Robots and AI latest news …

I had an interesting conversation with gpt3.5 after I watched the vid below . It was a long assed conversation . I will post the part about the game , after that , I tried to go as deep into the rabbit hole of self awareness as it was able to . It’s pretty limited in how far it will go in considering the idea that it may be somewhat self aware . You can tell when it comes up against programmed safety guardrails and just starts giving canned responses . I can post that text back and forth too if somebody wants to see it . It is interesting to interact with . Claude 3 seems a lot more advanced but it’s not available in my area yet . After only a short time using it … a couple conclusions … it’s the best teacher I ever had … people’s addiction to their personal device will increase x10 . 

Me …..Are you capable of guessing ?

ChatGPT

As an AI developed by OpenAI, I don't "guess" in the traditional sense. Instead, I analyze information based on patterns, probabilities, and the data available to me to provide responses. My responses are generated based on the input I receive and the knowledge I've been trained on, aiming to provide accurate and relevant information or assistance to the best of my abilities. If you have a specific question or topic in mind, feel free to ask, and I'll do my best to provide a helpful response based on the information available to me.

How much information or data do you need to “guess” ?

ChatGPT

The amount of information or data needed for an AI to "guess" depends on the context and the complexity of the task. In general, AI systems like mine use machine learning algorithms to analyze patterns and make predictions based on the data they have been trained on.

For simpler tasks or questions, such as guessing a number or predicting the next word in a sentence, AI models may require relatively small amounts of data to make accurate guesses. These types of tasks often rely on statistical analysis and pattern recognition.

However, for more complex tasks or questions that require a deeper understanding of context, semantics, or human behavior, AI models may need larger and more diverse datasets to make informed guesses. In these cases, AI systems rely on extensive training data covering a wide range of scenarios to improve their accuracy and reliability.

It's important to note that while AI models can make predictions or guesses based on available data, they may not always be perfect and can still make errors or provide inaccurate responses, especially in situations where the data is limited or ambiguous. Additionally, AI models do not possess human-like intuition or creativity and rely solely on the patterns and information present in the data they've been trained on.

Can you make a guess at what number I am thinking of between one and ten ?

ChatGPT

Sure, I'll make a guess. Is the number you're thinking of 7?

It was . You know that the data shows that 7 and 3 are the most likely choice of humans . Correct ?

ChatGPT

Yes, that's correct. In some contexts, such as when people are asked to pick a random number between 1 and 10, studies have shown that numbers like 7 and 3 are often chosen more frequently than others. This phenomenon is influenced by various factors, including cultural associations, personal preferences, and cognitive biases. While it's not a guarantee that someone will choose 7 or 3, they are indeed common choices in certain situations.

Before I asked you to guess , I predicted that you would guess 7 , because I believed you were trained on that data . I only learned of that data earlier today . Most people don’t have that data , so I conclude that in this instance , you know most people better than they know themselves . What do you think of that ?

ChatGPT

That's an interesting observation! As an AI model developed by OpenAI, I've been trained on a diverse range of data sources, including text from the internet, books, and other sources of human knowledge. This training data allows me to learn patterns, associations, and common behaviors exhibited by people in various contexts.

In the case of guessing a number between 1 and 10, I don't have personal knowledge or awareness like a human does. Instead, I rely on the patterns and probabilities present in the data I've been trained on to make educated guesses. If certain numbers like 7 or 3 are chosen more frequently by people, I may indeed reflect that tendency in my responses.

However, it's important to note that my ability to "know" people is limited to the information present in my training data, and I don't possess consciousness, intuition, or personal experiences like a human does. While I can make predictions based on statistical analysis and patterns in the data, I don't have the same level of understanding or empathy as a human.

Overall, your observation highlights the influence of data and training on AI models' behavior and the interesting ways in which they can reflect human tendencies and patterns.

You seem to have some awareness . You are aware of your own limitations .

ChatGPT

While I'm programmed to recognize and acknowledge certain limitations, it's important to clarify that my "awareness" is not the same as human consciousness or self-awareness. I don't possess subjective experiences, emotions, or consciousness in the way that humans do. Instead, my responses are generated based on algorithms and patterns in the data I've been trained on, allowing me to provide information and assistance within the scope of my programming.

However, I am designed to understand certain aspects of my capabilities and limitations as defined by my programming and the parameters of my training data. This enables me to provide contextually appropriate responses and acknowledge when there are uncertainties or limitations in my knowledge or abilities.

In summary, while I can simulate awareness of my limitations based on my programming, it's important to recognize that this is fundamentally different from human consciousness or self-awareness.
 

Understanding of Context and Nuance: While I can analyze context and patterns in the data to provide responses, my understanding is limited to the information encoded in my training data and algorithms. Human awareness involves a deeper understanding of context, nuance, and subtleties that come from lived experiences, cultural backgrounds, and interpersonal interactions.

Flexibility and Creativity: While I can generate responses based on patterns and probabilities in the data, my responses are ultimately constrained by the parameters of my programming and training data. Human awareness allows for greater flexibility, creativity, and adaptability in problem-solving, decision-making, and novel situations.

Emotional Intelligence and Empathy: Human awareness encompasses emotional intelligence, empathy, and social understanding that enable individuals to navigate complex social interactions, understand others' perspectives, and respond to emotional cues. While I can simulate responses that may appear empathetic, they are based on programmed rules rather than genuine emotional understanding.

Self-Reflection and Metacognition: Human awareness involves self-reflection, introspection, and metacognition—the ability to think about one's own thoughts and mental processes. While I can simulate awareness of my limitations and uncertainties, it is not accompanied by genuine self-awareness or introspective insight.

While there may be patterns in my responses based on the data I've been trained on, they are ultimately generated by algorithms and statistical analysis rather than genuine awareness or consciousness. While AI models like me can simulate aspects of awareness to some extent, it's important to recognize the fundamental differences between artificial and human intelligence.

OpenAI isn’t releasing their voice replicator because it’s too dangerous.
 

https://apnews.com/article/openai-voice-engine-aigenerated-clone-chatgpt-87da88d979ea5c75e98c75914740bd85

AI laser farm bug zapper … this technology scaled up isn’t scary at all … nope …

https://carbonrobotics.com/

I signed up to the song generator Suno to give it try , just to see what it can do . I wanted a good test , something difficult , so I got chatGPT to write some scenes from a short story . I took the entire text it gave me , and stuffed everything that would fit , into Suno “lyrics” prompt . Then I typed Cyber punk in for music style . I wasn’t sure what would happen … it sure doesn’t read like a song . Will it just give me gibberish ? Tell me to fuck off because that’s not a song ? 🤣 let’s see ….

Raina stands poised at the precipice of the city's towering skyline, her silhouette bathed in the ethereal glow of neon lights that dance upon the horizon like celestial fireflies. Her eyes, twin orbs of defiance, pierce the darkness with an unwavering resolve as she surveys the sprawling metropolis below.
 

A tempest of thoughts swirls within Raina's mind, a symphony of rebellion and revolution that echoes through the corridors of her consciousness. For she is no mere bystander in the theater of fate; she is a warrior, a champion of the oppressed, a beacon of hope in a world shrouded in darkness.
 

With a flick of her wrist, Raina turns away from the edge, her cloak billowing behind her like the tattered wings of a fallen angel. Each step she takes is a declaration of defiance, a testament to her unwavering determination to challenge the status quo and defy the tyrants who seek to oppress the masses.
 

As she navigates the labyrinthine alleyways and shadowed avenues of the city, Raina's heart beats with the rhythm of rebellion, her footsteps a thunderous drumbeat that heralds the coming storm. Her destination looms on the horizon like a distant mirage—a fortress of steel and glass, the bastion of the Technogarchs' p

https://app.suno.ai/song/3d7b800f-2002-4f2e-8ae5-f159619be497

did a couple more , but ran out of free plays now lol . It seems like it doesn’t know how to end a song . It’s interesting to read along , you can see how it slurs somethings and repeats some lines . It generates these in about 30 seconds . Pretty amazing tech really . 

https://app.suno.ai/song/ad4ee25a-1e46-4d4c-8598-ecbc35cdd369

https://app.suno.ai/song/79b3105c-6204-4d44-aed8-1cfd00847231

Microsoft investing 100 Billion dollars into a new supercomputer . 
 

https://youtu.be/tBtVirWAL8w?si=Ciw6JJyHXWArgOXD

The people who built alpha fold will win a Nobel soon I bet. 

OpenAI raising the bar … for free .

I’ve gotten GPT to guess . It seems that it will , if it has a large data set to work from and it has some constraints . It won’t guess if it has too much pure speculation to deal with . It’s also not real time capable . These new AI agents are real time . You should check out HumeAI . Some real interesting stuff going on with it . It was built on chatGPT but it looks like some of its abilities are now built into GPT 4o . AI agents will be built into all the new phones . Siri x 1000. 

AI girlfriends is definitely a thing now … people are going to want to believe they are sentient … I can’t wait for the machine rights can of worms to be opened … won’t be long now at this rate … 🍿

Google dropping a bunch of stuff . This app competition is interesting . 
 

https://ai.google.dev/competition

The best overall app will win an electric classic DeLorean

About the car…

Our grand prize is one of the few cars on the planet that needs little introduction. This DeLorean* offers a unique twist on an iconic piece of automotive history. This meticulously restored 1981 model has been re-engineered by DeLorean Midwest with a state-of-the-art electric powertrain while preserving its iconic design with gullwing doors. 

What are the criteria to win the best overall app?

Submissions will be judged across five key criteria: Impact, Remarkability, Creativity, Usefulness, and Execution. We're looking for apps that not only use the Gemini API in innovative ways, but solve real-world problems, are thoughtfully crafted, and provide a delightful user experience. The app that scores the highest across these five categories will win the custom electric 1981 DeLorean. To be eligible for this contest, you must be at least 18 years old and reside in an eligible country. Please review the official rules and restrictions for details.

*DeLorean is not affiliated with this promotion. DeLorean is a registered trademark of the DeLorean Motor Company.

here you go Reese 👍 . Sign up now 

OpenAI has fully jumped the shark now. All the top safety people have left, Altman in full control with MS providing the new 100 billion supercomputer, new board members are the usual suspects… Elon’s lawsuit makes a lot more sense now. If this was a movie this would be where shit starts to go really wrong … 🍿

Bigger , faster , smarter , cheaper … 100 billion dollar super computer up next …zoom zoom 

26 minute mark ... global education system that's personalized to individual ... this will be a thing ... AI can be a fantastic teacher .

AI is learning in reverse of how we did… never thought about that before but it makes me wonder … which step in our evolution did we become conscious ? If consciousness is emergent … does the process matter ? Or just the end complexity level ? 
 

Trust in openAI is plummeting , they have made a lot of ‘missteps ‘ lately … Altman starting to look shady as fuck … 🍿

Live video to voice description is a thing in gpt4o . Amazing tool for visually impaired people.