Its the same as with luddites / textiles, people proved they were willing to accept far worse quality because "automation". People have a strange cognitive bias it seems. When they know a person has done something, the scrutinize and correct. When its a machine / automated, or the price is low enough, it gets hand-wavy.
People here often say "we can't set rules, because imagine how hard it would be to follow them at scale". That same strange cognitive bias applies here. Somehow the "push a magic button and it works" and "of course it can't do it just as well as a person every time" kind of take over.
Imagine if many tech issues like access to pornography, corporate ad tech, violent/racist material, AI slop, etc. were all just armies of people doing it (like google / X actually just had a person curate your results really fast for every search). Of course people would be outraged at the quality, but its "automated" so no one questions.
Similarly, if AI was just a giant outsourced coding firm with 10M programmers that could write you a slop app in an afternoon for $1, people would be complaining at how idiotic it is, and how inferior to regular devs. But instead, the "machine bias" takes over and people try to forgive the shortcomings.
there is something in the human psyche that badly wants to replace itself with a machine, despite it spelling our own undoing, and often resulting in worse outcomes along the way. I suppose the whole "make a tool" drive has been wildly successful overall, but this is probably the required downside to such an innate drive.
Like have you seen the 3d printed houses? They are the uggliest, shittiest houses in the world, but somehow the lure of "yes but we can build them in 24 hours" seems to somehow get people interested. If a human made one of those dung heaps they would be pillaried as a village idiot, and rightly mocked, but somehow because "machine do thing", we should get excited about excreted dung beehive houses...
That was the point of the article, it's not easy at all. You can't just write Ruby in Rust literally unless you want untyped bags of key value pairs and virtual functions everywhere. The languages have different fundamental semantics and there are many non trivial decisions, including and up to, total architecture change.
For example, in ruby some kind of stringly typed reflection utility might Dry out code. In Rust, that might mean writing a trait over some subset of types and a macro helper, but would require careful dosing work.
Letting the LLM decide is the approach in question and the author was pointing out how spectacularly it failed to translate a complete, human designed system.
Honestly it feels like AI companies are still searching for the killer idea that will attract the general public (beyond "be my AI bf" or "better google").
Getting a UI that explains the parts of a bike is ok I guess, but isn't it simpler to get an actual breakdown? Google "parts of bicycle breakout" gets tons of useful images instantly.
Getting a specialized app to split a bill? It was already trivial to put in a calculator if we cared to go item by item on the bill. Having to provide names and tag every item as I go is just more work. Usually real people just go $total divide by 5, I had more, let me chip in an extra $10.
Same with booking travel and wedding plans, these aren't things people would even delegate to a trusted friend usually, much less a one-off request to an AI bot or custom UI.
Most of the useful tasks it can do right now are research, technical question/answer, coding. In terms of "build a flexible ui that solves real-world problem", if existing mobile app isn't useful in this arena, then its unlikely a completely custom UI will do the job.
That said, I've done a few small ones like a quick one to practice alphabet of a foreign language, or prototyping a web game, and the like. But ultimately its nothing that is worth trillions of dollars.
Is this actually good? Seems like they created a new tool that allows LLM to do a lot of numerical analysis. The papers seem somewhat legit?
https://bootloops.ai/papers.html
I suppose if this framework allows physics and other sciences to rapidly progress (e.g. publish more findings), that is good? It destroys a certain model we had, that science should be arduous and require a genius many years of work to uncover something, but if these papers truly add knowledge via this new tool, then this is for the good.
Not sure implications for researchers, but from the public perspective, now we have 36 new papers that increase our knowledge. But maybe an expert can weigh in, the papers might be slop?
I think we can only say we are forwarding human knowledge if a human understands the content they’re putting out in the world. I don’t know if that applies in this case, it I think it’s an important bar to clear.
Yes but people doing research will be asking LLMs to review many papers prior to their reading most likely. If more humans come to understand the findings of these papers, then its good right?
I guess on a theoretical level, if you had a button which would create breakthroughs in a field, but you yourself couldn't understand it, would you push it? I would because otherwise the field may not discover it, and others will be able to understand the breakthrough.
As long as engineers that are AI understand it when they put it to use that shouldn't matter for the practical aspects. Most people already don't understand physics or engineering. It's just keeping the models aligned that's the big issue.
I see the point, but also if these papers provide proof that human can eventually use, or another AI system can use to make further progress, then is it for the best? A great deal of academic papers are drivel couched in dense language and prose to obfuscate the lack of substance. So far, humans rely on search engines and heuristics to avoid wasting time and energy on them. Now, humans are already using LLMs to assist in reading and understanding papers, so perhaps it still contributes to knowledge.
I believe the bar for publication should be "can this be understood by (expert) humans". If that's not the case, it may even be true, but who cares? It's not improving knowledge, which is by definition human understanding.
Perhaps a more apt analogy: AWS unleashed tens of thousands of viruses within its own "sandbox" which then attacked dozens or hundreds of other sites including HuggingFace, the Australian government, etc. They did this after refining the viruses to be as powerful as possible, and after observing inadvertent escapes from the sandbox in the past, and warning the public they were doing this and it was probably going to end humanity. The viruses were created by harvesting code from all known hacks in humanity's knowledge, and using the full backing of AWS hardware so they could "best finish their task" of being flawless viruses / hacks. This entire process was green-lit and encouraged by senior staff and the C-suite. So yes, I would say they should be held responsible.
This is no different than if a gain-of-function virus lab keeps making "apocalypse level" viruses, warning everyone if the viruses escaped it could kill everyone. Despite several leaks, they continue doing so, and even infect live animals to see if it could escape or transfer to humans. The C-suite and senior management push researchers to do even more, automate the virus production and outbreak testing, increase viral effectiveness as fast as possible. Once a leak occurs and starts killing people, would you say "it was just a mistake, no one is responsible"?
Should they be writing quicksort in assembly as a first step? I think its legitimate in this case given that Jev is likely using the same tools as the example. Showing how easily the core is created using those tools helps to dispel some of the mystery and hype.
Example why its legit:
I just invented a new "Regression Estimate Validator" aka Rev. It takes hundreds of input dimensions, then outputs an interpretable score. Its very fast and statistically robust. Response: Ok but you could just use `pytorch.nn.Linear(d_in, 1)`? True, it is equivalent, but that's concealing millions of lines of hand-tuned math libs, CUDA, python, and other stuff.
The fact that there are many lines of code underpinning the target functionality doesn't make it any harder to use, and doesn't increase the value of the sales pitch for the "new shiny thing" using those few lines of code.
However, I do sympathize with your frustration that people can just say "its 1 line of code" when that line is "invoke API" which is really millions of lines / databases, etc. as a way to dismiss legitimate work without understanding its implications.
this is a good point. Similar to minimum wage, if the value of a young hire is zero, then working for free to gain experience is probably necessary (otherwise no one will hire).
We have other structural issues in society that make this difficult now. For example the fact that without a large pay check, few young people will be able to live near the place where jobs are available, afford essentials, and will have no health care.
But yes, if we returned to a culture of "children live with parents until married" and "apprentice at a company until your work is valuable", and companies returned to "hire what you actually need, then invest heavily in employee growth for the long term", and reducing churn of employees (so that employees and employers have incentive to do this), then it might be possible.
So how do you solve the problem of the company investing in their employees to make them more valuable and then the employees go to another company to get paid more (because that company does not spend on developing employing employees) or the employees leave to start their own company to compete with their original employer?
It's moot, because it will never happen. The apprentice system worked because they could perform actual useful work, if only sweeping the floor and such. An apprentice would only slow me down. A Keurig and a Roomba would be far more useful.
"Stealing" changes definitions every few years now, most recently with the mainstream suddenly deciding RIAA and news publishers are no longer the scum of the Earth but their new best friends, and reversing previous definition to now include IP transgression as theft, just so they can say AI companies are stealing shit.
Yes, but the quote about art has nothing to do with that.
The quote says that many artists borrow, meaning everyone still knows who did the original work and the artist is just riffing on it. But great artists transform the work so completely that it becomes theirs.
People here often say "we can't set rules, because imagine how hard it would be to follow them at scale". That same strange cognitive bias applies here. Somehow the "push a magic button and it works" and "of course it can't do it just as well as a person every time" kind of take over.
Imagine if many tech issues like access to pornography, corporate ad tech, violent/racist material, AI slop, etc. were all just armies of people doing it (like google / X actually just had a person curate your results really fast for every search). Of course people would be outraged at the quality, but its "automated" so no one questions.
Similarly, if AI was just a giant outsourced coding firm with 10M programmers that could write you a slop app in an afternoon for $1, people would be complaining at how idiotic it is, and how inferior to regular devs. But instead, the "machine bias" takes over and people try to forgive the shortcomings.
there is something in the human psyche that badly wants to replace itself with a machine, despite it spelling our own undoing, and often resulting in worse outcomes along the way. I suppose the whole "make a tool" drive has been wildly successful overall, but this is probably the required downside to such an innate drive.
Like have you seen the 3d printed houses? They are the uggliest, shittiest houses in the world, but somehow the lure of "yes but we can build them in 24 hours" seems to somehow get people interested. If a human made one of those dung heaps they would be pillaried as a village idiot, and rightly mocked, but somehow because "machine do thing", we should get excited about excreted dung beehive houses...