Artificial Intelligence
Generative AI: Creativity at machine speed. Generative AI isn’t here to analyze—it’s here to create. Trained on endless data, it produces everything from ads to art in seconds. For brands, it’s a content powerhouse. For creators, it’s limitless inspiration. For everyone, it’s the future of innovation. What will you make?
The Key to Productivity in the AI Era? It Might Be You
Perhaps the true catalyst for change in our work lives isn't just AI; it might be our own perspective.
John Maynard Keynes once prophesied a future of leisure, where technological advancement – science and compound interest – would slash our work hours. This was in 1930.
As we stand in the age of Artificial Intelligence 94 years later, this vision is tantalizingly close, yet still so far. So far, because the question isn't just about what AI can do for us, but also what we have to demand for ourselves.
Can we redefine success, not by the hours we spend at work, but by the fulfillment we glean from it?
Can we redefine productivity, not by the hours we spend at work, but by the inventiveness we use to solve for work? By the value we create? What is the real value difference between transporting a group of children in an aircraft designed by engineers and accompanying a group of children through kindergarten as an educator for 2 years?
This shift in mindset, coupled with AI's prowess, could unlock the door to a balanced life where productivity is a means to an end, not the end itself.
In this new era, it’s not about AI taking over our jobs, but about us taking control of our time which is limited by the death.
A collaborative, thoughtful approach to integrating AI in our work life holds the key to not just working less but living more. Because life limited only by our time – yes I’ll say it: “death is inevitable.”
Or are we too afraid of falling into an existential crisis? Are those who make our policies afraid that we might fall into an existential crisis?
Thus for the first time since his creation man will be faced with his real, his permanent problem-how to use his freedom from pressing economic cares, how to occupy the leisure, which science and compound interest will have won for him, to live wisely and agreeably and well.
— John Maynard Keynes, Economic Possibilities for our Grandchildren (1930)
Is Data Worth Nothing for Your LLM?
Maybe data is not the new oil.You can have all the content in the world and your LLM still sucks.
You can be the most popular social media platform—be the most popular website of 2021 surpassing even Google.
Youn can collect new user generated data at a neck-breaking speed. And that is just the video data with at least 25 images per second and sound. But on top of that video data you get textual data, descriptive data, usage data, user data and personal information.
And you still are unable to build a meaningful LLM?
Alex Heath journalistic investigations has uncovered that in the race of building generative AI ByteDance, the parent company of TikTok is secretly using OpenAI’s tech to build a competitor LLM:
TikTok’s entrancing “For You” feed made it an AI leader on the world stage. But that same company is now so behind in the generative AI race that it has been secretly using OpenAI’s technology to develop its own competing LLM, including for training and evaluating their model.
So is data worth nothing, or is it?Because what counts is the quality of the data. Building high quality training sets is crucial and cumbersome.
Introduction to Large Language Models
This talk by AI researcher Andrej Karpathy is the busy person's intro to LLMs and a great introduction for anyone who wants to understand LLMs, “the core technical component behind systems like ChatGPT, Claude, and Bard.”
Andrej Karpathy's talk "Intro to Large Language Models" provides a comprehensive understanding of the current state and future potential of LLMs.
He emphasizes the models' simplicity in structure yet complexity in capability, highlights the expansion beyond text generation, underscores the importance of security, and anticipates future developments in AI that are critical for informed business decisions.
Sora text-to-video by OpenAI
The new text-to-video AI model from OpenAI is capable of generating up to minute-long videos from text prompts, including complex scenes with accurate details and emotions.
Sora understands interactions in the physical world, but for now it has limitations, such as difficulties with physics simulation and specific cause-effect scenarios.
Despite these challenges, Sora represents a significant advancement in AI's ability to create dynamic and interactive video content.
What is the most exciting Leap from GPT-4 to GPT-5?
One year after GPT-4 was released we are all excited for a monumental leap forward with GPT-5.
People who have close ties to or are investors in OpenAI seem to be traveling to San Franscico right now as their Twitter status updates show. Is GPT-5 close to its public release?
Sam Altman, CEO of OpenAI, shared his excitement for the most recent podcast with Lex Fridman.
"I'm excited about being smarter," Sam remarked, highlighting the holistic improvement of GPT-5 over its predecessor.
Altman emphasizes that the forthcoming iteration isn't just about isolated advancements. Instead, it's "getting better across the board."
I'm excited about being smarter.\ The really special thing happening is that it's not getting better in one area and worse at others.\ It's getting better across the board.\ — Sam Altman
Lex Fridman eloquently suggests that the true magic lies in the notion of being "heard and understood" by an AI.
He hopes that GPT-5 will increase the moments of "intellectual connection," that mirror these magical moments of connection one feels when another person just "gets you."
This is insofar fascinating as Lex highlights that it isn't solely about raw intelligence or the mechanical processing of language. But the moments when the AI can grasp the deeper questions behind your questions, providing responses that make you feel profoundly understood, even when articulated imperfectly, in the way only a close friend or colleague only grasps one.
…it's just such a good feeling when it got you like what you're thinking about. And I look forward to it getting me even better.\ — Lex Fridman
Both Sam and Lex share the excitement for an AI that does more than just compute or predict. An AI that can truly understand you, that can truly "get" you, making the interaction with it feel more like a dialogue with a thoughtful partner rather than strings of unsatisfactory query-response transactions.
As Fridman aptly puts it, "all of us love being understood, heard, and understood."
You feel like there's an understanding in your crappy formulated prompts that you're doing, that it grasps the deeper question behind the question that you posed. Yeah, I'm also excited by that.\ — Lex Fridman
The anticipation surrounding GPT-5 centers on this very concept—AI that not only hears but understands and responds in a way that feels deeply connected to our thoughts and intentions.
With Altman’s remarks, that the enhancements are not confined to a single aspect but extends across the board the evolution from GPT-4 to GPT-5 promises AI interactions getting more empathetic, more intuitive, and more aligned with a human-like experience and not just a technical upgrade.
Creating a Custom GPT Phonetic Alphabet
Creating a Custom GPT: See how easy it is to create a Custom GPT yourself!
Taking the use case of a phonetic spelling application as an example to create a Custom GPT that you can use as your personal assistant.
BTW, you know what a fun prompt for the “Phonetic Speller” Custom GPT is?
Do you say "gif" or "gif"?
It told me that “[…] the creator of the GIF format, Steve Wilhite, has stated that the correct pronunciation is with a soft "g" (/dʒɪf/), like "jiff".”
I guess the longstanding debate over its pronunciation settled then?
Can somebody please fact check that?
For the entire instructions see Custom GPT documentation here.
Watch the video on YouTube and subscribe for more builds.