rooftopzen

6 ms·

31.
by rooftopzen 1y ago
Share IT is from 2024, but the 2017 tax cut bill is interesting (lots of emdashes there that deviate from the avg) - you’re correct on the additional need for text analysis in this case. Bills I’d found from earlier in 2024 that are publicl
32.
by rooftopzen 1y ago
Exactly the point of the post (900 pages falls in line with “flooding the zone” mentality), but we don’t have data on the intent.
33.
by rooftopzen 1y ago
Lol can you elaborate
34.
by rooftopzen 1y ago
First, see below for Toll et al 2020 and I used autocorrect for grammar. Sorry you were dismissive before looking it up, is more a reflection of your bias. https://liu.diva-portal.org/smash/get/diva2:1591409/F
35.
Trump's "Big Beautiful Bill" likely created with AI – what does this imply?(jonathanbennion.info)
24 points by rooftopzen 1y ago|24 comments
36.
by rooftopzen 1y ago
Compares emdashes per page in the bill vs the same from the average bill sent to Congress
37.
by rooftopzen 1y ago
Important topic but is expected behavior (questionable research if implying this is something that happened randomly): 1) weights change when fine-tuning so applied safety constraints less strong 2) asking a model "what it would do&quo
38.
by rooftopzen 1y ago
lol no comment - the post states: >> In the end, all models are going to kill you with agents no matter what they start out as.
39.
by rooftopzen 1y ago
Also historically happens during wartimes: https://niemanreports.org/press-access-to-satellite-images-i...
40.
by rooftopzen 1y ago
Bad idea. Will be textbook case of boomer toy executives misunderstanding nuances of AI: https://www.psychologytoday.com/us/blog/parenting-matters/20...
41.
by rooftopzen 1y ago
The guys comments on AGI do not follow scientific method - my hypothesis is Meta wants to have enough AGI influencers so that they can dictate definition as goalposts change, not quite based on skill set obv.
42.
by rooftopzen 1y ago
Seems AI generated, if not, nothing new here. Post regurgitates info known for long time and misses largest issues of nuance of “LLM-as-a-judge” as if written in 2023 and audience is living under rock (why?): >> This is where LLM-eva
43.
by rooftopzen 1y ago
>"When it comes to vector search, it's not just about matching words. Understanding the meaning behind them is equally important." This statement ^ is clearly incorrect on its premise -semantic meaning is already vectorize
44.
by rooftopzen 2y ago
If we need a fraud detection algo what tech bro demographics would be stat sig?
45.
by rooftopzen 2y ago
Pypi downloads used to be more granular (can't recall when this changed), but still a publicly avail dataset in GCP
46.
Querying public PyPI downloads to fact-check AI hype(jonathanbennion.info)
3 points by rooftopzen 2y ago|1 comments
47.
by rooftopzen 2y ago
AI development no longer exponential; it appears growth is now linear - is this a surprise?
48.
Querying AI Trends: OpenAI and Azure Peaked in March, Amazon Peaked in June(aiencoder.substack.com)
1 points by rooftopzen 2y ago|1 comments
49.
by rooftopzen 2y ago
Assuming prompt injection issues will never be prevented 100%, is there any solution or will AI tools be stuck here?
50.
AI Prompt Injection: How I hacked Priceline's new AI tool Penny in 2 min(aiencoder.substack.com)
2 points by rooftopzen 2y ago|1 comments
51.
by rooftopzen 2y ago
Here's a write up the differences of GraphRAG - TLDR is it's expensive but can improve accuracy (possibly pertaining to LLM definition of the nodes), so if it's worth it to your app. https://medium.com/@jona
52.
by rooftopzen 2y ago
Did you read it? That paper mentions "Comprehensiveness" and "Diversity" lift to a degree of "substantial improvements over naive RAG baseline" - could they be any more vague?
53.
by rooftopzen 2y ago
AI safety != protection from AGI (lots more to unpack under AI safety umbrella)
54.
Knowledge Graphs in RAG: Hype vs. Ragas Analysis(aiencoder.substack.com)
145 points by rooftopzen 2y ago|21 comments