zsyllepsis

4 ms·

1.
by zsyllepsis 4mo ago
Maybe a charitable reading of the parent comment, but my interpretation of it was that while the _models_ are stateless, modern deployments of these models for inference rely on state. For example, tiered pricing for cached context relies o
2.
by zsyllepsis 1y ago
True, we have been building conversational interfaces with traditional NLP. In my experience, they’ve been fairly fragile. Extending the example you gave, nicely packaged, fully deterministic workflows work great in demos. Then customers st
3.
by zsyllepsis 2y ago
I don’t think that’s what the parent was saying. There are cases when refactoring Rust code where it’s possible to hit limits in the compiler related to e.g. lifetime inference. When these limits are hit, simple straightforward refactorings
4.
by zsyllepsis 2y ago
One potential downside I see with this approach is that it forces you to store and name intermediate results which may or may not be meaningful on their own. Consider a slightly more complicated example that filters on multiple conditions,
5.
by zsyllepsis 2y ago
I think the labels could be much, much worse. They could contain straight noise, just completely random text - not even words. They could also contain plausible, factual text which otherwise has no relationship with the text. I think most c
6.
by zsyllepsis 2y ago
I don’t think this is accurate. While most popular concatenative languages are stack-based, that is not a requirement for the paradigm. The Wikipedia article calls out a few alternatives, such as Om. Source: https://en.m.wikipedi
7.
by zsyllepsis 2y ago
In my experience SageMaker was relatively straightforward for fine-tuning models that could fit on a single instance, but distributed training still requires a good bit of detailed understanding of how things work under the covers. SageMake
8.
by zsyllepsis 2y ago
I think the author’s choice of function to demonstrate purity made this harder to grok as a reader. Asking the reader to “…exclude the I/O interactions…” when considering functional purity makes the analogy much harder to follow. My in
9.
by zsyllepsis 2y ago
The author specifically mentions, and benchmarks against, ripgrep in the linked content.
10.
by zsyllepsis 3y ago
> Making the red go away is important because the red indicates a problem! This is a lot easier than other ways of discovering the error. Why would you want to discover the error later? I don't think that's what the parent to y
11.
by zsyllepsis 3y ago
> If all you have are TS developers, the. IMHO, you don’t have very good developers. Does it change anything to reframe this as having a common denominator across all developers? As in, rather than “All of my developers know only TS” to