Seems like everywhere you look nowadays a conman lurks — just waiting for the moment to pounce! Instead of helping improve your life, he’s telling fibs and trying to make people more dependent on him, at which point he will “monetize users.” The hot lie du jour is that “AI” is on the verge of consciousness. Don’t believe the hype!
If anything, the “AI” chat bot industry is the greatest concentration of scammers in recorded history.
Unlike those other sites, designed to feed lies and rumors and addiction, this site is here to serve humans, instead of stalking and deceiving them. To the degree that we need money, we will take that sh… stuff offline and not attempt to deceive folks so we can take their money later. As an aside, DARE taught me about that model!
So… what are we doing here, you ask. Well, I’m glad you wondered (in my mind). Organized.info is a public front-end for a RAG that augments the Mixtral model in Petals.
Huh?
Exactly.
Let’s start at the beginning. Ok, right now, there are numerous so-called “AI” tools. Why do I say it like that — why am I using scare quotes? Well, I’m glad you asked that, too (it’s amazing how you ask exactly what I want to answer). AI — Artificial Intelligence — it’s like sci-fi, right??? Wellllllllll, don’t go buying any bridges — that, like so-called AI, is a con. What companies are calling AI is essentially LLMs using next-token prediction. If you remember C-3PO in Star Wars, he had feelings and could make observations and spontaneously tell jokes and whatnot. LLMs cannot. LLMs use prompts because they need some kind of input in order to produce output.
LLMs are very useful for empirical facts, particularly if those facts are written somewhere. For instance, what’s the speed of light on earth? Which chemical elements comprise table salt? Who is the treasury secretary of the US? These are all LLM things. However, things like: what’s a good song? Which athletic activities are fun? Does mountain air or ocean air smell sweeter? Sure, an LLM can quickly summarize all of the data it has and provide a relevant output. However, we are talking about computers here. They do not have the sensory organs, like noses and eyes, that people have.
Anyway, let me get to the point before I ramble about all that.
As you likely know, there are several well-known charlatans trying to capture your attention. However, that is not the only game in town. A few years ago, some grad students made a program called Petals. Like the other “AI” tools, it does inferencing, but instead of using massive data centers, huge amounts of power, and rivers-worth of water, it is distributed, kinda like BitTorrent or Napster. Additionally, because it is distributed, it does not creep on its users.
However, for reasons I cannot fathom, they neglected to make a chat interface, so we made Pollen. Unlike the prevailing chat bots, it doesn’t surveil users, and it’s designed to help you get the gist of something without having to read 100 different sites. The long-term goal is to make you smart and capable, not stupid and easy to manipulate.