From Pine View Farm

Geek Stuff category archive

Sliding into the Singularity 0

A man and a computer are sitting at a desk in book store next to a sign reading

Click for the original image.

Share

Artificial? Yes. Intelligent? Not So Much. 0

Truthful? Pigs. Wings.

At Psychology Today Blogs, Timothy Cook (no relation to Tim Apple) offers a four-step process for tricking AI into revealing its biases and fabrications.

Given the hype and hyperbole about these robotic search engines and given how many browsers and websites are trying to hornswoggle us into letting AI bots do our thinking for us, it is a worthwhile read.

Share

Artificial? Yes. Intelligent? Not So Much. 0

A learning aid? Impediment, actually.

Timothy Cook (no relation to Tim Apple) argues that, rather than helping students learn, AI, with its built-in bias towards certainty (often based on stuff AI makes up out of thin air, I will add), will stunt their education. Specifically, it will inhibit their development of critical thinking skills.

He identifies four specific dangers.

      Students lose the ability to sit with “I don’t know.” . . .
      They learn intellectual dishonesty as a strategy. . . .
      They develop intolerance for complexity. . . .

      Most dangerously, they lose their authentic voice. . . .

Follow the link for a detailed discussion of this issue.

Share

Republican Thought Police 0

El Reg reports that the Republican thought police are now trying to ban “woke” AI*. Here are a couple of snippets (emphasis added); follow the link to put them in context.

In an email, Joshua McKenty, former chief cloud architect at NASA and the co-founder and CEO of Polyguard, an identity verification firm, told The Register, “No LLM knows what truth is – at best, they can be trained to favor consistency, where claims that match the existing model are accepted, and claims that differ from the existing model are rejected. This is not unlike how people determine truthiness anyway – ‘if it matches what I already believe, then it must be true.'”

(snip)

“In the LLM world, attempts to ‘un-wokeify’ LLMs have literally produced an AI that named itself MechaHitler,” he said. “This isn’t just a problem in how LLMs are constructed – it’s actually a problem in how humans have constructed ‘truth’ and ideology, and it’s not one that AI is going to fix.”

_________________________

*That is, AI that doesn’t reflect and perpetuate their racism, bigotry, and prejudices.

Share

Artificial? Yes. Intelligent? Not So Much. 0

Citing precedent? Nah. Just making stuff up.

So egregiously that it provoked a judge into kicking some lawyers off a case. Here’s a bit from the story at Above the Law:

When the dust settled, three attorneys — two partners and an of counsel — earned a public reprimand, got disqualified from the case, and referred to the state bar. The court also went out of its way to release without sanction the associates on the case and the firm itself. In a profession where accolades accumulate at the top and responsibility gets pushed downhill, it’s refreshing that the associates who declared that they had nothing to do with the fake citations inserted by a partner avoided punishment for being dragged along for the ride.

Share

The Roll-Back 0

Methinks Bruce Schneier makes a valid point.

However much he wants to, Trump is not going to be able to roll back the clock, but he’s going to do a heck of a lot damage along the way.

Share

The Outers of the Outed 0

At El Reg, Brandon Vigliarolo, using the recent incident at a Coldplay concert as a springboard, argues that we are living a a surveillance state of our own creation. A snippet:

We have cameras everywhere, our personal data has become one of the most valuable commodities in the world, and we’re all perpetually ready to use that tech to make those we feel have violated the social contract pay publicly for their transgressions.

One more time, “social” media isn’t.

Share

It’s All about the Algorithm 0

In a longer article looking at how hate metastasizes, Steven Stosny includes this fascinating and not at all surprising tidbit:

About nine years ago, I opened two accounts on YouTube. In one, I clicked only on progressive-leaning videos, and only on conservative-leaning videos in the other. The two separate algorithms that brought me videos without searching for them described vastly different worlds. Many videos in the two camps used the same news clips, but gave them starkly different interpretations. These videos were an indication that we don’t disagree about facts so much as interpretations of facts.

The common thread in most of the cultural and political posts sent to me by algorithms has been, you guessed it, hate.

Follow the link for context.

Share

Executive Ethics 0

Caption:  Phil's struggle to stay on the oliggarch diet continues.   Image:  Two executives stand in front of a food truck labeled

Click to view the original image.

Share

Artificial? Yes. Intelligent? Not So Much. 0

At Psychology Today Blogs, John Nosta wonders whether we should stick with “not so much,” or should we (these are my words, not his) continue to invite the singularity over for dinner.

Where is Neo when you need him?

Share

Artificial? Yes. Intelligent? 0

A false idol? Sacrifices welcome.

Joe Pierre points out that “(f)or some, deifying AI chatbots as a god-like form of super-intelligence can lead to psychosis.” A snippet:

Anecdotal reports of “AI-induced psychosis”—many of them coming from friends and family—have documented a startling number of people who have developed grandiose and paranoid delusions emerging in the context of conversations with artificial intelligence (AI) chatbots like ChatGPT.

While of the topic of dubious digital deities, check out the discussion about AI on this week’s episode of Le Show. It starts at about the 36-minunte mark.

Share

Geeking Out 0

Mageia v. 9 with the Plasma desktop. The desktop menu is active. The wallpaper is from my collection.

Screenshot

Share

Artificial? Yes. Intelligent? Not So Much. 0

Obedient? Good little Grok is always listening to his master’s voice.

Share

Twits on Twitter X Offenders 0

See foot. Shoot foot.

Share

Artificial? Yes. Intelligent? Not So Much. 0

A reliable witness? That bridge in Brooklyn is still on the market.

Share

Artificial? Yes. Intelligent? Not So Much. 0

Susceptible to suggestion? We’ll find out.

Security expert Bruce Schneier reports that “(a)cademic papers were found to contain hidden instructions to LLMs.”

One example he provides is an implanted instruction for “no negative reviews.” Follow the link for more examples and a link to the original study.

Share

Artificial? Yes. Intelligent? Not So Much. 0

Devolutionary? Quite possibly.

Or you can read the transcript.

Share

Artificial? Yes. Intelligent? Not So Much. 0

A good listener? Could you repeat that once more all over again, please?

Share

Facebook Frolics in the Surveillance Economy 0

The EFF explores the Zuckerborg’s latest scheme for assimilating you and suggests some steps you can take to protect yourself from assimulation. (Audio only. Pretend you are listening to a radio show. Remember radio shows? If you don’t, see the OTR category on the sidebar, over there ——>.)

You can read the synopsis here.

Share

Artificial? Yes. Intelligent? Not So Much. 0

Courting disaster? Why, on bended knee.

Above the Law reports on a “hallucinated” A. I. precedent that, had it not been caught by an appeals court, would have led to a wrong outcome. Here’s a bit:

They cited MORE fake cases to defend their first set of fake cases. Epic. A perpetual motion machine of bullshit, if you will. Seeking attorney’s fees based on a fake case was a nice touch. Probably should’ve thought of that at the trial court level, it probably would’ve worked.

And, in more news of our wandering blithely and stupidly into a singularity of our own making . . . .

Where is Neo now that we need him?

Share
From Pine View Farm
Privacy Policy

This website does not track you.

It contains no private information. It does not drop persistent cookies, does not collect data other than incoming ip addresses and page views (the internet is a public place), and certainly does not collect and sell your information to others.

Some sites that I link to may try to track you, but that's between you and them, not you and me.

I do collect statistics, but I use a simple stand-alone Wordpress plugin, not third-party services such as Google Analitics over which I have no control.

Finally, this is website is a hobby. It's a hobby in which I am deeply invested, about which I care deeply, and which has enabled me to learn a lot about computers and computing, but it is still ultimately an avocation, not a vocation; it is certainly not a money-making enterprise (unless you click the "Donate" button--go ahead, you can be the first!).

I appreciate your visiting this site, and I desire not to violate your trust.