But that qualifier in stupid because there’s no start or stopping point for ultra processed versus all foods. Is cheese an ultra-processed food? Is wine?
There actually is a stopping point , and the definition of ultra processed food versus processed food is often drawn at the line where you can expect someone in their home kitchen to be able to do the processing. So, the question kind of goes whether or not you would expect someone to be able to make cheese or wine at home. I think there you would find it natural to conclude that there's a difference between a Cheeto, which can only be created in a factory with a secret extrusion process, versus cottage cheese, which can be created inside of a cottage. And you would probably also note that there is a difference between American cheese which requires a process that results in a Nile Red upload, and cheddar cheese which still could be done at home, over the course of months like how people make soap at home. You can tell that wine can be made at home because people make it in jails. I have found that a lot of people on Hackernews have a tendency to flatten distinctions into a binary, and then attack the binary as if distinctions don't matter. This is another such example.
There actually is no agreed-upon definition of "ultra-processed foods", and it's much murkier than you make it out to be. Not to mention that "can't be made at home" and "is bad for you" are entirely orthogonal qualities.
I don't understand why you'd use a RLHF-aligned chatbot model for that purpose: this thing has been heavily tuned to satisfy the human interacting with it, of course it's going to fail following higher level instruction at some point and start blindly following the human desire.
Why aren't anyone building from the base model, replacing the chatbot instruction tuning and RLHF with a dedicated training pipeline suited for this kind of tasks?
Because the pretrained chatbot is the flagship product of an AI company in 2025. They want to sell this product to customers who can't spell RLHF, never mind have the (substantial) resources to do their own training.
If Anthropic were getting into the vending machine business, or even selling a custom product to the vending machine industry, they'd start somewhere else. But because they need to sell a story of "we used Claude to replace XYZ business function", they started with Claude.
Gold is a terrible unit of international money because the supply isn't flexible enough to accommodate any growth in international trade.
Contrary to popular belief, during history gold has always had limited role in the monetary system, because it was too scarce to really be useful (in most of human history, Silver, not gold was the cornerstone of trade, and trade itself was a tiny part of economic activity in an era where most of it was subsistence farming). It's only when banking and paper money replaced silver that gold took a bigger role in the monetary system. The gold standard is in fact an invention of the late 19th century and it didn't last long before it disappeared progressively (the first world war being the beginning of the end).
Unfortunately for us, it just happened to be the period when a bunch of influential economists grew up (particularly Ludwig Von Mises), and like every human being they assumed that the system they grew up with was special and what came after was decadent, an idea that has unfortunately since then become widespread in the general population.
Most people wrongly assume that the key property for a commodity to become the basis of a monetary system is scarcity, but in reality scarcity is a drawback. Money must be abundant enough (too abundant is bad, but too scarce is even worse).
Your daily reminder that neural network weights aren't creative work and as such aren't subject to copyright protection in the first place. The “license” is purely cosmetic (or rather, it has an internal purpose: it's being put there by the ML scientists who want to share their work and have to deal with the corporate reluctance to do so).
Neither do you need and IDE, syntax highlighting or third party libraries, yet you use all of them.
There's nothing wrong for a software engineer about using LLMs as an additional tool in his toolbox. The problem arises when people stops doing software engineering because they believe the LLM is doing the engineering for them.
Every IDE I've used just worked out of the box, be it Visual Studio, Eclipse, or anything using the language server protocol.
Having the ability to have things like method auto-completion, go-to-definition and symbol renaming is a net productivity gain from the minute you start using it and I couldn't imagine this being a controversial take in 2025…
> I don't know what “tarpit” you're talking about.
Really? You don't know software developers that would rather futz around with editor configs and tooling and libraries and etc, etc, all day every day instead of actually shipping the boring code?
This entirely misses the fact that Spain became a European superpower during that period because war was then done by mercenaries…
> Any wealth created for the state mostly just went into paying for wars because the inflation worked its way up through salaries.
Wealth created for the state just went into paying for wars because that what mattered to the early modern aristocracy and that's what they wanted to pay with their additional money. From the point of view of the Spanish elite of the time, their “wealth” increased dramatically during that period, it's just that this don't fit your or my criteria for wealth.
You only have to show that it fails for square (or any non-spherical shape) cow in vacuum or for spherical cow in air.
reply