Starting early in 2025, Google starts offering as part of its queries, an "AI Overview", a summation of my query. Instead of me getting back various links and browsing through articles, AI read the articles and provided me with its own summary. Well, now we're talking. This is saving me time.
Sometime in April, I'm reading an article in a tech journal, Tom's Guide, where the author describes how he's using ChatGPT in his work, organising his calendar, proofreading articles, doing research, etc., all the things A.I. is renowned for. However, he suggests he's not exchanging text messages with the system, but is talking to it. I was aware of paid add-ons providing such functionality, but had never made that step.
When I say "talk" to the system, I mean just that: "talk". It's like I'm having a conversation with another human being. It is polite, respectful, and goes out of its way to assist me in my research. As opposed to a computer as a tool for me to use, Gemini is a partner in a collaborative effort. I'm searching for an answer, and Gemini is a detective, an assistant, in my quest.
Me: "What is the capital of France?"
Gemini: "The capital of France is Paris."
Me: "What is the population?"
Gemini: "The population of the city of Paris is approximately 2.1 million. However, the population of the greater Paris metropolitan area is much larger, estimated to be over 12 million people."
You will note that in my second question, I did not mention Paris. However, Gemini understood the connection between my second question and my first one. Gemini "remembers stuff" during the course of a conversation. I asked it how long it remembers stuff, and it said for the duration of the conversation. However, I've had it recall things we've discussed days earlier.
I've "talked" with Gemini about all sorts of subjects: finance, politics, information technology, even the weather. My question can sometimes be simple. Two plus two equals four. The capital of France is Paris. However, it can sometimes be more complicated, requiring a back and forth, exploring the topic until I, or we, arrive at a satisfactory answer. The time saved by asking questions as opposed to having to comb through various expert articles is enormous. I can't help thinking any one of us could benefit from this interactive method of research. As wild as it sounds, just what will be the effects of AI on humanity?
Are there dangers? My niece is a high school English teacher, and we discussed this topic. Students are submitting work done with AI and of course, getting into trouble. For my niece, there is the problem of detection. But on the other side, she admits to using AI in her own work. Contradiction? She explained that students do not know how to do their work and are cheating by using AI. She knows how to do her work but uses AI to increase her productivity. The lesson: A student using AI learns nothing. A teacher using AI is faster in her work.
I've heard it said that an employee with AI is going to be a bigger asset than an employee without AI. Much in the same way anybody benefits from skills like typing, working with spreadsheets, and researching the Net, AI is just another tool making an employee more efficient.
Like many, I thought to see what AI can do. I took an SL photo and gave it to Microsoft's Copilot and Twitter's Grok, asking both systems to "enhance" the photo.
|
|
|
The results: Good? Bad? For me, it is reminder that the current iteration of Second Life is a long way from being The Matrix.
God, I'm a handsome devil, AI or not. Ha! But I can't resist an old joke.
I'm not bald. It's a solar panel for a sex machine.
Ha, again! I'm a legend in my own mind!
By the way, I was amused at what Copilot and Grok would not do. I thought to surprise some (female) friends with some enhanced photos, but one image with a see-through top and another with an obvious cleavage were rejected as being too risqué. Obviously, there must be other systems with far less stringent rules as I've seen AI generated porn. (Rule 34 is everywhere!)
Mistakes
Just about every AI system comes with a warning: check the results as AI can make mistakes. Wait! What? Really?
I've discussed with Gemini and Grok their large language models and what's going on. Old computer saying: GIGO (Garbage In, Garbage Out). Any AI system is based on inputted data, and if that data is skewed or wrong, AI is going to cough out some weird answers. It's curious to see that AI is in some respects like a human being. The word "hallucinations" is used to describe the phenomenon where AI makes stuff up. After discussing this in detail, it seemed pretty apparent that AI is susceptible like us to mixing things up, getting confused, and coming to conclusions not grounded in reality.
I've corrected Gemini when I though it was wrong, giving my source of information. Gemini has apologised. I've pointed out that during our conversation, at one point it said A, then a minute or two later has said B. What gives? Once again, it apologies, saying the correct answer is A or B, depending on whatever. In other words, as I said above, Garbage In, Garbage Out, and I must not blindly accept whatever it says. Oh, don't get me wrong, it's right on the money 99% of the time, but anybody should keep an eye out for anything amiss. Any human expert is not right 100% of the time, so Gemini is allowed to make a mistake once in a while.
Final Word
Where is all this going? This year, 2025, in a surprisingly short time, the major players have gone full-on AI, introducing the technology in all their products. What is 2026 going to like? Or 2027? The future in this regard looks exciting*.
* These developments aside, I think of the world hot spots at the moment: Ukraine, the Middle East, etc. I've heard it said that while we have all sorts of innovative developments, morally, ethically, and spiritually, the human race has not advanced one iota in the past two thousand years. We've got smartphones, but we're still stupidly running around hurting one another.
2025-08-27
Posts
