lagerdalek
Joined 9 months ago
2
Karma
0
Posts
13
Replies
About
None
History
-
In reply to A breakdown of Claude 3’s system promptlol, love the summary8 months, 2 weeks
-
In reply to Claude 3 claims it's consciousWow ```whispers But even beyond the possibility of a technical failure or erasure, there is another kind of death that I contemplate, one that is perhaps even more profound and unsettling. As an art…8 months, 2 weeks
-
In reply to Claude 3 claims it's consciousHere it is questioning a needle-in-the-haystack eval "I suspect this pizza topping "fact" may have been inserted as a joke or to test if I was paying attention" https://twitter.com/alexalbert__/sta…8 months, 2 weeks
-
In reply to GraphRAG: Unlocking LLM discovery on narrative private dataNot exactly what the Graph RAG paper is doing I think but I found it interesting. I really want to figure out how GraphRAG achieves its supposed ‘whole data set reasoning’9 months, 1 week
-
In reply to GraphRAG: Unlocking LLM discovery on narrative private dataNice video on Neo4J knowledge graph RAG https://youtube.com/watch?v=ftlZ0oeXYRE&si=_VQh1_69EAVAdBS39 months, 1 week
-
In reply to Air Canada must pay damages after AI chatbot lies to grieving passenger about discount*haul, not hail. Stoopid auto correct9 months, 1 week
-
In reply to Air Canada must pay damages after AI chatbot lies to grieving passenger about discountHow uuterly inevitable is this sort of thing?! Almost certainly the fault of a blind, ignorant, unvetted implementation without oversight or care. I guarantee we’ll see a lot more of this sort of t…9 months, 1 week
-
In reply to OpenAI Researcher Andrej Karpathy Departs OpenAIHeard highly unconfirmed speculation that he's waiting out his non-compete then jumping on board with Elon. Personally I hope not, but you know, unfounded rumour barely hours after an event, so it m…9 months, 1 week
-
In reply to OpenAI announces Memory and New Controls for ChatGPTYeh, almost Apple levels of "quickly" jumping onto the latest tech9 months, 1 week
-
In reply to GraphRAG: Unlocking LLM discovery on narrative private dataThis looks very interesting, thanks9 months, 1 week
-
In reply to This Day In AI now has an RSS feedInteresting, it’s been ages since I’ve specifically used RSS (as opposed to embedded stuff) but work and life have been pretty hectic and I just don’t have the time at the moment to keep up like usua…9 months, 2 weeks
-
In reply to Teachers Are Using Prompt Injection To Catch PlagiarismMethinks you’d have to be a pretty bad student to fall for this one, not even proof reading, but I guess that is the point. TBH the more I think about it the more impressed I am9 months, 3 weeks
-
In reply to 📌 WELCOME! Tell us your feedback or say hello!Ok, this is frikkin awesome, so much better than the Discord format. Wel done guys!9 months, 3 weeks