Back So Soon?

So, yes, I’ve been away and perhaps overwhelmed with all that is happening at once. There have seemingly always been disparate situations demanding varying levels of concern, but now it seems many are cocked to pop off.

*****

A cerebral vapor lock struck when I started trying to understand the whiz-bang tech phenomenon of late: AI. So much has been said, including the primary fear that it threatens to replace humans. And, as mechanical advancements have replaced people in some blue-collar work, AI is likely to do the same in some white-collar work.

To begin with, I started reading about large language models (LLMs) and Generative AI. We’ve had ChatGPT publicly available for about a year and a half, and it impresses with its well-written and informative responses and astounds with its well-written garbage. Bit by bit, the garbage should recede.

Generative AI can use machine learning to analyze and conceptualize patterns, allowing it to generate (wonder where they got the name) original output, including graphics and music (art?). Large Language Models use Natural Language Processing (NLP) to comprehend and compose text that reads like something a human wrote. One that I looked at starts with a correlative database relating every word to other words producing word vectors. With those vectors known, transformation steps can produce word predictions.

 Some extensive training is required for any AI system, and some will be better at some things than others. The creation of word vectors for any one language seems by necessity to be all-encompassing and requires considerable effort. And no human is likely to be able to match an AI system’s global recall while competing on correlative competence remains a question. Undoubtably, AI systems should generally be able to outperform humans at many things.

But replace? I’ve received ads for AI that will help me write my blog. Do I want that? No.

While AI can produce plausible, well-crafted text and beautiful, complex graphics, it cannot duplicate people on a few levels. The human brain is an organic computer subject to physiology in ways that AI is not, sometimes ailing, sometimes experiencing the waxing and waning of hormones, aging, etc. To say nothing of emotions driving thoughts. Qualia are described as subjective, conscious experiences such as feeling cold, tasting bitterness or smelling something sweet. Or smelling something pleasant. Some mimicking of qualia is possible with sensors and training, but all of those examples are qualitatively subjective, and training would reflect the trainers’ tastes. I guess that would be the seed of an AI’s “personality(?).”

Whatever language is used in training the AI and whatever culture whence it comes will potentially produce huge differences in outputs. AI systems developed in Russia, Japan, and Mexico may not predict similarly.

So those are some human traits that will be a challenge for AI to match if that’s desired. But another facet to consider is awareness. One definition of being aware that I’ve read describes it as being able to detect a change in the environment and respond to it. That’s a very basic level, as the plant on my kitchen table will reorient itself toward the sunlight if I move it. And there’s a lot of software that meets that criterion.

I’ve also pondered whether intelligence implies consciousness. That started with thinking about “market intelligence”, which produces decisions without any one nexus of consciousness.

Do we then think of AI as conscious or merely aware? I’ll posit at this point that it lacks self-awareness. An individual can be self-aware, but AI is not an individual.

A popular, recurring theme with intelligent computer systems is that once they can maintain and produce themselves, they’ll no longer need humans (as with Skynet in the Terminator series). But I think that without self-awareness, AI will lack the survival instinct, and without humans asking AI to do something, it may just be idle.

What do you think?

And suddenly I'm reminded of René Descartes', "I think, therefore I am."

*****

This morning I’m reading excerpts from a New York Times Op-Ed  Why Covid Probably Started in a Lab that seems to validate the idea of the COVID-19 virus having originated in a Wuhan lab that was paired on the same page with An Object Lesson From Covid on How to Destroy Public Trust.

From the latter:

Under questioning by a congressional subcommittee, top officials from the National Institutes of Health, along with Dr. Anthony Fauci, acknowledged that some key parts of the public health guidance their agencies promoted during the first year of the Covid-19 pandemic were not backed up by solid science. What’s more, inconvenient information was kept from the public — suppressed, denied or disparaged as crackpot nonsense.

…

Officials didn’t just spread these dubious ideas, they also demeaned anyone who dared to question them. “Dr. Fauci Throws Cold Water on Conspiracy Theory That Coronavirus Was Created in a Chinese Lab” was one typical headline. At the hearings, it emerged that Dr. David Morens, a senior N.I.H. figure, was deleting emails that discussed pandemic origins and using his personal account so as to avoid public oversight. “We’re all smart enough to know to never have smoking guns, and if we did we wouldn’t put them in emails and if we found them we’d delete them,” he wrote to the head of a nonprofit involved in research at the Wuhan lab.

…

Beaches and schools are open again, but the most severe ramifications of these failures may last for decades, because they gave people cause to doubt the word of scientific and public health authorities [emphasis added – Ed].

While the central point bolded above is spot-on, the writer seems to be more distressed by the fact that many people now distrust the government rather than that the government has proven untrustworthy. She acknowledges that there was false information given regarding six-foot separation and whether the virus could be airborne but she’s all onboard with masks and vaccines. And while she recognizes the hardships people experienced with the lockdowns, she makes no mention of the derision endured by those wanting to know more about vaccine safety or the potential benefits of hydroxychloroquine or ivermectin.

On the blog Cofee & Covid 2024 where I read of these Op-Eds a commenter observed:

The good news is that this shows that the old narrative has lost credibility and those running the show know they are in danger of losing the upper hand. But the fact it's from these two, and in the New York Times to boot, should make clear exactly what's going on. It's a tactical retreat, not an admission of guilt  [emphasis added – Ed].

P.S. After being amongst the chorus who ridiculed its use, Leftist media figure Chris Cuomo has now done an about-face on ivermectin  since he’s begun taking it.