Outlier360 Project retweeted
Very interesting...
Today in @AnythingLLM we made any LLM capable of being a reliable AI agent. Whether your LLM provider lets provides tool-calling or not.
Here I take Llama3 running on @ollama and get a Perplexity level experience running all locally on my desktop.
Real-time Web browsing and scraping
Chart generation
RAG and document summarization
Local file creation
All local, all in a single app without writing a line of code.
Outlier360 Project retweeted
Discussing AI/AI news with really great people…
It’s getting wild…
Truly interesting times…
@ceobillionaire
@MetaPrime001
@JohnNosta
@David10x3
…and others…
By: @RoundtableSpace
This quoted post is unavailable.
Worth reading...
17 life-learnings from 17 years of The Marginalian themarginalian.org/2023/10/2…
Outlier360 Project retweeted
Replying to @aniketvartak @DavidDeutschOxf
And brains can be self-programmed...
Outlier360 Project retweeted
If you can not program it, you don't understand it. - @DavidDeutschOxf
Outlier360 Project retweeted
Arete - 'The duty to yourself to be the best you can be; in all that you do'
A Noise and Excuse Removal System...
Outlier360 Project retweeted
Invisible Russian Roulette - [Inductive Reasoning]
Thinking that because you've done a particular thing or things a certain amount of times and got good results, that's the way it will turn out again and again...
You might get a bullet in the future and it may not necessarily be from a future action.
All because you never checked the chamber of your thinking...
Outlier360 Project retweeted
Thomas Edison didn’t just invent the incandescent light, he envisioned and created a system for lighting the entire world...
Outlier360 Project retweeted
AI World Models are crucial…
Right now, LLMs are building World Models, but not of how the world/universe/reality works… but world models of how humans understand the world/universe/reality/each other…
This implications of this seems to be poorly understood.
@ylecun
My tests in Spain with AI shows the AI itself isn’t ready for such a device.
ChatGPT’s world model has many holes that will frustrate users.
Fast Company’s review rubs in AI’s holes in even more ways.
Walk around a city talking to AI and you quickly get frustrated because these systems simply don’t have enough data or are hooked up to enough APIs.
For instance just asked GPT for Supercuts hours. It did fine at that but wasn’t hooked up to its online appointment system.
This is the kind of thing a personal AI will need to do to get everyday consumers excited about using it.
So far that kind of utility is in the future.
It was the promise of the original Siri, which never really arrived.
That said the ability to talk to a device and have it arrange anything is a huge interest of mine so will watch closely.
You buying one? Why?
If I get one it will be to post about it but that is an expensive few X posts. My gadget budget is smaller than an AI pin. :-)
Outlier360 Project retweeted
Talking in the #AITownHall teapawt.net/i/spaces/1PlKQDwePayxE
Outlier360 Project retweeted
Where there's a Way there's a will...
Where there's a Will there's a way...
What if most people are Stupid...?
What if even the most intelligent people are Stupid...?
Is there a way to avoid doing stupid things...?
youtube.com/watch?v=3O9FFrLp…
How to fool the brain…
…and like Richard Feynman pointed to… we are easily fooled…
The reason Why the Brain is easily fooled is crucial to being less fooled…
…and also to #OutlierLearning















