I can’t imagine having a hobby that involves passing by, and in some cases climbing over, the exposed remains of others who died doing that same activity.
In other words STPA is a design review framework for finding some less obvious failure modes. FMEA is more popular but relies on making a list of all of the knowable failure modes in a system, but the failure modes you haven’t thought of don’t make it on the list. STPA helps fill in some of those gaps of failure modes you haven’t thought of.
A notebook is a REPL with an inline wiki. Of course it is not intended for running production code, it’s just an R&D environment to document, share, and test ideas
Article mentions people were raiding beaches for free building materials like sand and stone for concrete, which would be a problem at a large volume. Enforcing the same rules on individual souvenir collectors seems excessive.
Way too soon! He recently did an interview with Dave and Krist from Nirvana about the 30th anniversary of In Utero. They described recording prank phone calls during the recording sessions.
Strongly recommend watching Andrej Karpathy’s “Lets build GPT-2” videos on YouTube which dives into an actual PyTorch implementation, then download the code and study it carefully. Then study “Spreadsheets is all you need” to see what the internal data structures look like.
LLMs have a limited context size, i.e. the chat bot can only recall so much of the conversation. This project is building a knowledge graph of the entire conversation(s), then using that knowledge graph as a RAG database.
OpenSearch perhaps? The search query results returns a list of hits (matches) with a text_entry field that has the matching excerpt from the source doc
That’s pretty much correct. An LLM is often used rather like a forecast model that can forecast the next word in a sequence of words. When it’s generating output it’s just continuously forecasting (predicting) the next word of output. Your prompt is just providing the model with input data to start forecasting from. The prior output itself also becomes part of the context to forecast from. The output of “think about it step-by-step” becomes part of its own context to continue forecasting from, hence guides its output. I know that “forecasting” is technically not the right term, but I’ve found it helpful to understand what it is LLM‘s are actually doing when generating output.
For this type of unikernel project C makes sense. I’m a fan of both C and Rust. I like that Rust prevents typical code safety problems, but I like that C hardly changes over time whereas younger languages like Rust are constantly changing. It’s plenty possible to write correct, clean, memory-safe, and understandable code in C, especially if verified by extensive fuzz testing and code scanning.
I wonder how deeply the board and the rotating CEOs actually understand the technology and practicality of achieving AGI, or whether they are falling over themselves due to buying into the LLM hype.
I can believe it. There are those skilled at performative intellectualism, i.e. sounding very smart and insightful, but the performance itself is the product, it is all they sell and offer. Whenever I hear a public speaker appealing to supposed ancient wisdom of obscure tribes and traditions who had better-than-modern diets, laws, monetary policy, medicine, footwear (or lack iof), etc then that tells me that it’s just more hucksterism