversial but much admired. These 4 photographs
were part of a series called "Immediate Family",
originally shown in 1992.
Every year 2 Master's students from the Fashion Department at the Royal Academy of Fine Arts
Antwerp
are awarded prizes for outstanding graphic quality. This year, Annaelle Reudink's collection won the prize. Titled
"Too Many Me’s, Not Enough Hangers, Professionally Undecided (working title) or Who Am I Wearing?", these works are lively,
vibrant, and just plain fun.
Karen LaFleur calls herself a
Techspressionist Moving Image Artist. She successfully integrates digital imaging with elements of improvisational
dance movement, producing complex and colorful imagery.
c. Corinne Whitaker 2025
want to know more about the art?
front page , new
paintings, new
blobs, new
sculpture, painting
archives, blob
archives, sculpture
archives,
photography archives, Archiblob archives, image
of the month, blob
of the month,
art headlines,
technology news, electronic
quill, electronic
quill archives, art smart
quiz, world art
news,
eMusings,
eMusings archive, readers
feast, whitaker
on the web,
nations one,
meet the giraffe,
studio map, just
desserts, Site
of the Month, young
at art,
about the artist?
copyright 2025 Corinne Whitaker