Now our understanding of Musk has changed and the direction is… frightening.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
xAI Launches Grok-3 AI Model, Claims Superior Performance Over GPT-4
- Thread starter MacRumors
- Start date
- Sort by reaction score
If the AI bubble does burst, don’t expect that it means that AI goes away. It will be like the Internet/.COM bubble of 2002. Some companies went under, others shifted their business models and we went on using the internet more and more.
Last edited:
AI as a Tool, Not a Replacement: AI is designed to augment human abilities, not replace them. Just like calculators, computers, search engines, or even written books, AI provides valuable assistance but doesn’t replace the need for critical thinking, creativity, or problem-solving. It can help people access information more efficiently or provide insights, but humans still need to interpret and apply that knowledge effectively.
Improved Cognitive Efficiency: Rather than making humans more inept, AI can actually enhance cognitive function by freeing up mental energy for more complex and creative tasks. By automating repetitive or data-heavy tasks, humans can focus on higher-level thinking, innovation, and decision-making. AI helps optimize the use of human intellectual resources.
AI Encourages Continuous Learning: AI can be a powerful learning tool. It can provide feedback, explain concepts, and present different perspectives, fostering a deeper understanding of subjects. Instead of making people lazier, AI can support lifelong learning by acting as an educational assistant that adapts to an individual's needs.
AI Doesn't Provide All the Answers: While AI can offer answers, it often requires humans to ask the right questions. AI’s output depends on the input and context provided by the user. This means users must still engage their critical thinking and decision-making skills to use AI effectively. Additionally, AI isn't infallible and often needs human oversight to ensure accuracy and appropriateness.
AI Can Empower Humans: AI can democratize access to knowledge and expertise, making it easier for people to tackle complex problems without requiring specialized training. People in less privileged environments or with fewer resources can benefit from AI, enabling them to solve problems they otherwise couldn’t. Far from promoting ineptness, this could level the playing field and allow for more people to engage with sophisticated ideas and tasks.
AI's Role in Creativity and Collaboration: In fields like art, music, and literature, AI is increasingly seen as a collaborative partner. AI can propose ideas, inspire new directions, or help overcome creative blocks. It’s not a crutch for laziness but a partner that enhances human creativity and productivity.
In essence, AI is not inherently a crutch but rather a tool that, when used wisely, amplifies human potential. Like any tool, the way it’s used determines whether it’s a benefit or a detriment. Instead of making people inept, AI has the potential to enhance human capability, encourage learning, and enable individuals to work smarter, not harder.
Note: This answer brought to you entirely by AI.
Yes. I did read somewhere about the perils of training AI on AI generated data and it quickly spiralled into gibberish. ("Model collapse" and "inbreeding"). It can be useful for anonymized data, but creatively is problematic.
But that's just it...AI is a fix for those who know how to use it and best understand it.
Philosophically, I rarely think about such questions in any universal way (for example...is the AVP a good product is filtered primarily though the lens of "is it a good product for ME?" not some universally generic metric).
So when someone is asking is AI a fix? For what? For whom? In what situations?
Those situations are already overwhelmingly positive for those who need it and understand how best to use it. And the use-case for AI is already an order of magnitude larger than those who understand how it could assist them in their work already.
Note: This answer brought to entirely by AI.
This made me smile..
I gotta got the AI boilerplate "vibe" in the first couple sentences
Yeah, see... it's hard to imagine these opinions aren't making it into his AI
For "subjective" use cases yes. Even though on perhaps a tangent, I'd suggest that most humans don't lean into their own subjectivity, but are looking for an objective way to live. To be "right" in the decisions they make, rather than to be uniquely creative.
AI researchers talk about self-learning models that can receive "rewards" by which a reward is judged by being "right." Something that can be measured. So in STEM uses, there will be little need for humans to verify the results.
Not using anything Muskrat produces or is affiliated with
I dislike Mr. Musk because he and I have radically different moral values, and while I tend to understand the subjective nature of my values, my perception of him is that he thinks of his values as objective facts. I tend to see him as blind to his own biases. Many people, both on the Right and the Left, think of their values as objectively true, and I have similar problems with both viewpoints.
The idea he seems to hold that his AI model is going to be more "true" than any other model is a good example of why I dislike Elon. To speak of an AI model as being "true" would suggest that he's adjusted the parameters to fit his own view of what is true.
The idea he seems to hold that his AI model is going to be more "true" than any other model is a good example of why I dislike Elon. To speak of an AI model as being "true" would suggest that he's adjusted the parameters to fit his own view of what is true.
Last edited:
Tried the paid subscription for Grok a few months ago and it was pathetic. Even with paid subscription, it ran out of prompt limits in almost no time. Has nothing to do with the current trend where it’s cool to bash anything with Elon, it’s just that grok is/was simply a pile. Not sure this new and improved Grok would change my mind on ever using it again. Hell, even the free version of ChatGPT has an incredible amount of prompt use in comparison
Last edited:
I wonder if this new "maximally truth-seeking AI" thing will be able to tell us the name of the 150-year-olds receiving social security, 'cos Musk can't. (Rather, it might point out the use of 1875 in Cobol.)
The real questions to ask it is what happened on June 4th 1989 and why does other AI say Hong Kong is a country founded in 1997 🧐
Even though you are both coming at this from very different examples, the larger point about the misuse of these systems (by anyone) is a real problem
One that I would hope we can all agree on
Never. I don't trust anything that has to do with Musk.
Great answer. Much more thoughtful than any other response.
You should do more than that. When was the last time he made a "bold" declaration about a new product from any one of his companies that was actually factually real? Better to presume false until demonstrated otherwise.
Oh, I think a huge part of it is that. It's his behavior that gave them no reason to stop being reserved about an already-held opinion.
Well, if Musk said it must be true, right? I mean he promised the cars he was selling 8 years ago would be self driving by the end of the year and now look at all the autonomous class 5 Teslas driving around.
Last edited by a moderator:
The "Deep Search" that takes its answers from X's fake news is great!
FWIW I am not a Musk apologist by any means (check my post history 😂), and have not used it myself to confirm (and have no intention of using it), but people I trust in the AI space who have used it are saying it is very impressive - and some of these people would be much happier if they could say it sucked.
I’ve used ChatGPT and Grok. Grok is hands down better. So many people won’t use it because of Musk but they are narrow minded.