Mea Vita: Carpe Diem: AI

Showing posts with label AI. Show all posts

Tuesday, November 25, 2025

AI Meets Expectations: Personal Turing Test

Left: What I got. Right:What it should look like

I’ve had two personal tests that I’ve applied to every new AI model, and for years none of them passed.

One test was to generate an image of an absinthe fountain with glasses and absinthe spoons. Last year, AI got this right.

The other test was to take a multi-hours MP3 that I recorded at monthly jazz jam sessions and break it into individual tracks based on either silence between the songs or applause. This week, ChatGPT finally got it right, but I have yet to see success with other AI models.

Wednesday, October 22, 2025

ChatGPT's Atlas Web Browser

Yesterday, ChatGPT released a standalone Web browser, called Atlas, with ChatGPT integrated into it. You can download it here for macOS only. Other operating systems are coming soon.

As a web browser, it seems OK.
But it really shines when you're on a web page, like a specific restaurant on Yelp, and ask it how long it'll take you to drive there.
"What time should I leave to arrive here by 6:15 today, accounting for live traffic updates?"

Also, since you're logged into your own ChatGPT account, your Atlas website queries are saved in ChatGPT.

Perhaps, one day, Apple will integrate AI into their operating systems so it can answer questions like this based on what you're looking at on your screen.

Wednesday, August 27, 2025

AI Calendar Invites

I live near a religious facility and they recently left a door hanger on my front door to let me know about upcoming events that will create parking issues in my residential neighborhood.

I uploaded a photo of the door hanger to ChatGPT and asked it to create a ZIP file of calendar invites for each event.

Boom! It worked perfectly. I downloaded a ZIP file with more than two dozen calendar invites that I added to my calendar.

Very impressive and helpful.

Sunday, June 29, 2025

AI for Eating

Last night, I was at an Italian restaurant and I wanted to know how these two sauces would look on my cheese stuffed gnocchi.

These photos the are AI-generated images I received after uploading the menu photo above and asking what both dishes would look like.

AI-generated Images

I ordered the second option, the Gnocchi Ragu della Nonna, and what I received matched the image perfectly.

My dinner was spot on

Saturday, June 7, 2025

AI for Data Science with San Diego Data

I asked some popular AI chatbots to analyze the open source data that San Diego provides.

I started off analyzing the data for parking meters, since that's a hot button issue for the city, recently. Two separate chatbots gave me similar answers when I asked what's the most amount of money that these 10,978 parking meters could generate if they were in use continuously for the times they're in effect. The answers came back remarkably close: $75,772.75 and $75,796.75. That's an impressive $24 difference between the two AIs. Close enough for me.

Crimes committed within half a mile of a local high school

Next, I downloaded the 190 MB file of crimes committed in San Diego since 2020 and asked it to plot on a map all of the crimes committed within half a mile of a local high school. I was genuinely impressed with the static HTML file it generated which included listing the crime with each pin. Upon digging deeper, I noticed that I didn't see my police report for my stolen catalytic converter from three years ago. But after a bit of investigation, I realized it wasn’t included in the SDPD’s dataset, so the omission wasn’t the AI’s fault.

Sunday, March 23, 2025

AI Let Down

AI is great at manipulating text and documents, and it does a very nice job at examining photos and transcribing text from images. But today AI let me down with what seemed like a simple task: to take a single, three-and-a-half-hour MP3 recording and break it down into individual tracks.

I used to belong to 3rdSpace, which was a coworking space for creative people. They had monthly Jazz Jams along with other music events. I would typically turn on my audio recorder on my iPhone and let it record the entire event as a single audio file. Occasionally, I’d then take that MP3 file, load it up in GarageBand, and slice it into individual music tracks. But more so than not, instead of doing the work myself, I’d upload the single MP3 to AWS’s Mechanical Turk and pay someone a few dollars to slice up the music into individual tracks.

This morning, I thought, for sure, that one of the big AIs could do this for me. After trying ChatGPT, Gemini/NotebookLM, and DeepSeek, they all failed me. The closest was ChatGPT, which kept apologizing for timing out on me. Gemini referred me to other applications, and DeepSeek was beyond its depth.

Perhaps one day; but, for now, AI’s sweet spot is language, not action.

Wednesday, October 2, 2024

Mind Blowing AI Podcast Creation

This podcast was created for free, entirely by an AI based on approximately 1,200 words of text: http://joemoreno.com/patriotism.mp3

Everyone I’ve shared this with has been absolutely amazed.

Those two voices, in this podcast, are completely AI generated. The source of the text the AI used was this blog post from earlier today.

Try It Yourself

Here's how I did it (you'll need a Google/Gmail account).

1. I visited Google's NotebookLM AI at: https://notebooklm.google.com

2. I clicked the Create button.

3. I then uploaded the text of my document.

4. Under the "Audio Overview" I clicked on the Generate button to create the "Deep dive conversation" with two hosts. It took a few minutes to complete and then I downloaded the WAV sound file. I added one additional, unnecessary, step of converting it to an MP3 using the Apple Music app on my Mac (formerly known as the iTunes app).

Ta-da. That's all there is to it.

Since this is generative AI, every time you use the same content to generate a new podcast, you'll get a different result.

Use Cases

I'm thinking of other uses for this such as turning a long article into a podcast to listen to during a commute or perhaps some work related reading that you can listen to as a podcast while commuting.

Wednesday, August 28, 2024

Hallucinations: How Many ‘R’s Are in the Word Strawberry?

Ask your favorite AI chatbot, "How many ‘R’s are in the word strawberry?"

Most will respond with, "The word 'strawberry' contains two ‘R’s." Obviously, that's wrong – the correct answer is three.

This is the difference between knowing and understanding.

AI models tokenize words. Tokenization is the process of breaking down a stream of text, such as sentences, into individual words and then assigning values to each word in multiple dimensions. An AI model doesn't break down a word into letters, so current models don't use introspection to know what letters make up a word. While an AI model could break down words into letters, the juice is not worth the squeeze when it comes to memory and storage requirements.

In the world of AI, this seemingly confidence, yet random guess, is called a hallucination.

Monday, November 20, 2023

Is the Move Away From Artist to AI a Repeat of the Industrial Revolution?

Is the move away from artist to AI a repeat of the industrial revolution? Progress, through innovation, is a hard force to stop.

The industrial revolution is when the master/apprentice system of building was replaced by product cloning. This era shifted us away from a time when unique crafts were built by a master and taught to an apprentice. [These craftsmen may have called themselves artists – but their work was not the purest form of art. Art with function is actually design.]

Keep in mind that art may seem “unneeded”… it might be easy to dismiss art. “Why do we need the David or the Mona Lisa? Who needs a particular song?" What if we got something equally as artistic, but different? Would anyone miss it?

Alas…Think you don’t need art? Try to get through a pandemic without the artists’ creativity to fill your mind. The artist will take you from deep in their soul to a world away.

The medium and methods may keep changing – sometimes drastically – but the artist will always be there to awe and entertain us. And we will evolve to adapt. We have an excellent track record for surviving unprecedented times. But we can't always beat the odds. The house always wins. So, we must skeptically embrace transformational ideas. A new, good idea ineffectively adopted is a bad a idea.

Endnote: This is my first blog post using custom generated AI images to accompany a piece.

Monday, March 20, 2023

How Will Education Adapt to AIs like ChatGPT?

It's so very interesting to see how society adapts to technology. I'm sure, a hundred years from now, people will look back at us and say, "What was the big deal about AI?" since they would have grown up with it.

When the camera was invented, it was thought that a photo couldn't be copyrighted because there was nothing "created." Obviously, that legal opinion has changed.

We teach Driver's Ed in school, because teens need to learn how to drive in society, but we don't teach "horse maintenance" because we no longer use horses as out primary means of transportation. I'm sure our grandparents were looked down upon by their grandparents for learning to drive and not riding a horse.

Public education teaches personal hygiene, like how to brush your teeth. Now they teach digital hygiene like how to pick a strong password or not leak private data.

It'll take some trial and error, but I'm sure K-12 will eventually figure it out, but only after society, as a whole, figures it out.

Thursday, December 8, 2022

Depth-first Recursion with AI in 103 Seconds

Once, when I was interviewing with the iTunes U team at Apple, I was asked to write a depth-first search, using recursion, in a language of my choice. I chose Java and proceeded to sketch out a tree on the whiteboard while keeping track of my stack on the side of the board.

It took about 10 or 15 minutes. Then the hiring manager and I walked though the code and I was thrilled that I passed, especially since recursion is not my strongest area of coding.

Today, I used ChatGPT which was released eight days ago. It came up with three different solutions in less than two minutes. This is fascinating.

Tuesday, February 1, 2022

The AI Challenge: Feeling and Understanding

A key goal of AI is to reproduce human behavior at scale. AI seeks to reason on input and explain on output while improving (learning) with each experience.

Computers are very good at processing data to turn it into information through logic. Computer systems can then store this information at a global level and attempt to turn it into knowlege by applying it locally. But it's currently very challenging for computers to turn knowledge into wisdom in a personal context.

AI Challenge

The challenge with artificial intelligence (AI) is that computers are inherently thinking machines, trying to imitate humans. However, people are not thinking machines, we are feeling machines. Simply look at a baby which experiences life through feeling, not thinking. If a baby's hungry or tired, they cry. When a baby's entertained, they laugh.

Teaching a robot how to walk is vastly different than teaching a baby how to walk. Robots use precision (digital) logic to balance and walk while babies feel their way to becoming bipeds. I'm not advocating that people always go with gut feelings – there's an old saying, don't believe everything you think – but it seems that we need a more fundamental layer that's missing from AI.

Sunday, November 26, 2017

The Future of AI

Things we enjoy in the real world, we don’t like online. I like it, when I walk into my local coffee shop, that they know my favorite food or drink. But, it seems creepy when I visit Facebook and see ads for products I searched for, days ago, on Amazon.

Like all technology, we’ll learn how these interactions work and we’ll get used to it, especially those who will grow up with it. But we sometimes forget too quickly; especially if we didn't live through it. Even our grandparents are too young to remember a time when people scoffed at teenagers, living a hundred years ago, because they learned to drive a car instead of ride a horse.

AI of tomorrow, like the iPhone X, will know who you are from more than touch (PIN or Touch ID). Like another person, the iPhone X can now see literally see and recognize who you are. And it already knows exactly where you are. Perhaps, its camera and microphone will guess more about what you're doing before you ask, if we can trust it.

Living with AI will be like living with a new species. Almost human, but not quite, probably even after passing the Turing Test.