ChatGPT launches a $200 pro subnoscription.
This makes a lot of sense.
They only need 1% of their 300 million MAUs to purchase pro to have the money to train GPT5
I won't be surprised if Sora costs even more
They need to do this quickly, before everyone else, to capture the premium market
Paying $200 is good practice. That means your gonna use it to its fullest and launch your side project in less than a month to cover the cost. Big productivity boost.
This makes a lot of sense.
They only need 1% of their 300 million MAUs to purchase pro to have the money to train GPT5
I won't be surprised if Sora costs even more
They need to do this quickly, before everyone else, to capture the premium market
Paying $200 is good practice. That means your gonna use it to its fullest and launch your side project in less than a month to cover the cost. Big productivity boost.
YouTube
OpenAI o1 and o1 pro mode in ChatGPT — 12 Days of OpenAI: Day 1
Sam Altman and some members of the OpenAI team introduce & demo o1 and o1 pro mode in ChatGPT and discuss the ChatGPT Pro plan.
(from left to right): Sam Altman, Hyung Won Chung, Max Schwarzer, Jason Wei
(from left to right): Sam Altman, Hyung Won Chung, Max Schwarzer, Jason Wei
👀2
The biggest threat to Open AI and AI companies in general is big conglomerate unicorns publishing their flagship product, like sora, as an opensource side quest.
Sam Altman is in a tough position. He might be in a situation where its AGI or bust.
Here is Tencent's Opensource Video generation model
Sam Altman is in a tough position. He might be in a situation where its AGI or bust.
Here is Tencent's Opensource Video generation model
⚡2
YouTube
On Memory as a Self-Adapting Agent
We discuss Michael Levin's paper "Self-Improvising Memory: A Perspective on Memories as Agential, Dynamically Reinterpreting Cognitive Glue." Levin is a scientist at Tufts University, his lab studies anatomical and behavioral decision-making across biological…
Evolution creates problem solving agents
We are temporary metabolic patterns
The way nature decodes and encodes information resembles the auto-encoders we use for LLMs
Memories have Agency
Dissociative Personality / Multiple Personality disorder happens when agents go rogue and gain too much power
The mind is a master rationalizer. We will rationalize whatever knee jerk chemical reaction is happening in our brains.
One of the most fascinating podcasts I've listened to in a while
will to code
Evolution creates problem solving agents We are temporary metabolic patterns The way nature decodes and encodes information resembles the auto-encoders we use for LLMs Memories have Agency Dissociative Personality / Multiple Personality disorder happens when…
He raises an interesting question.
The caterpillar is literally getting cooked in its cocoon. It breaks down and reassembles every tissue in its body.
Yet somehow even when every cell is getting rearranged memory persists. The butterfly can still know what the caterpillar knew.
So there's only one way to explain this. Memories have agency. Not just that but they have a life of their own, they want to survive and adapt 🤯
That's not it, All cells have agency, every cell in your body is trying to solve a problem that guarantees its survival 🤯🤯
All cells exhibit neuron-like behavior 🤯🤯🤯
How is it that the memories of a caterpillar that undergoes Metamorphosis persist for the butterfly?
The caterpillar is literally getting cooked in its cocoon. It breaks down and reassembles every tissue in its body.
Yet somehow even when every cell is getting rearranged memory persists. The butterfly can still know what the caterpillar knew.
So there's only one way to explain this. Memories have agency. Not just that but they have a life of their own, they want to survive and adapt 🤯
That's not it, All cells have agency, every cell in your body is trying to solve a problem that guarantees its survival 🤯🤯
All cells exhibit neuron-like behavior 🤯🤯🤯
You are a musical symphony of billions of agents solving a problem
You are the solution to a problem that has endured for Eons
will to code
Listen to podcasts with background music Try it | GitHub
On my long walks, I usually tune in to either classical music (+ Ethio Jazz) or podcasts. But last weekend, I had this idea: Why not mash them together? Like, why not listen to Andrew Huberman with Mulatu Astatke in the background.
[Insert any Podcast] + Ethio Jazz = Literally Heaven ✨
Fr Podcasts need to have background music. Spotify could easily add this feature. Just like they hire video editors they need to hire a film composer like Hans Zimmer
Anyways, after trying this out, I can confidently say I’m never going back to the silent podcast life 😂
Its too awkward. The music just puts me in such a good mood. Instead of walking I'm skipping lol
[Insert any Podcast] + Ethio Jazz = Literally Heaven ✨
Fr Podcasts need to have background music. Spotify could easily add this feature. Just like they hire video editors they need to hire a film composer like Hans Zimmer
Anyways, after trying this out, I can confidently say I’m never going back to the silent podcast life 😂
Its too awkward. The music just puts me in such a good mood. Instead of walking I'm skipping lol
👌3
will to code
Listen to podcasts with background music Try it | GitHub
I built the web app with Svelte which was a great development experience. If you have never tried svelte and use react stack, watch this video, then try Svelte.
Also I tested out WindSurf's AI capabilities, it has really good agentic execution on tasks. Excellent for small tasks but can't build full stack apps yet.
Just like big LLMs are plateauing in their abilities I think we are seeing the same with ai coding agents. I don't think there is an LLM better than Claude Sonnet at coding and I don't think there will be any time soon. Its so good that the only thing holding it back is our inability to articulate what we want. It can't read our minds.
Also I tested out WindSurf's AI capabilities, it has really good agentic execution on tasks. Excellent for small tasks but can't build full stack apps yet.
Just like big LLMs are plateauing in their abilities I think we are seeing the same with ai coding agents. I don't think there is an LLM better than Claude Sonnet at coding and I don't think there will be any time soon. Its so good that the only thing holding it back is our inability to articulate what we want. It can't read our minds.
Google casually stating we might live in a multiverse 🤭
Maybe it's the same way our brains do computation. There's a book about this called Irreducible, which argues that we won't achieve AGI with classical computation, and how our brain queries the future, or according to Google multiple universes
I love this timeline
Also I should read David Duetsch
(The Beginning of Infinity)
Maybe it's the same way our brains do computation. There's a book about this called Irreducible, which argues that we won't achieve AGI with classical computation, and how our brain queries the future, or according to Google multiple universes
I love this timeline
Also I should read David Duetsch
(The Beginning of Infinity)
The fact Vs Code is yet to make anything that can correspond to a response to Cursor, whom by the way forked their entire codebase, gives me hope
A small group can out maneuver and outpace a much larger group
The prize, whatever it maybe, is reserved for the small group that will organize best and act fast
A small group can out maneuver and outpace a much larger group
The prize, whatever it maybe, is reserved for the small group that will organize best and act fast
This media is not supported in your browser
VIEW IN TELEGRAM
When I say "Technology" this is what I mean
🤯3
Gemini_Generated_Image_yifdrxyifdrxyifd.jpg
377.7 KB
Prompt: A swirling nebula forming the shape of an open eye, gazing into the infinite cosmos, with subtle hints of ancient symbols woven into the stardust.
🔥2
I knew Joe would talk about this
This is like when cavemen discovered fire
"Wait it keeps you warm at night and cooks your food 😱"
In 20 years we will have simulated universes 😐
Also we might be in a simulation. Imagine what an alien species that's just 1000 years ahead of us would be able to build
This is like when cavemen discovered fire
"Wait it keeps you warm at night and cooks your food 😱"
In 20 years we will have simulated universes 😐
Also we might be in a simulation. Imagine what an alien species that's just 1000 years ahead of us would be able to build
YouTube
Googles Quantum Computer is Communicating with Multiple Universes? - Joe Rogan
Julian Lennon is a Grammy-nominated singer/songwriter, photographer, author, and filmmaker. His new fine art photography coffee table book, “Life’s Fragile Moments," is available now, as is the Spike Stent remix of his 1998 song "I Should Have Known."
Watch…
Watch…