If you say that it's hard to get the state to put you in jail, then the only way I can reconcile that with facts is that people in the USA commit crimes X10 times more than in other developed countries.

Do you think that's true?

The point is what are the typical use cases for the tool / what are the agreed upon areas of application?

Making the LLM do math with large numbers, I would argue, is not in its typical use case, thought it's at the border.

Asking an image generator model to calculate numbers before running an image sounds definitely NOT like a reasonable use case (do people need it? Will people try using it for this purpose?)

Rigorous understanding of what is over fitting, techniques to avoid it and select the right complexity of the model, etc, are much newer. This is a statistical issue.

My point is that forecasting isn't curve fitting, even thought curve fitting is one element of it.

Is there something in particular that made you conclude that or are you going just with how it felt?

For what it's worth, it didn't seem to me.

That being said, they do have issues with some nationalities. For example, the average American is way too loud for the average japanese place. Even if they think they are being polite, they just talk too loud and too much for japanese sensibilities.

The point is that you can't abstract away the details of back propagation (which involve computing gradients) under some circumstances. For example, when we are using gradient descend. Maybe in other circumstances (global optimization algorithm) it wouldn't be an issue, but the leaky abstraction idea isn't that the abstraction is always an issue.

(Right now, back propagation is virtually the only way to calculate gradients in deep learning)

On the other hand, research on "common intelligence" AFAIK shows that most measures of different types of intelligence have a very high correlation and some (apologies, I don't know the literature) have posited that we should think about some "general common intelligence" to understand this.

The surprising thing about AI so far is how much more jagged it is wrt to human intelligence

One way to peak into the state is to use bayesian models to represent the "belief" state of the bandits. For example, the arm's "utility" can be a linear function of the features of the arm. At each period, you can inspect the coefficients (and their distribution) for each arm.

See this package:

https://github.com/bayesianbandits/bayesianbandits

Why do you think this? The rest of the comment is just rephrasing this point ("llms isn't suited for AGI"), but you don't seem to provide any argument.

Say you are receiving records from users and different intervals and you want to eventually store them in a different format on a database.

Streaming to me means you're "pushing" to the database according to some rule. For example, wait and accumulate 10 records to push. This could happen in 1 minute or in 10 hours. You know the size of the dataset (exactly 10 records). (You could also add some max time too and then you'd be combining batching with streaming)

Batching to me means you're pulling from the database. For example, you pull once every hour. In that hour, you get 0 records or 1000 records. You don't know the size and it's potentially infinite

Why do you think the managers/business owners need an excuse to lay people off? If it's legal and economically beneficial to them, they'll fire people. Having AI won't help them as an excuse. In fact, I would say it sounds like a much worse excuse than "the economy is on a rough spot" or something like that

My use case for fastAPI is very specific (we only maintain APIs for ML models), so I'm curious to learn about this.

The computing cost, on the other hand, is a continuous improvement. If (and it's a big if) a computer can do your job, we know the costs will keep getting lower year after year (maybe with diminishing returns, but this AI technology is pretty new so we're still seeing increasing returns)