zby

12 ms·

151.
by zby 2y ago
I am trying to register from Warsaw, Poland - but I always get a 'City not found' error - I've tried Warszawa, Polska and Warszawa, Poland too. It reports the error already at the first letter and never changes. I am using Fi
152.
Separating Data from Instructions in Prompting(zzbbyy.substack.com)
6 points by zby 2y ago|2 comments
153.
by zby 2y ago
In general I don't - but I have some minor cases where I try to fix the JSON - see the tests: https://github.com/zby/LLMEasyTools/blob/main/tests/processo... and https://github.com&#
154.
Show HN: LLMEasyTools – Tools for LLM agents (Python lib)(github.com)
2 points by zby 2y ago|3 comments
155.
by zby 2y ago
The difference between a state and a company is that it is much more difficult to change the first.
156.
by zby 2y ago
But you can create one? If many people wanted to join a cooperative - then it should be easy to do?
157.
by zby 2y ago
Well if you don't allow the traditional company governance and everything must be a cooperative - then these employees cannot choose a traditional company governance. But my comment really was about restricting the choice of a potentia
158.
by zby 2y ago
If you cannot choose to work for an organization that is governed in a different way - then this is a restriction on your rights.
159.
by zby 2y ago
And yet people overwhelmingly prefer to work for these non-employees owned companies instead of working in cooperatives - how do you explain that? They could take those humongous profits for themselves but they don't do that. My answer
160.
by zby 2y ago
To take over the market companies need to attract customers - i.e. make their lives better. Usually there are more consumers than employees - so I think that measure works quite well with optimising for humans. Profit is a side effect of ta
161.
by zby 2y ago
If the model collapse means that the text produced by it is not statistically identical to the garbage that fills the Internet - then I guess a collapse is the goal.
162.
by zby 2y ago
I guess you can imagine an LLM that contains all information there is - but it would have to be at least as big as all information there is or it would have to hallucinate. And also you Not to mention that it seems that you would also requ
163.
by zby 2y ago
Traditional FTS returns the whole document - people take over from that point and locate the interesting content there. The problem with RAG is that it does not follow that procedure - it tries to find the interesting chunk in one step. Ev
164.
by zby 2y ago
A prompt here is a textual presentation of a data structure - this is an area that I am exploring right now. I wonder if it would not be equally effective if the LLM was getting a JSON or Python representation instead of the presumed Markdo
165.
by zby 2y ago
My review: https://zzbbyy.substack.com/p/quick-review-of-claudette
166.
by zby 2y ago
Letting the LLM to decide what to do is a powerful technique. For example one pass RAG is very limited: https://zzbbyy.substack.com/p/why-iterative-thinking-is-cruc... To make it iterative you need the cede the control
167.
by zby 2y ago
In practice this means function calling - the LLM chooses the function to call (and its parameters). Usually in a loop with a 'finish' function that returns the control to the outside code. You can do that without function calling
168.
by zby 2y ago
I am always suspicious with frameworks. There are two reasons of that. First is that because of the inversion of control they are more rigid than libraries. This is quite fundamental - but there are cases where the trade off is totally wort
169.
by zby 2y ago
Seems like the idea of Agentic Rag ( https://zzbbyy.substack.com/p/agentic-rag ). I have created a very bare bones implementation as an example at https://github.com/zby/LLMEasyTools/tree/m
170.
Agentic RAG(zzbbyy.substack.com)
1 points by zby 2y ago|0 comments
171.
by zby 3y ago
I had a knee injury just this February and I was instructed to cool my knee with ice for weeks (I don't think they meant it constantly - but at least from time to time).
172.
by zby 3y ago
Yeah - the idea is simple - but there are so many variations as to what makes a good chunk. If it is a program - then lines are good, but maybe you'd like to set the boundaries at block endings or something. And for regular text - then
173.
by zby 3y ago
What I want is a dynamic chunking - I want to search a document for a word - and then I want to get the largest chunk that fits into my limits and contain the found word. Has anyone worked on such thing?
174.
by zby 3y ago
I have a library with similar api but in python: https://github.com/zby/LLMEasyTools . Even the names match.
175.
by zby 3y ago
They are the Turing Machine without the (infinite) tape. What is your favourite abstraction of computer?
176.
by zby 3y ago
Is there any good definition of low code and no code solutions?
177.
by zby 3y ago
ReAct was used before OpenAI introduced functions. It might be less reliable - but you can add error checking and retry.
178.
by zby 3y ago
Another abstraction layer library is: https://github.com/BerriAI/litellm For me the killer feature of a library like this would be if it implemented function calling. Even if it was for a very restricted grammar - like
179.
by zby 3y ago
Fediverse has a wrong model of news aggregation - this has been talked here many times: https://news.ycombinator.com/item?id=36549218 My take is that it conflates what is civil and what is on topic .
180.
by zby 3y ago
I wish there was some more documentation on dynamic parameters - how they interact with the Workbook editor? A related question - I want to have chains of function calls - so that the prompts should contain all previous function calls and f

More ›