Dagmawi Babi
6.43K subscribers
14.8K photos
1.96K videos
231 files
2.06K links
Believer of Christ | Creative Developer.

Files Channel: https://t.me/+OZ9Ul_rSBAQ0MjNk

Community: @DagmawiBabiChat
Download Telegram
So I did the only next logical step. Exported this channel's data up until last night and use that for the podcast.

(I am so proud of my tool cause it took like 5 seconds or less to generate the book)

#NotebookLM #AIML #Gemini
#DeepDive #Podcasts #AIPodcast
@Dagmawi_Babi
Audio
And using that PDF file here's the episode NotebookLM generated! It's just so good! 🀯

(It's like a really good introduction about who I am and what this channel is about tbh)

#NotebookLM #AIML #Gemini
#DeepDive #Podcasts #AIPodcast
@Dagmawi_Babi
While still in the topic of my own content I made my most read blog post "To my FWB" a podcast and this might be the most entertaining episode ever. They grilled tf out of me and it's just a very crazy episode. Like it genuinely got me hooked that I would like another episode.

#NotebookLM #AIML #Gemini
#DeepDive #Podcasts #AIPodcast
@Dagmawi_Babi
Audio
This is hilarious, absolutely interesting and you just forget that you're listening to an AI generated conversation.

And yes, they grill me so hard I wanted to shout back πŸ˜‚

#NotebookLM #AIML #Gemini
#DeepDive #Podcasts #AIPodcast
@Dagmawi_Babi
I also used what I considered to be my best blog post which's about Bible Translations and generated an episode.

I think it's the best resource about Bible history focused on it's translations.

#NotebookLM #AIML #Gemini
#DeepDive #Podcasts #AIPodcast
@Dagmawi_Babi
Audio
The Episode did not disappoint, it was seamless and gem after gem! It's such a good listen I recommend this to people who want to know about Bible translations, it's THAT good! πŸ”₯

#NotebookLM #AIML #Gemini
#DeepDive #Podcasts #AIPodcast
@Dagmawi_Babi
Audio
HIGHLY RECOMMEND YOU LISTEN TO THIS! πŸ”₯ πŸ‘

#NotebookLM #AIML #Gemini
#DeepDive #Podcasts #AIPodcast
@Dagmawi_Babi
It's actually very powerful and I'm glad Google is the one who made it cause you can also link a YouTube video and it'll generate an episode for you.

So ofcourse I generated one for one of my most favorite speeches by Steve Jobs
β€’ youtube.com/watch?v=UF8uR6Z6KLc

#NotebookLM #AIML #Gemini
#DeepDive #Podcasts #AIPodcast
@Dagmawi_Babi
Audio
Steve Jobs' 2005 Stanford Commencement Address

#NotebookLM #AIML #Gemini
#DeepDive #Podcasts #AIPodcast
@Dagmawi_Babi
I generated a whole ton of deep dive episodes across a variety of topics and resources. But I won't spam this main channel with the files so I've uploaded all of them to my files repository channel where you can join and check out.

Here're the episodes...

Andrej Karpathy's Blogs
β€’ Software 2.0 β€” Podcast
β€’ A Survival Guide into PhD β€” Podcast
β€’ A Cognitive Discontinuity β€” Podcast
β€’ A from scratch tour of Bitcoin β€” Podcast

Research Papers
β€’ Attention is all you need β€” Podcast
β€’ Chains of Affection β€” Podcast
β€’ The Oldest Open Problem in Math β€” Podcast

C.S. Lewis Books
β€’ Mere Christianity β€” Podcast
β€’ The Screwtape Letters β€” Podcast

Other Books
β€’ Measurement by Paul Lockhart β€” Podcast

Enjoy these resources and episodes and if you've created some share with our community! πŸ’›

#NotebookLM #AIML #Gemini
#DeepDive #Podcasts #AIPodcast
#Papers #Blogs #Resources
@Dagmawi_Babi
Please open Telegram to view this post
VIEW IN TELEGRAM
Audio
Before NotebookLM I was impressed by this text to speech model.

(This's a narration of a short sci-fi article Andrej Karpathy wrote called Forward Pass)

#AIML #ShortStory #Scifi
#AndrejKarpath #TTS
@Dagmawi_Babi
Also I am in love with Gemini Flash. It's soooo fast and compact yet so powerful it's so impressive.

Not to mention it's so cheap I could afford it.

#Google #Gemini #LLM #AIML
@Dagmawi_Babi
Intellect 1 is the first decentralized training run of a 10-billion-parameter model, inviting anyone to contribute compute.
β€’ app.primeintellect.ai

Everything will be open sourced; the training framework and dataset.

AI Model specs:
β€’ 10B parameters, 6T+ tokens dataset
β€’ Llama Architecture and tokenizer
β€’ Dataset mix: Fineweb-edu, DLCM, Stack v2, OpenWebMath

The goal is to create the most powerful open source AI model. This excites me so much. Not to mention you can train your own models and it's so cheap.

#Intellect1 #AIML #OSS
@Dagmawi_Babi
This media is not supported in your browser
VIEW IN TELEGRAM
Llama 70B model running at 3,200 Tokens/sec πŸ”₯

#LLAMA #AIML
@Dagmawi_Babi
Google flexing Gemini's context size by allowing it to remember all the conversations you've had before. You can ask what you talked about last week and it'd know.

#Google #Gemini #AIML
@Dagmawi_Babi
This media is not supported in your browser
VIEW IN TELEGRAM
Mercury, the first commercial-grade diffusion large language model (dLLM)! This's so interesting.

Most of the LLMs you've been seeing are clones as far as the core modeling approach goes. They're all trained autoregressively", i.e. predicting tokens from left to right.

Diffusion is different - it doesn't go left to right, but all at once. You start with noise and gradually denoise into a token stream.

Most of the image / video generation AI tools actually work this way and use Diffusion, not Autoregression.

#Mercury #AIML
@Dagmawi_Babi