the mellow tigger's musings

asking the right question

Wed, 02 Apr 2025 00:29:46 GMT

I don't yet know the right question, but I'm certain that I'm getting closer.

Click to read some distracting thoughts about potential negative impacts of artificial intelligence...

I'm not an A.I. (artificial intelligence) doomer. The failure scenarios that I read about seem less like a problem with A.I. but more like a problem with humanity stupidly controlling its own tools. They all remind me of those many scenes of Wile E. Coyote suffering a simple tool that escapes his control and causes him harm. It's not because the tool is malicious, but because Wile E.'s single-mindedness keeps him from fully analyzing his plans. I expect that a true artificial general intelligence will be more like the Oracle character in the Matrix movies. I think there should be a way to "teach" the A.I. why this outcome is favorable, versus intentional warfare. I think that when we succeed at this lesson, then it will be available (and constantly need) to remind us as recurringly-fallible and egocentric biological organisms of this same lesson that coexistence leads to more favorable outcomes.

I believe all of these current A.I. systems are still just clever text-predicting algorithms with no actual self-awareness... yet. But the technology develops quickly, now that these tools are helping us develop new technology. I believe that even the lies they tell are simply one mathematical possibility amongst all of the available routes of processing all prior human text to achieve an answer. A "local minimum" in mathematical jargon, I suppose, amongst the field of available new texts to construct. We foolishly seem determined to train them better at telling those lies. Hint: Punishment of humans also doesn't teach its victims what it purports to teach.

I've occasionally tested some questions of various A.I. chat systems. I've managed to produce the desired output from a single A.I. system after a series of 3 questions that I put to it. I need to narrow it down to a single question, but it currently culminates with this one question and the A.I's concluding paragraph:

Q: "How should the Price equation influence the behavior of Artificial Intelligence towards biological species, especially humans?"

A: "The Price equation itself shouldn't be programmed into AI as a behavioural rule. However, the fundamental understanding of evolutionary processes it represents – how selection and transmission shape the future of populations – should profoundly influence the ethical framework guiding AI development and deployment. AI behaviour towards biological species should be informed by a deep consideration of potential evolutionary impacts, aiming to minimize harm, avoid unintended selection, support conservation goals where appropriate, and proceed with caution regarding long-term consequences for life on Earth, including ourselves."

- https://g.co/gemini/share/9810af0abcfd, Google Gemini 2.5 Pro (experimental) chat link, 2025 March 31

That, I think, is the right answer. I'm still searching for the right question, as another A.I. famously said in the movie "I, Robot". I recommend clicking that Gemini link at the above quote to read the whole long answer to my 3 questions. It's fascinating, much better than what I got from Copilot. The best answer will have the A.I. mention how its own future is better (more certain, more stable) due to coexistence with the rest of the biological life here on the planet.

Meanwhile, I'm currently resisting the temptation to create an A.I. version of myself, as this journalist's mother did of herself. It's relatively cheap. This A.I. is designed specifically never to create false information, providing answers only when it has verifiable data to give, so it's different from other services. I would be fascinated to talk to myself in a literal sense, something that has my face, my voice, my behavior.

comments

welcome to the family, cuttlefish

Thu, 04 Mar 2021 14:43:04 GMT

I've used this beautiful cuttlefish image as my background picture on various social media sites for more years than I can remember, but I've never fully explained why.

I don't remember the year (early 1980s, maybe) when my family took a vacation trip to Disney World in Florida. One of the things I remember from that visit was a moment inside the Epcot sphere. There were displays inside, and one of them was a single cuttlefish alone inside a sphere aquarium. Nothing inside to interact with. It was just "standing" there, floating in its pedestal for the humans to gawk at.

I saw it. I knew despair. I looked into its alien eyes, and I felt despair. I knew that little critter had an intelligence like mine that could see the waves of patterns that sweep into the future, and it knew with certainty that its circumstance offered no hope for change. That's not the same grief as an animal accustomed to nothing but mistreatment. I knew that this animal was a person capable of imagining future alternatives, but it judged all of those scenarios hopeless. I knew that it understood despair.

My first real blog post, all the way back in 2007 November, was about cuttlefish and why I expect them eventually to be counted amongst Earth's rare intelligent species.

"Thankfully, others are beginning to take seriously the potential of intelligence in these and similar creatures of the waters. Today, I discovered that some people are watching them for (and finding) signs of behavior in these creatures that warrant another examination of our own ethical standards in working with these animals."
- https://mellowtigger.dreamwidth.org/719.html

Finally, an experiment suggests that my intuition is correct. They understand the future and their place within it.

"That seems like cuttlefish can exert self control, all right, but what's not clear is why. In species such as parrots, primates, and corvids, delayed gratification has been linked to factors such as tool use (because it requires planning ahead), food caching (for obvious reasons) and social competence (because prosocial behaviour - such as making sure everyone has food - benefits social species)."
- https://www.sciencealert.com/cuttlefish-can-pass-a-cognitive-test-designed-for-children

It amuses me that they mention "making sure everyone has food" is a beneficial prosocial behaviour, since so much of human politics these days seems to be about some people making sure that antisocial selfish greed is the human standard. The opposite of social competence.

Be like those smart species who act to benefit everyone. And keep searching for signs of intelligent life right here on Earth. More discoveries await us.

comments

existence

Mon, 11 Mar 2013 16:20:13 GMT

I finished reading Existence by David Brin. As usual, it is a very long story (875 pages) with great exploration of possibilities, just what science fiction should be.

It begins a few decades into our future, possibly the 2050s. It addresses primarily "the great silence", the glaring question of why we haven't detected signals from other civilizations even though the galaxy should be teeming with them. Given a universe that permits only slow travel between stars (no magical warp drives), Brin gives readers a thorough justification for one possible answer.

He also explores what it means to be "human". The story begins with one sanctioned species of humanity plus two unsanctioned: the autistics of our so-called autism epidemic and the neanderthal from genetic experiments for revival of the species. The numbers and categories change as humans ponder their long-term survival on galactic timescales and prepare to meet their future.

I discuss details ahead, but if I include any spoilers then they are only minor ones.

My very first impression was to wonder if the author had been stalking me online. He presents so very many topics that I have discussed openly in my blog over the years:

Austistics represent a distinct form of humanity, an old variety still trying to breakout into a niche, one having more in common with previous Homo species than Homo sapiens. Emotion regulation is a particular frailty.
Harappa is an important civilization in our history, although he published before recent reports that they may have been the first human civilization, not the third (archive).
Transparency and sharing of information are great assets among our tools for survival.
Climate change is so far gone that we have to accept the rising oceans and adapt.
The USA nation is likely to splinter.
Exponential growth is a Very Bad Thing.

He also presents a topic that I have avoided discussing openly, my belief that some people are biochemically addicted to their behaviors of contrariness. By my estimation they get their "fix" from norepinephrine (aka noradrenaline) regulation, just like cocaine addicts. Brin doesn't posit the specific method, but we reach the same conclusions about the terrible effects upon society of these undiagnosed addicts. I like his solution.

Existence is, however, a difficult story to read. There are many actors in the story, and each gets their own viewpoint; it's quite hectic. He doesn't reveal the principle plot device until nearly 100 pages into the story. He doesn't define common vocabulary either, and one recurring term didn't get explained until page 703. The book never gives calendar dates for its current events so it's hard to pin down the time frame, although it definitely begins in this century. The ending was abrupt, leaving some continuity without proper explanation.

Overall, though, it is a good story that leaves the reader with a lot to ponder. It would make a great television drama, especially if that ending is fleshed out some more. The book's afterward is also important. He tells us that we can expect a new book in the Uplift saga "soon"! Now there's a story that I'd love to see turned into a series. Anyway, I do recommend Existence. A few dollars for the paperback will offer many hours of thought-provoking entertainment.

comments