_akhe

Sure and V8 is written in C++ and JS is just the scripting layer

As if there's no downside to building everything in C

I did this with my indoor pots, had a couple of rolly pollies, was like oh cool I'll give them this rock and stick to chill on.

Fast forward 2 years - The Interpot Federation of Rolly Pollies have sent me a declaration of war. They have a standing army of 4 million rollies and a mercenary crew of centipedes.

Their demands are as follows:

- Keep the strawberry tops flowing

- More water but not TOO much water

- The kids want a pump track to roll around on

JavaScript is a hell of a lot faster than Ruby, PHP, or Python, yet Python is mostly used for AI/ML. I think V8 is plenty "fast enough" (why wouldn't it be)?

Naively said. Everything ends up in JavaScript.

Awesome work! Would love to see a video demo, mainly to see how the speech sounds, delays, and general answers.

This is the perfect project for HN - great readme/setup tutorial too.

Amazing work, feels like a step forward for LLM usability.

Would be interesting if there was a web browser that managed the download/install of models so you could go to a site like this, or any other LLM site/app and it detects whether or not you have models, similar to detecting if you have a webcam or mic for a video call. The user can click "Allow" to allow use of GPU and allow running of models in the background.

> Is it nostalgia, or just if it aint broke don't fix it.

I see - yeah I get that, like classic cars, sailboats, SQL, etc.

Good points - sleep walking, even general absent-mindedness. In its most extreme form: https://en.wikipedia.org/wiki/Philosophical_zombie

Kind of a tough read, felt a bit like dancing around a point or not quite getting to one. It seems like the author learned how to use a power drill recently then quizzed a random chat on how they work (something he didn't know 5 mins before), then concluded that the UX is not good based on those results. Might be just the wrong audience, both the author and those in the chatroom might not be the target customer of the power drill, and whoever the target customer is might prefer clear numbers on a dial like that - probably because of paradigms set forth in power tools.

I consider it similar to an audio engineer staring at a complex screen of panels, dials, scales - to a random person it wouldn't be intuitive at all, but to the target customer that's exactly how they are surfacing the important parts of their work to a control space.

By the way - the author ultimately did figure out how to use the drill, can't help but wonder if anyone in the chat had to use a drill for some reason, they would simply watch the same video and easily figure it out too - or more likely just turn it, see the difference and be like "Ah ok" without even understanding what torque is.

Article could use an ending or something definitive - this piece wants to be more insightful. Maybe a few more visuals to keep the reader close to the author's train of thought.

Yes it exists and is probably some biological product of evolution, but the brain is apparently not like a heart that pumps, or a lung that is essentially a bag that pulls oxygen out of air, it's doing something else that isn't explained using any mechanical analogy (AKA the materialist worldview).

In materialism everything we know is machines and mechanisms: From atoms to trees to buildings. It works for explaining most things, but maybe there are some things that aren't machine-like, leaving us with no way to understand it (at least not using those concepts).

Think of it like - you can program a computer to do all kinds of things, but it will never "feel pain" because that is a totally different class of phenomenon that we have no idea how to produce - no amount of code will ever make pain or joy happen to the computer - it's not a code problem. "Feeling of pain" etc. are the primitives of our experience, these little experiences add up to make up our lives and selves, but we have no idea what they are or how to produce them. We only know adjacent physical things that happen alongside it like neurons firing - which not only doesn't even begin to explain it directly but is itself a concept of the consciousness machine we're trying to explain.

Thanks I appreciate it! Maybe when I start producing stuff with it and putting it on itch.io and other places, just need to get it out there outside dev circles :)

Apps? So you would use this OS for the amazing consumer experience? And in your mind is that comparable to a Linux distro like Raspbian or Ubuntu in terms of availability of apps?

Strange way to gauge simplicity too "this screen has 10x less pixels!"

> The question is what do you want to use it for?

Exactly, people replied as if those apps and use cases are actually used today.

So it's just nostalgia right?

> Rent could be in the form of token gifts like a rose or an arrowhead

2024: That will be thousands of dollars every 30 days please. HANG ON. Let me see if you qualify first. You don't. I hereby decline you, not only that but we recorded the fact that you were declined, making it harder the next time you try, kick rocks.

Raspberry Pi is amazing. Everything from the size and simplicity of hardware to that of Raspbian. Super moddable, modular, surprisingly performant. As a lot of others do, I use one for most of my sites and apps as a home server. 15+ apps, sites, and APIs running on this thing 24/7/365 over WiFi.

What would be the benefit of using RISC OS over Raspbian, or even Ubuntu Server? Is it pure nostalgia like running Windows XP on a Pi?

FWIW nobody has created a great JavaScript framework experience yet. Closest we have for RAG is LlamaIndexTS or Langchainjs but both are full of bugs and have little LLM support. Their whole approach to supporting LLMs is writing bespoke wrappers for each.

Maybe it's the same on the Python side, but it feels like nobody has nailed the perfect LLM wrapper library yet. I would focus on dev experience - make it dead simple to load in files and use it from 0-1.

Yeah but that's like saying what propels a car forward is a green stoplight, it's just a correlation you can observe that could have nothing to do with it - might not even be indirectly causal.

When you think deeply about the Hard Problem, it's really The Impossible Problem because there's no way to solve it within our subject:object and "self" paradigms. Yet the problem persists.

Take something like the memory of the first time I tied my shoes - maybe you can find the exact neuronal correlations to that experience I have when I think of that, but they won't be that memory, nor do they even give up the ingredients to engineer that in some alternate setting. The memory itself seems like a completely different phenomenon that the neurons (themselves, concepts in consciousness) can never explain on their own.

Has nothing to do with a pedestal, it's a scientific observation.

> a chaotic thalamo-cortical loop

The aroma of coffee is not explained by the words "chaotic thalamo-cortical loop" - those are words used to loosely describe which neurons activate when I experience the aroma of coffee, but they are not the aroma of coffee, and there's no proof that they are even causal to that experience. It's a logical fallacy because many correlations are not causes but effects, and some are entirely unrelated. Example:

The green stoplight correlates with cars driving, but they don't cause a car to drive. No amount of learning about stoplights will teach you about how a car moves. From far away you simply noticed that the green light is loosely correlated with the cars moving, but that's all you know. That's what is meant by the phrase correlation is not causation - "False Cause Fallacy".

LLMs have nothing to do with consciousness, that's artificial intelligence which as I defined above is not the same as having a sensation. Intelligence (and selfhood, and other psychiatry) are all happening within the conscious process, they shouldn't be conflated with it.

> There's no hard problem of consciousness

If there's no problem - meaning, it's well-understood to the point you can reproduce it - then why has nobody been successful at producing it in isolation? The answer is, we don't know what consciousness is to the point we can reproduce it.

To deny consciousness outright is to deny the first-person observation that you're having right now. Empirical science is based on observational evidence, and there is observational evidence of an experience from your perspective - the fact of your experiential existence owes a grokable explanation - even if it is caused by or correlated with neurons in brains.

Sure, that's a different conversation though. That's like talking about a new brand of beer and then saying "I don't like this beer because there are so many alcohol related deaths and accidents, this will only contribute to more" - while true and you may have a point, there is still a separate conversation where people appreciate beer and engage in connoisseurship etc.

To talk about what entertainment, art, music, etc. might come out of AI I think is a conversation worth having for anyone into those things, either producing or consuming, macro ethics aside.

An attorney with takes like "they don't put the facts from cases into SQL databases to query" yikes! They literally do

> you've clearly never used the software I use

> this tool doesn't know my workflow

> etc.

LLMs already train on knowledgebases like WestLaw. You really think there will never exist an LLM for legal research, etc.? That much is probably happening now, I just haven't heard of the startup.

> Get a f*king grip

So a defendants PII, plea, criminal history, time served, etc. are not important to a defense attorney?

I built a JavaScript library for creating LLM-based game NPCs, here's a video demo: https://github.com/bennyschmidt/ragdoll-studio/tree/master/e...

You drop all their knowledge into a file, like .txt. or .html and it uses RAG to learn everything in the file, so if it's a vendor things like shop inventory, prices, or if it's a storyteller/questgiver you can give it knowledge the player must uncover, etc.

It uses the LLM for natural sounding conversations and instructions ("speak in the style of..." etc.)

It's always exciting when new correlates are discovered, but this doesn't really uncover anything about consciousness per se. At least not as classified by David Chalmers and others who use the term The Hard Problem of Consciousness - which seems to show that consciousness is impossible to understand in our current paradigm. Understand meaning - to the point we could reproduce it.

The Hard Problem is contrasted with the so-called Easy Problems - which are still very hard to solve, and many of them still aren't - but they are ultimately of a class of solvable problems, like neural correlates, the role of dopaminergic pathways, and so-on. These everyday neurological findings are needed to make progress toward understanding the brain, but they don't really have anything to do with consciousness.

Some definitions:

  let intelligence mean the acquisition and understanding of knowledge

  let self mean who you refer to when you say "I" or "me" 

  let consciousness mean raw sensations/happenings: *the aroma of coffee*, *the taste of chocolate*, *the sound of a drop of water* 

Intelligence and selfhood fall under the Easy Problems - both philosophically and neurologically. We have machines that for all intents and purposes are intelligent. Even though it's not perfect, we know what kind of problem that is and how to solve it mechanistically. While nobody has invented a true "self" yet, if we ever find out exactly what conscious experiences are, scaffolding those experiences with self/identity becomes an organizational task (an Easy Problem). Still a very hard one involving neural correlates, cortical pathways, etc. from which we may derive new algorithms and paradigms, but ultimately solvable.

So that leaves the mystery of raw experiences ("the aroma of coffee") which do not seem solvable in any scientific sense. You could say that a digital camera that detects faces in its viewfinder is intelligent, but it doesn't "see" the face, we know that it only unconsciously "detects" the face. So a mystery arises when we examine the human eye, because we see basically the same hardware as the camera: Lens (lens), sensor (retina), wire (optic nerve), chip (visual cortex). Even after mapping the entire visual cortex to all visual features, it just doesn't get us any closer to understanding that raw experience of "sight" that happens when you look at something. It's not the same thing.

I applaud all progress toward understanding the brain, and sometimes wonder why the whole of humanity are not all focused on solving problems like death and aging (by way of solving consciousness), as it seems like pretty much the most important possible thing to all of us!

Favorites on the topic: Donald Hoffman, John Searle, David Chalmers, Dan Dennett, Colin McGinn, Susan Blackmore, Sam Harris.

Sorry - straw manning on internet comments is so bad I shouldn't have even gone there with the crypto analogy, couldn't help because I see parallels with regards to general reception.

I agree with what you said here 100%.

Working with it daily I can't help but be slightly more optimistic though. I see LLMs as being a major component of future apps. You have servers, databases, game engines, and now there's this generative token thing you can use for... quite a lot - without an internet connection no less. It will only get better.

The fact that RAG isolates specific document data in a db and is based on regular database querying IME solves the problem with regular LLM accuracy, but yeah ofc still could be some errors like with anything

Seems like it could be a useful fact check service provider, wherever that might apply (social media apps, news sites, etc.).

Would come down to how nice the libraries are for developers, and how good the SaaS UI and pricing is.

It's not fizzing, you're going to be able to say:

  render 40 hours of Harry Potter film based on every detail of all books

or:

  create a sequel to Titanic where Leo DiCap comes back from the dead and rezzes Godzilla 

and you'll be able to watch that. Imagine that in every content vertical, we haven't even scratched the surface.

The first words of the article

  Lisp programmers claim that

You already know you're in for a fun ride lol

Raspberry Pi at home for several apps and websites. Vercel for front-end React websites that maybe have minor backend work.

No need for anything else. Vercel covers most front-end use cases, and when I need persistence/backend I have my home server. In a couple things I'll use Vercel + MongoDB Atlas so it's fully cloud, but still free. If it's something that wouldn't be free I move it to the Pi.

Basically https://replicate.com/

Because it happens when running your own models on localhost too. I have ollama and all the ones they support, but there are some on HuggingFace I run through llama.cpp inside apps where I won't have ollama installed, Replicate also has Stable Diffusion models, not just chat ones, and OpenAI which is its own thing. So it could potentially all be unified under a provider like that.

Haven't actually tried Replicate because I'm just running locally for free, but probably would try to find a single cloud provider for all deployments, like a Heroku of LLMs.

This is awesome! Very very good work. Simple, friendly product design and I love that it corrects both facts and grammar.

Not only is this an impressive use of LLMs, it's highly relevant to our social media of today, and I can imagine use cases for user-facing apps like giving context to a user comment (which may or may not be factual).

Are there API client libraries where I can see developer usage? That's likely how I would use a product like this.

Very impressive work for high school - I was making Lunar Lander in VB6 with MS Paint graphics at that age :) very cool to see work of this quality from a 16-year-old.