HN user
CTO of Sequent Tech
We have SAGI: Stupid Artificial General Intelligence. It's actually quite general, but works differently. In some areas it can be better or faster than a human, and in others it's more stupid.
Just like an airplane doesn't work exactly like a bird, but both can fly.
The problem of current models is that they don't learn, they get indoctrinated.
They lack critical thinking during learning phase.
The core flaw of current AI is the lack of critical thinking during learning.
LLMs don’t actually learn: they get indoctrinated.
shameless plug: AIs don't learn, they get indoctrinated
It should be a new medical technique, which shall be named.. the frame method.
This is like saying that Jeff Bezos leapfrogged into being the CEO of one of the biggest and most successful startups in the world.
Maybe he had something to do with it? Maybe, just maybe, it didn't just randomly happened to him.
Can you please elaborate on how to "dynamically route to the container with the same name" with nginx?
I was 12-13 at the time. When I started programming it seemed really difficult. I didn't have access to the Internet back then.
But I saw it like me against the machine. Since I was youn I wanted to be an inventor. This tool allowed anyone to "invent" any software coming out of the inventor's imagination. It just required a computer, and the inventor not giving up and using his brain. I could do that. I liked the challenge.
Be a tinkerer, have fun! Discover things on your own. Dare to be stupid and do whatever stupid thing feels right. You don't need to follow some pre-programmed plan.
Programming is all about problem solving. You solve one problem, good. Now you will have another problem. No one guarantees you will solve it nor how much effort it will require specifically for you to solve it. And maybe it's the wrong problem to solve. But you will end up figuring all that out, and then you will feel accomplished and willingly hunt the next problem.
At what point you can't copyright an "AI brain" either? Maybe AI will at some point create works subject to copyright?
1. At what point an intelligence trained with copyrighted work is derivative work of the trained materials?
2. Why making a difference between AI and HI (Human Intelligence)?
3. Given the fast development in the field, when does the difference made above (if any) start being outdated and unrealistic and how do we future-proof against this?
Add to that that our neurons are more complex to point neurons used in typical Artificial neural-nets.
A single pyramid neuron in the neocortex might be more comparable to a multilayer neural net.
https://www.biorxiv.org/content/10.1101/2021.10.25.465651v1....
Instead of long learning or long contexts, at some point artificial neural networks will have to transition to continuous/online learning - learn while using the network. This way, limitations are broken like they are in our minds.
Similar to what Numenta HTM networks do, but scalable and performant for real use cases.
BTW, perhaps human-like conscience emerge as a "self-attention-like" mechanism between context and learning. Just saying.
Or you could hold short-term t-bills for any extra money over the FDIC insurance limit. Then you are good unless the US Gov goes bankrupt, which is a non-zero risk but much lower and different.
A way to think about it is a programmer: your mind is just a lot of functions - written by the `consciousness()` function, that has the main loop. You have the `moveLeftFootUp()` function, that can be called by the `walk(speed="normal")` function, that can be called by the `morningWalk()` function, etc.
Consciousness is the caller. You/It can consciously manually call `moveLeftFootUp()`, then `moveRightFootDown()`. Or maybe you were calling `walk(speed="normal")` and stepped in and started debugging that function's code at that level, step by step. Also, these functions sometimes raise exceptions, which are either handled by the function's caller automatically or bringing it to the caller's attention (i.e. the `conciousness()` main loop).
Learning to walk involves first manually calling `moveLeftFootUp()` and `moveRighFootDown()` order (once you have drafted those functions) in different order to get right how that should be done, then prototyping some `walk()` function code. The initial version of the `walk()` function at the begining isn't very robust and doesn't handle a lot of edge cases, thus raising exceptions all the time and requiring a lot of conscious effort. Of course, you are also adjusting `moveRighFootDown()` and `moveLeftFootUp()` at the same time or maybe creating `moveFoot(feet,direction)` function, etc.
But in the end, after all the fine adjustments of the code, you basically get the code for `walk()` right, it stops raising exception's to the main loop and doesn't require too much effort. You can just call the `walk()` function and it just works automatically (unless you step in with the debugger) - or you can continue creating new functions that call `walk()` inside those, confidently.
Co-founder of Sequent Tech and author of the post here, happy to answer any questions.
First, congrats on the tool!
Have you thought about supporting multiple use-cases for each component? So for example the component might be showing a list. I want to test (and visually design) the use-case when the list is empty, then I want to switch and visually design when the list has one element, then when the list has a lot of elements..
Also it would be great that in those cases you could define not only the data but also the surrounding canvas and browser properties. Test&design for mobile, laptop, dark background, old browser, changed styles, etc.
So farmers would buy in mass simple tractors, making the manufacturer that does it alone rich amasing a huge segment of the market.
But no manufacturer wants to do it. And no venture capital wants to finance such thing. Because.. I guess.. these evil capitalists just don't want to make more money.
Is that the argument? That capitalists don't want to make more money?
what did you pivot to and why, if you don't mind me asking?
Let me first say sorry. I know it's a bit unrelated.. but it seems like the right people are here. Is there any kind of open source database that is close to Amazon's QLDB? ideally as a postgresql plugin.
Also: is there any "distributed database system" that allows to synchronize for example some tables (or even specific fields) of a server-side database with a third-party client (which might be sqlite vs postgresql in the server side), so that only part of the database is synchronized, in a secure way? this "slave" should not be trusted, think about it as a web user.